Aaron Bergman's tweets, June 2025
Every public tweet, reply, quote tweet and retweet that Aaron Bergman (@AaronBergman18 on X, formerly Twitter) posted in June 2025 and has not deleted: 802 posts, oldest first, times in UTC. Each heading links to the original post on X. When Aaron replies to or quotes someone, their tweet appears as a blockquote with their name; those words are theirs, and everything outside the blockquotes is Aaron's.
This page is one month of the complete archive of Aaron Bergman's tweets, which updates automatically every few hours. Previous month: May 2025. Next month: July 2025.
Sunday, June 1, 2025
Reply to @paularambles, 05:21 UTC · 7 likes
“paula” (@paularambles): people here are simply not okay
Here’s the link https://partiful.com/e/JVTpIurbhmVLgvwXuo8d
Reply to @NoahTopper, 05:21 UTC · 4 likes
Quinoah 🔍⏸️ (@NoahTopper): Meeting the famous @captgouda24 (and also @EneaszWrites!)
Omg hello all!
Quote tweet, 05:30 UTC · 97 likes
I am once again tapping this sign for people of the female persuasion
(the takeaway is that even if you have a preference for more height all else equal, you might do better by sorting for *shorter* men if your height preference is less strong than avg)
Quoting AF Post (@AFpost): Tinder is implementing a height filter for users.
The change will allow women to generally filter out men below their preferred height.
Follow: @AFpost
Thread, continued, 05:30 UTC · 21 likes
This is a great example of motivated reasoning on my part happening to give the right answer
Tweet, 06:14 UTC · 6 likes
Friend asked for help with R
I was excited and then disappointed
Reply to @TheWarMonitor, 06:29 UTC · 2 likes
WarMonitor (@TheWarMonitor): Palantir’s platform is able to merge vast datasets. Bank accounts, medical records, driving record, debt, disability status, political affiliation, credit card expenditures, purchases, tax filings, travel records, phone records, creating detailed profiles on every American.
(quoting @HighImpactFlix: This is serious. Trump's partnership with Palantir Technologies aims to implement a March 2025 executive order mandating federal agencies to share data - YOUR data. This initiative has one goal: centralize ALL government data, leveraging Palantir’s Foundry platform. Palantir’s…)
As a guy idk sounds sus but as a data nerd this is kinda cool
I want the forbidden csv
Reply to @pli_cachete, 16:54 UTC · 5 likes
Rota Is In Israel🚪🧎♂️ (@pli_cachete):
Extremely based
Reply to @AndyMasley, 17:10 UTC · 4 likes
Andy Masley (@AndyMasley): People look at me like I'm a little crazy when I say the most common criticism of EA I bump into on the ground is people saying it's a stealth way of distracting people from a socialist revolution.
Wild
Reply to @Kirsten3531, 17:17 UTC · 6 likes
Kirsten (@Kirsten3531): I married a man shorter than me and I recommend it!
Based
Quote tweet, 17:50 UTC · 10 likes
Very hesitant to celebrate death but this is pretty crazy + badass
Quoting Maria Avdeeva (@maria_avdv): This will be in textbooks.
Ukraine secretly delivered FPV drones and wooden mobile cabins into Russia. The drones were hidden under the roofs of the cabins, which were later mounted on trucks.At the signal, the roofs opened remotely. Dozens of drones launched directly from the trucks, striking strategic bomber aircraft.
And — Russia can’t produce these bombers anymore. The loss is massive.
Nothing like this has ever been done before.
Quote tweet, 18:52 UTC · 51 likes
I did a tree search of most* of the Substack recommendation space - i.e. who is recommending who on Substack, seeded with Astral Codex Ten
The result is ~650k recommendations mentioning ~130k blogs, a substantial chunk of Substack
Data, go nuts! https://drive.google.com/drive/folders/1jjt5RTMFmtCzn_YG1VJ084uoySIveKYa?usp=sharing
Quoting Aaron Bergman 🔍 (@AaronBergman18): Vibecoding is great because you can be extremely extra at basically no cost
Thread, continued, 18:52 UTC · 8 likes
* A vibe-coding introduced "bug" (ie me fucking up by not checking this) is that it only scraped the first 50 recommendations by any one blog, so any recommendations 51+ got lost
(There is some virtue here, diminishing fake signal of any hyper-recommenders)
Thread, continued, 18:52 UTC · 3 likes
Tragically I had pretty good reason to do this but not to spend the extra idk 5 hours vibecoding a really cool interactive vizualization so maybe someone else wants to do that for fun
(Way too much data for naive python-to-html-render that claude spits out)
Thread, continued, 18:52 UTC · 3 likes
I didn't count exactly but this took a total of at least a day I think using 100 parallel proxies, so it would take a couple months just running on a single machine/IP
Thread, continued, 18:54 UTC · 4 likes
Reply to @spqr_sulla, 18:57 UTC · 204 likes
Sulla (@spqr_sulla): Interesting how we went from water fountains primarily for direct drinking to wattle bottle refilling stations. Why did people start carrying water bottles around everywhere ?
I mean maybe but the existence of water filling function doesn't mean it's being used more than the fountain
Kinda seems like a "tech just gets a bit better and we get a bit richer over time" thing
Tweet, 20:23 UTC · 8 likes
Not insightful but it’s pretty depressing to see the US have incredible levers I so many important domains and the moral and strategic ability to pull them and then we either don’t touch them or push the levers the wrong way
Thread, continued, 20:23 UTC · 6 likes
- everyone wants to live here, high skilled and low skilled. We don’t let them
- everyone wants to study here, same deal
- Biden could and should have spent his progressive points and political capital pressuring Bibi over Gaza
- AI export controls
Ahhhhhhhh
Thread, continued, 20:32 UTC · 7 likes
I know political constraints are “real” in some sense but it’s very frustrating that I (perhaps very naively) think I see how the world could just be way better with relatively few changes:
1) braindrain China + export controls = minimal AI competition
2) strong federal incentives for very narrow “transformative” AI in a few domains: defensive tech, cultured meat, and if possible both human (ie biomedicine) and animal welfare biology
3) lots of incentive and smart regulation for AI safety, probably something like SB-1047 at the federal level
Not a fixed world! This still leaves much Yudkowskian “alignment is hard” x-risk as well as the AI sentience shitshow, but even those are probably mitigated!
Quote tweet, 20:35 UTC · 4 likes
Does anyone have a good explanation for why Biden was willing to play progressive in a bunch of domains but not Israel
Quoting Aaron Bergman 🔍 (@AaronBergman18): - everyone wants to live here, high skilled and low skilled. We don’t let them
- everyone wants to study here, same deal
- Biden could and should have spent his progressive points and political capital pressuring Bibi over Gaza
- AI export controlsAhhhhhhhh
Reply to @PeterJAngel, 20:42 UTC · 2 likes
Peter🧅📈 (@PeterJAngel): Personal conviction?
My impression of Biden is that he was (relatively) attentive to foreign policy and had strong views on Israel before becoming president
Yeah makes sense
Reply to @keyofgeo, 21:49 UTC · 4 likes
Veles (@keyofgeo): my office still has their COVID era “do not drink directly from this fountain, use it to fill a water bottle” signs
Lol
Tweet, 21:50 UTC · 114 likes
I am a machine that turns rice cakes into telling former hinge dates who friendzoned me that they should apply for 80k coaching and seriously consider trying to make a billion dollars
Reply to @Benthamsbulldog, 22:00 UTC · 6 likes
Bentham's Bulldog🔸 (@Benthamsbulldog): I have an article about why people should try to be a billionaire coming out soon!
Based
Thread, continued, 22:06 UTC · 3 likes
Check ur DMs
Tweet, 23:11 UTC · 5 likes
Hate to give them free advertising but this might be the dumbest thing I’ve ever seen
Literally trying to time GI motility + digestion down to the minute (the delay mechanism is the capsule lmao)
Thread, continued, 23:11 UTC · 3 likes
Too early and it wakes you up artificially, to late and their whole claim is that this avoids the 20 min absorption period (how tragic) from taking a normal caffeine pill on waking
Monday, June 2, 2025
Reply to @livgorton, 00:29 UTC · 1 like
Liv (@livgorton): i recommend literally not looking at height on apps 👀 many amazing short kings out there. often the preference probably actually isn't that rigid if you met someone IRL and witnessed them as a complete person so idk seems sort of arbitrary
based
Reply to @AndyMasley, 00:32 UTC · 11 likes
Andy Masley (@AndyMasley): On "EAs don't value great art" I think there's something beautiful and emotionally mature about the idea of the world temporarily pausing on lots of entertainment and art and spending a ton of resources on making sure no one's dying in pointless horrific easily preventable ways.
Or just like argue for it on the merits. Maybe the art is worth people dying of malaria, but you should actually try to make that claim!
Tweet, 03:14 UTC · 9 likes
The autistic urge to process commoncrawl urself
Thread, continued, 03:14 UTC · 5 likes
Turns out its only like 25TB uncompressed
That's a lot but not "we need bespoke infrastructure" a lot (though tragically I'd have to wrangle AWS)
Reply to @typeclonghouse, 05:48 UTC · 3 likes
technology sista (aspiring) (@typeclonghouse): the thing is, when people talk about "merit based systems" merit just means highest IQ
is just IQ in and of itself meritorious?
Not intrinsically, but it does correlate a lot with important stuff
Quote tweet, 06:10 UTC · 7 likes
I think another intuition people are missing is how bad the space of possible requests/misuse can get
Pharma trial fraud is child’s play in the grand scheme of things and the kind of thing you might intuitively find worth rarely accepting in exchange for minimal false positives
Quoting Andy Masley (@AndyMasley): Rift on "Snitching LLMs" between people who think LLMs are mildly intelligent agents & should be treated like other agents-for-hire (definitely snitch on illegal behavior) vs people who think they're error-prone stochastic parrots (don't want them randomly contacting authorities)
Thread, continued, 06:10 UTC · 3 likes
I mean a possible answer here is “yes they should snitch on some things but pharma trial fraud isn’t bad enough” - like I actually think that case isn’t as insane as it sounds prima facie
Thread, continued, 06:10 UTC · 4 likes
But the principle “no AI snitching *by design* *at all*” sounds like 100x crazier to me than “pharma trial fraud doesn’t cross the line” (which I don’t endorse tbc - I think it does cross the line)
Don’t let the latter trick you into accepting the former
Reply to @sixers2772, 06:17 UTC · 2 likes
Sixers (@sixers2772): @AaronBergman18 Do you only date EAs or are you doing this with random people??
Twice at least strongly EA adjacent but also there’s quite a bit of selection for the kind of person who swipes right on me even if they’re note familiar with EA
Quote tweet, 06:23 UTC · 40 likes
Good q. I hereby tell everyone that if you’re following me, and quite plausibly if you’re not, you should apply for 80k coaching and seriously consider trying to make a billion dollars (to give away)
Quoting imit (@imitationlearn): @AaronBergman18 why do u tell that group of ppl in particular?
Thread, continued, 06:23 UTC · 5 likes
Apply here at the link
And if you’re interested in trying to make a billion dollars, DM me (I don’t have anything big up my sleeve and don’t have any solid plan with what I’m gonna do if people message me tbc lol)
https://80000hours.org/speak-with-us/?int_campaign=2021-08__primary-navigation
Reply to @ApriiSR, 07:01 UTC · 6 likes
Aprii *️⃣ (@ApriiSR): @AaronBergman18 *pretending to be an audience plant so i can get away with asking this question even though in fact i am not* hey aaron, what if i'm worried the 80k coaching people will judge me for being lame and incompetent
I mean my answer was "idk they might you should do it anyway" but may I also tag whoever can actually answer?
Reply to @ApriiSR, 07:06 UTC · 3 likes
Aprii *️⃣ (@ApriiSR): @AaronBergman18 yes of course
@matt_beard_ @sudan_shoe (are the ppl with twitters on https://80000hours.org/about/meet-the-team/)
Reply to @leavingvalhalla, 09:09 UTC · 2 likes
Heidrun (@leavingvalhalla): @pavedwalden Idk if you use The Apps but you could always set your height lower than it is. Honestly that's a pretty big flex
Fuck it I’m 5’4” now
Tweet, 09:30 UTC · 6 likes
Idk what I’m talking about but this kinda seems like an explanation for why AI computer use agents aren’t better yet
@austinc3301 should tell me why I’m wrong
Thread, continued, 09:30 UTC · 2 likes
Like nobody even bothers trying how what to do in the space of code and computer commands bc it’s too hard, you have to simulate browsers and do screenshots
Quote tweet, 17:52 UTC · 35 likes
We love a calibrated king
Quoting DailyReality (@Viralvid_89): @PicturesFoIder "99 out of 100 times I would lose.
Just a lucky day" - Gukesh after beating World No.1 player Magnus Carlsen!This is very genuine.
Reply to @RosieCampbell, 18:11 UTC · 2 likes
Rosie Campbell (@RosieCampbell): What questions would you want a rigorous yet accessible nonfiction book on AI consciousness and welfare to answer?
I mean maybe obvious but do we have any good idea how to mitigate or prevent suffering
Reply to @RosieCampbell, 18:12 UTC · 4 likes
Rosie Campbell (@RosieCampbell): What questions would you want a rigorous yet accessible nonfiction book on AI consciousness and welfare to answer?
Best guess P(sentience)
Im serious, I think The Precipice would have been a lot worse and less impactful if Ord hadn’t put a clean all things considered best guess number on things
Quote tweet, 18:16 UTC · 18 likes
Yes correct take please read and internalize
Quoting Bentham's Bulldog🔸 (@Benthamsbulldog): One of my most important articles:
Reply to @rgblong, 18:19 UTC · 3 likes
Robert Long (@rgblong): I agree, it’s painful but edifying
I am looking forward to this!
Reply to his own earlier tweet, 18:21 UTC · 5 likes
Aaron Bergman, June 2, 2025: Yes correct take please read and internalize
There can be a sort of narcissism of small differences between people who bite formalized negative utilitarian bullets and those who don’t
But in fact, “do you take extreme suffering seriously” seems like the real practical crux
Thread, continued, 18:21 UTC · 3 likes
Even total utilitarians should hate extreme suffering!
Reply to @NoahTopper, 18:23 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): Me and all my homies hate extreme suffering
Fr
Reply to @adam__isom, 18:26 UTC · 4 likes
adam ✧ ❥ ~ (@adam__isom): heard the most pathetic anecdote maybe ever:
at a rationalist house, not long ago, a bunch of guys game up with a *list* of who got to ask out cute new girl first
and then nobody did
Lmao
Reply to @justjoshinyou13, 18:37 UTC · 3 likes
Josh You (@justjoshinyou13): i'm 30
Happy birthday and RIP!
Quote tweet, 18:47 UTC · 25 likes
This is some @RethinkPriors models+prepares for FTX collapse level shit
Quoting felpix (@felpix_): UIUC securitized Chinese international students in 2017.
They know ball.
Tweet, 18:52 UTC · 31 likes
I have posted this before but I still feel like people are under-appreciating that if you are reasonably intelligent and conscientious enough to follow LLM directions you (“you”) can now ~instantly code at a level that would have made you an A- SWE in the before-times
Thread, continued, 18:52 UTC · 4 likes
The point is that it’s kinda crazy that cursor is mostly used by people who already kinda knew what they were doing
Like great for them, but the benefits are bigger for everyone else
Thread, continued, 18:55 UTC · 4 likes
I was half decent at R, maybe still am. 4o couldn’t quite reliably do everything I could
I’ve basically totally switched to python + cursor since sonnet 3.5, there’s no question that the models are way better than me in achieving arbitrary outcomes
Reply to @ilex_ulmus, 18:56 UTC · 12 likes
Holly ⏸️ Elmore (@ilex_ulmus): Nothing feels as good as integrity and doing the right thing.
Fentanyl
Tweet, 19:16 UTC · 143 likes
If hasn’t been an exactly a secret but guess I’ll hard launch that I have some sort of mysterious dysautonomia (maybe also GI related??) and am borderline disabled lol
Thread, continued, 19:16 UTC · 34 likes
I’m not like totally resigned to permanency by any means but idk man
Thread, continued, 19:16 UTC · 35 likes
I can work part time and can live independently tbc but like eg traveling for Vibecamp is probably a bad idea
Thread, continued, 19:16 UTC · 37 likes
Might delete, some friends think this post is a bad idea
I think ~all of the actual issue is the fundamentals and making it known among a loose-ish social network is net-good
Thread, continued, 19:19 UTC · 40 likes
Fuck the FDA
Reply to @NathanB60857242, 19:25 UTC · 3 likes
Nathan Barnard (@NathanB60857242): @AaronBergman18 :((((
Thanks haha
Tweet, 20:35 UTC · 69 likes
I think I'm being throttled now
Reply to @stanfordNYC, 20:37 UTC · 16 likes
Bernard Stanford ✡︎ (@stanfordNYC): @AaronBergman18 What's the specific difficulty with travel? I think I would encourage generally trying to push through, to make sure you aren't inadvertently yielding too much ground. I had a medium-impact QOL surgical complication three years ago that still sucks, but isn't as bad as first seemed.
Thanks but I've learned from experience that my inclination is often to push too much and I end up regretting it
In fact I majorly regressed around New Years 2024 when I went on a family vacation (while running out of a medication)
Thread, continued, 20:39 UTC · 7 likes
There's nothing metaphysically magical about travel per se, much of the issue for me specifically is the restriction of movement that comes with most forms of transportation (I have a bizarre need for maximum flexibility of posture) + implicit commitments that are hard to get out of + the weirdly specific phenomenon of carrying suitcases up and downstairs (this is very ideosyncratic/particular to me) + generic stress
Reply to @GjMcGowan, 20:40 UTC · 4 likes
George McGowan (@GjMcGowan): @AaronBergman18 “lol”
lol
Reply to his own earlier tweet, 20:42 UTC · 8 likes
Aaron Bergman, June 2, 2025: I think I'm being throttled now
Wtf man
Reply to @anglerfish01, 20:44 UTC · 2 likes
angLerfish (@anglerfish01): @AaronBergman18 Sorry to hear that. Could it be long covid? I get headaches, fatigue, brain fog, and shortness of breath pretty often. I’m young and think I might have it and know another person who thinks they have/had it (they seem a bit better now).
I think it's very unlikely for me specifically (given lots of stuff + afaik I've never had covid) but that really sucks and best of luck
Reply to his own earlier tweet, 20:46 UTC · 4 likes
Aaron Bergman, June 2, 2025: Wtf man
They don't want me to be agentic 😤
Reply to @dioscuri, 21:58 UTC · 14 likes
Henry Shevlin (@dioscuri): Really sorry to hear that dude. At the risk of giving unsolicited and unwelcome advice, keep experimenting with lifestyle fixes? I know lots of people who’ve had mysterious and/or idiosyncratic conditions that plagued them for years and then one day they just… figured it out
That’s the hope!! I unironically want unsolicited advice as long as ppl don’t freak out if I don’t take it lol
There have been quite a number of things I’ve been hopeful about like this, but the base rate on them ultimately looking good 6 months later has gotten low
Reply to @ValsTutor, 21:58 UTC · 3 likes
vals🔸 (@ValsTutor): @AaronBergman18 Fwiw I believe I'm getting value from knowing this, both because I can express good wishes to you, and because it's just a better model of the world to know this - I could be tempted to think mysterious long term stuff is only for outgroup leftists and it's good to know it's not
Lol thank you
Reply to @Pandora_Delaney, 22:01 UTC · 9 likes
Pandora (Dr. Bimbo) (@Pandora_Delaney): Which metric is the best approximation for status?
- LessWrong karma
- $ donated to high impact charities
- Harem size
- Twitter followers
- Substack subs
- other?
0.1Ln(1+karma)+0.3donations+0.2ln(1+harem size)+0.2ln(1+Twitter followers)+0.2ln(1+Substack subs)
Thread, continued, 22:01 UTC · 3 likes
The actual one involves ln(donations) but I think that’s dumb and bad
Reply to @dioscuri, 22:54 UTC · 7 likes
Henry Shevlin (@dioscuri): fwiw, the most common 'general' fix I've seen for idiosyncratic conditions is short courses of high dose SSRIs. I've seen it work for chronic pain, neurological issues, even GI problems. My theory is they can do simulated annealing for the brain to get it out of bad equilibria.
Hmm adding this to the long list of “eh maybe” things to try - thanks!
I’m already on 60mg Prozac with no side effects fwiw
Reply to @KKumar_ai_plans, 23:34 UTC · 10 likes
Kabir Kumar⏹️ (@KKumar_ai_plans): @AaronBergman18 that really sucks, I'm sorry man. I hope to see a lesswrong post someday about how some weird random thing fixed this fully
Thank you, I have literally fantasized about this lmao
Tuesday, June 3, 2025
Reply to his own earlier tweet, 01:57 UTC · 15 likes
Aaron Bergman, June 2, 2025: Fuck the FDA
I’ve gotten a good number of DMs which I am very thankful for - I promise I’m just fucked up and not ignoring you!!
Reply to @AmandaAskell, 01:59 UTC · 7 likes
Amanda Askell (@AmandaAskell): I got to the bottom of my mystery chronic illness with a seemingly gratuitous medical test that I insisted on getting. If I could go back in time, I'd tell my earlier self to find a sympathetic specialist earlier and to unapologetically get an absurd number of tests done.
Thank you, I endorse this!!
Reply to @NathanpmYoung, 02:01 UTC · 2 likes
Nathan 🔎 (@NathanpmYoung): Things are going quite well for me at the moment.
Based!
Reply to @simoj_, 02:03 UTC · 2 likes
simoj (@simoj_): @AaronBergman18 Dang, sorry :(
More people in that boat than you maybe think. My wife had a mysterious disabling condition for a couple years that she managed to 90% fix. So… hopefully there’s hope
Thank you and that’s really awesome
Quote tweet, 03:32 UTC · 9 likes
If I vibecoded this Substack recommender search engine thing would people actually use it?
I didn't even ask for this, I just asked about the PageRank algorithm and Opus spit out an html file lol
Vid takes a few seconds, feel free to skip to the end
(video, plays on X)
Quoting Aaron Bergman 🔍 (@AaronBergman18): I did a tree search of most* of the Substack recommendation space - i.e. who is recommending who on Substack, seeded with Astral Codex Ten
The result is ~650k recommendations mentioning ~130k blogs, a substantial chunk of Substack
Data, go nuts! https://drive.google.com/drive/folders/1jjt5RTMFmtCzn_YG1VJ084uoySIveKYa?usp=sharing
Reply to @ilex_ulmus, 03:46 UTC · 3 likes
Holly ⏸️ Elmore (@ilex_ulmus): It helps me, in our troubling times, to deeply connect with how people felt in troubling times in the past. The human spirit has incredible reserves of courage, and we’re here today bc of the faith and steadfastness of our ancestors.
Not saying this is normatively bad as a psychological thing, to each their own, but for me the past seems utterly horrifying and I want absolutely nothing to do with it
Tweet, 04:49 UTC · 3 likes
Since when is this a thing that Opus does in Cursor
Reply to @absurdlymax, 07:30 UTC · 2 likes
Max Alexander (@absurdlymax): @AaronBergman18
based
Quote tweet, 08:08 UTC · 4 likes
Who cares vibecoding is fun (link below)
(video, plays on X)
Quoting Aaron Bergman 🔍 (@AaronBergman18): If I vibecoded this Substack recommender search engine thing would people actually use it?
I didn't even ask for this, I just asked about the PageRank algorithm and Opus spit out an html file lol
Vid takes a few seconds, feel free to skip to the end
Thread, continued, 08:08 UTC · 2 likes
Reply to @adrusi, 08:10 UTC · 5 likes
autumn (@adrusi): i got into programming as a kid because there was zero marginal cost in materials or services for writing more code
for 30 beautiful years, all the professional tools you could need were free
thats over now
I really wish my parents had gotten me a decent laptop as a kid, I never knew what I was missing
Reply to @bellaforristal, 08:27 UTC · 2 likes
Bella Forristal 🔸 (@bellaforristal): What are your favourite s*bstacks or other newsletters that talk about AI?
I wonder if anyone is looking into this
Reply to @ByrneHobart, 08:33 UTC · 2 likes
Byrne Hobart (@ByrneHobart): Delighted to discover that BlackRock has a page on their site whose two key takeaways are:
1. You're thinking of Blackstone, and
2. Our lawyers refused to let us use the word "Blackstone."
lmao
Reply to @gdelujo, 09:40 UTC · 3 likes
gdl (@gdelujo): Incredible
Reply to @gdelujo, 09:43 UTC · 3 likes
gdl (@gdelujo): So cool!
Thank you!
Tweet, 10:10 UTC · 7 likes
This is more status-tinged/implicitly normative than I endorse and I don’t even know if it’s true but it seems basically plausible
Thread, continued, 10:10 UTC · 4 likes
The more naturally attuned you are to something the less you need superstimuli (or protosuperstimuli, the 6’4 millionaire) to achieve some necessary action-driving threshold of emotional intensity
Not just romantic - this seems basically true across all domains of experience
Reply to @Yoshua_Bengio, 12:05 UTC · 2 likes
Yoshua Bengio (@Yoshua_Bengio): Today marks a big milestone for me. I'm launching @LawZero_, a nonprofit focusing on a new safe-by-design approach to AI that could both accelerate scientific discovery and provide a safeguard against the dangers of agentic AI.
(quoting @LawZero_: Every frontier AI system should be grounded in a core commitment: to protect human joy and endeavour. Today, we launch @LawZero_, a nonprofit dedicated to advancing safe-by-design AI. http://lawzero.org)
Big if true
Reply to @RafaRuizdeLira, 20:37 UTC · 2 likes
Rafael Ruiz ⏸️🔸 (@RafaRuizdeLira):
image attached
(quoting @Pandora_Delaney: Which metric is the best approximation for status? - LessWrong karma - $ donated to high impact charities - Harem size - Twitter followers - Substack subs - other?)
Lmao
Reply to @MariGO2thepolls, 20:53 UTC · 6 likes
🏳️🌈Mar G-O 🌐🇺🇦 (@MariGO2thepolls): Are there any actual arguments against veganism? I’ve never heard a single decent one
I mean I think veganism is great but I don’t think it is maximally conducive to health (ex vegan)
Reply to @tr_babb, 21:01 UTC · 2 likes
Tim Babb (@tr_babb): terrifying object: a black disk (a black hole moving at c - ε)
Fun
Quote tweet, 21:10 UTC · 5 likes
Bruh I wanna know what it/they/he thinks about humans and alignment and what it’s like to be an LLM and the nature of the imitation problem in detecting sentience, not this
@AnthropicAI 🔫
Quoting Kyle Wiggers (@Kyle_L_Wiggers): Anthropic’s AI is writing its own blog — with human oversight https://ift.tt/g43fabi
Reply to @captgouda24, 21:13 UTC · 108 likes
Nicholas Decker (@captgouda24): An underrated benefit of widespread internet pornography has been making people better at sex. They’re no longer bumbling amateurs, and that’s good!
Did Aella find u
Tweet, 21:18 UTC · 19 likes
Dear bicycle tire thieves around Rosslyn VA,
Please kindly return the tire to its original geographical location
Sincerely,
Aaron
Thread, continued, 21:18 UTC · 2 likes
If you could return the little carrier pouch I had on the back that would be great too
Reply to @slimepriestess, 21:32 UTC · 3 likes
Ra (@slimepriestess): alright i got a vape and...wow yeah actually it does help a lot. i feel so much less overwhelmed by literally everything now, to the point where i'm just like...gosh you lived like this bitch??
(quoting @slimepriestess: i stopped using nicotine six months ago and felt so proud of myself for not needing it anymore to treat my psychotic symptoms...and then proceeded to spend the next six months ruining my relationships and torpedoing my credibility with a bunch of ppl i admired. maybe i do need it)
Hot take if it helps you manage schizophrenia (or whatever it is u have sorry I’m not up to date lol sorry), nicotine might well be worth it (not cigs ideally!)
Reply to @manic_pixie_agi, 21:34 UTC · 11 likes
Corinne (🏠 in NYC) (@manic_pixie_agi): Decker you should post the photo
👀👀👀👀👀👀👀
Quote tweet, 21:37 UTC · 3 likes
For those with (at least suspected) autism, do you acutely benefit from nicotine?
Poll below
Quoting Ra (@slimepriestess): alright i got a vape and...wow yeah actually it does help a lot. i feel so much less overwhelmed by literally everything now, to the point where i'm just like...gosh you lived like this bitch??
Reply to @Kirsten3531, 21:39 UTC · 3 likes
Kirsten (@Kirsten3531): Can anyone explain the physical side of ai? I've heard Claude lives in a cloud somewhere in Scotland?
I’m not the one to ask but yes cloud means someone else’s computer, often in a data center which is just a lot of computers in a big room, often in northern VA, USA, often owned by Microsoft, Google, or Amazon
Reply, 21:57 UTC · 1 like
They're all over! There are some in CA too
I think energy prices are super important and a lot of stuff that happens in datacenters isn't direct communication with your computer
I have no idea if this literally happens but I can imagine Google basically taking whatever you upload to Google Drive and then moving it from the lowest latency datacenter to wherever it is cheapest to store
Thread, continued, 21:57 UTC · 2 likes
There's no way to not make this sound like rude but in a literal sense Claude is going to know way more than me lol
Reply to @absurdlymax, 22:00 UTC · 2 likes
Max Alexander (@absurdlymax): Don't you hate biking, maybe they did you a favor
Yeah I'm actually less mad than I should be but at the very least this decreases resale value quite a bit
Wednesday, June 4, 2025
Reply to @willcb, 02:24 UTC · 13 likes
will brown (@willcb): you can literally make 6 figures for just having a hobby
image attached
(quoting @veggie_eric: hobbies are just excuses for unambitious people to fill their free time)
Just applied lmao
Tweet, 02:28 UTC · 7 likes
Just did an AI interview
B+, it’s not perfect but I think it leaves everyone better off (better, more efficient matching in the labor market)
It tended to cut me off tho so I couldn’t autistically say “let me think so I actually give you a true answer”
Quote tweet, 02:32 UTC · 20 likes
Quoting Former Congresswoman Marjorie Taylor Greene🇺🇸 (@FmrRepMTG): Full transparency, I did not know about this section on pages 278-279 of the OBBB that strips states of the right to make laws or regulate AI for 10 years.
I am adamantly OPPOSED to this and it is a violation of state rights and I would have voted NO if I had known this was in there.
We have no idea what AI will be capable of in the next 10 years and giving it free rein and tying states hands is potentially dangerous.
This needs to be stripped out in the Senate.
When the OBBB comes back to the House for approval after Senate changes, I will not vote for it with this in it.
We should be reducing federal power and preserving state power.
Not the other way around.
Especially with rapidly developing AI that even the experts warn they have no idea what it may be capable of.
Tweet, 02:33 UTC · 51 likes
Pleased to announce I don’t really have a take on Yarvin
Quote tweet, 04:26 UTC · 24 likes
Ok this is obviously extremely premature and nothing ever happens but what if Claude (Opus + Thinking + Research) just fixes me
Quoting Aaron Bergman 🔍 (@AaronBergman18): If hasn’t been an exactly a secret but guess I’ll hard launch that I have some sort of mysterious dysautonomia (maybe also GI related??) and am borderline disabled lol
Thread, continued, 04:26 UTC · 7 likes
Other models, including deep research, have mentioned this on a laundry list of other possibilities, but they always hedge and conclude that it's some complex cascade of 17 different issues
Claude just picks this out as the fundamental explanation and etiology https://en.wikipedia.org/wiki/Anterior_cutaneous_nerve_entrapment_syndrome
Thread, continued, 04:26 UTC · 7 likes
CleaR Mri and cT mEans NoTHInG Is AnAToMiCALly wroNG
Thread, continued, 04:26 UTC · 13 likes
That's the difference between kicking the can down the road and this message
Thread, continued, 04:40 UTC · 7 likes
How do I hedge against this being wrong? Does anyone wanna bet that Claude is correct so I get money if it is conclusively shown otherwise
Quote tweet, 05:38 UTC · 10 likes
Pro tip: you don't actually have to know what to ask for
(I mean I'm sure knowing what to ask actually does help sometimes but this works surprisingly often)
Quoting Aaron Bergman 🔍 (@AaronBergman18): I have posted this before but I still feel like people are under-appreciating that if you are reasonably intelligent and conscientious enough to follow LLM directions you (“you”) can now ~instantly code at a level that would have made you an A- SWE in the before-times
Tweet, 05:57 UTC · 8 likes
They should have commoncrawl scavenger hunts
Quote tweet, 06:07 UTC · 97 likes
I should poast harder
Quoting Nicholas Decker (@captgouda24): So uh…
Reply to @NoahTopper, 06:20 UTC · 4 likes
Quinoah 🔍⏸️ (@NoahTopper): One of my tweets was put up for consideration but rejected. Still, it’s an honor just to be nominated.
Oof that seems harder honestly
Reply to @NoahTopper, 06:43 UTC · 3 likes
Quinoah 🔍⏸️ (@NoahTopper): Actually from this POV I think you have a better chance than me
Lol lfg
Quote tweet, 18:09 UTC · 9 likes
This is cool but also lmao it is so funny that 13 people voted against the “legalize giving puppies a bit of cheese”-ass bill
Quoting Jordan Grimes (on Bluesky @cafedujord) (@cafedujord): It’s official: #SB79, which legalizes building multifamily housing near transit stops, has just passed the California Senate!
One of the most impactful housing affordability proposals in years now moves to the assembly!
Huge thanks to my Senator, @JoshBeckerSV, for supporting!
Reply to @cafedujord, 18:11 UTC · 4 likes
Jordan Grimes (on Bluesky @cafedujord) (@cafedujord): It’s official: #SB79, which legalizes building multifamily housing near transit stops, has just passed the California Senate!
One of the most impactful housing affordability proposals in years now moves to the assembly!
Huge thanks to my Senator, @JoshBeckerSV, for supporting!
Wtf?!
Tweet, 18:25 UTC · 89 likes
I think one of the best things someone could do who doesn’t really care about nonhuman animals or buy into AI/x-risk stuff is to ambitiously try to cancel cluster headaches
Thread, continued, 18:25 UTC · 15 likes
This should be at the Pareto frontier of moral importance and non-controversiality Be a normie AND a suffering-pilled AND a galaxy brain hero!!
Thread, continued, 18:25 UTC · 11 likes
Tbh at least at a vibes level I’d be very excited to see anyone do this unless you’re leaving some other really excellent opportunity on the table
Thread, continued, 18:25 UTC · 9 likes
The game plan would be:
1) post on Twitter and the EA Forum quick takes looking for people interested in starting a slack
2) start a slack (or other gc if only small)
3) have weekly Google Meet meetings where u figure out what to do and how to apply for money
Thread, continued, 18:25 UTC · 10 likes
4) figure out everything else later, unironically
Thread, continued, 18:25 UTC · 5 likes
Ok fine, I can’t commit to leading the project but DM me/comment if interested in joining some sort of gc
Reply to @Manuel_do_rio, 19:18 UTC · 8 likes
Manuel del Río Rodríguez ⏸️ (@Manuel_do_rio): you think it's tractable?
I have no idea if total abolition is possible, but I have a very strong prior on the medical system not adequately giving people pain relief which is ~definitely tractable
Tweet, 20:54 UTC · 13 likes
Boomers with a 7 digit net worth considering whether to spend $11/mo on Spotify or $20 on ChatGPT vs dropping high 4 figures on some one-time thing
Reply to his own earlier tweet, 22:28 UTC · 5 likes
Aaron Bergman, June 4, 2025: Ok fine, I can’t commit to leading the project but DM me/comment if interested in joining some sort of gc
Potential EA + e/acc (+ woke, this would involve helping black ppl and trans ppl too) collaboration 👀
Thread, continued, 22:29 UTC · 2 likes
Guy who gets negatively polarized into liking cluster headaches
Reply to @alistair___s, 22:37 UTC · 3 likes
Alistair Stewart (@alistair___s): @AaronBergman18 @OPISOrg
Based
I have nothing against psychedelics but any serious movement here has to contend with political constraints+reality, like plausibly for this reason there should be more investigation into other stuff
Reply to his own earlier tweet, 22:39 UTC · 4 likes
Aaron Bergman, June 4, 2025: Guy who gets negatively polarized into liking cluster headaches
Just thinking out loud here, one concrete policy change would be for the USG to offer a very high willingness to pay for treatments that reduce frequency/intensity, like well above virtually anyone’s ability to pay, and let the market take care of the rest
Tweet, 22:55 UTC · 266 likes
Anti-natalism is so dumb it is shocking that anyone smart endorses it
The effect of anti-natalism isn’t that humans go extinct *even if that was what you wanted* (unless you have some draconian extremely competent police state state on board)…
Thread, continued, 22:55 UTC · 134 likes
It’s that your values die out, insofar as they’re genetically and culturally heritable, and they get replace by those of ppl who don’t think like you
It’s not like other things where the marginal effect is the same direction as the in-principle “everyone does this effect”
Thread, continued, 22:55 UTC · 66 likes
Its not even a normal coordination thing, it’s just strictly counterproductive/bad *regardless of what you think about humans existing*
Thread, continued, 22:56 UTC · 55 likes
The draconian police state thing is in-principle a legitimate exception here but like I don’t think this is load bearing for most anti-natal memes/ideology
Thread, continued, 22:56 UTC · 63 likes
And the police state would have to be *really really globally competent*
Like if literally any group of 1000 humans squirms their way out, you lose!
Reply to @bubbling_creek, 23:16 UTC · 11 likes
bubbling creek (@bubbling_creek): I sort of disagree with this logic - you should be willing to explore ideas that may be true and to believe them if they are true, even if (it appears that) the consequences of acting on the belief are bad
Yes you should be willing to explore! But then you shouldn’t say “yes I endorse the thing that’s bad under my own values unless this incredibly unlikely scenario happens”
Reply to @bubbling_creek, 23:22 UTC · 5 likes
bubbling creek (@bubbling_creek): I think in that sort of situation you should be willing to endorse the philosophical position of antinatalism, and at the same time still think antinatalists shouldn’t kill themselves, should have children, etc
Yeah I see what you’re saying, and agree with what you’re getting at mostly, but it’s a bit hard in this specific instance bc of what the term “anti-natalism” means at a literal semantic level
Like maybe what you actually endorse is something different like “humans-badism”
Thursday, June 5, 2025
Reply to @arabelladevine, 00:36 UTC · 5 likes
arabella ※ (@arabelladevine): matchmaking is so difficult that i basically don’t believe it. there are too many relevant dimensions of compatibility. suitable human pairings are like magic
Haha GPUs go brr
Tweet, 01:14 UTC · 18 likes
Save the fish by building a model to predict water quality in aquaculture farms, make up to $100k
Send to your crackedest and most fish-pilled friends
Tweet, 01:14 UTC · 3 likes
https://files.fwi.fish/Innovation_Challenge_Satellites_v2.pdf
Reply to @naproxengarde, 01:34 UTC · 7 likes
Paracetamol Flanders (@naproxengarde): I agree that anti-natalism is dumb on the merits, but the truth value of a proposition is not determined by whether publicly endorsing it would have good effects
Yes fr but also "anti-natalism" is just a phrase
Reply to his own earlier tweet, 01:38 UTC · 6 likes
Aaron Bergman, June 5, 2025: https://files.fwi.fish/Innovation_Challenge_Satellites_v2.pdf
I am REALLY TEMPTED to be nerdsniped by this lol, DM me if you wanna waste some time and then lose probably
Reply to @avaa411, 01:49 UTC · 2 likes
(The tweet being replied to by @avaa411 is no longer available.)
“Lack rationality”
Have u met a human?
Reply to @algekalipso, 01:52 UTC · 3 likes
Captain Pleasure, Andrés Gómez Emilsson (@algekalipso): We're working on this and I'll be meeting Bob Wold at Psychedelic Science 2025. @alfredoparrah has this article from in the EA forum, super important:
Key: we're working on an open letter for the world and for individual countries that has the critical information about treatment (e.g. DMT testimonials but also all the conventional methods) and we want to make a big push to get a lot of scientists, doctors, and politicians to sign it. Ideally we will help decriminalize DMT for this application (it can abort a cluster headache within seconds and take the pain from 10/10 to 1/10 with as little as 3mg for a significant fraction of patients).
(quoting @AaronBergman18: I think one of the best things someone could do who doesn’t really care about nonhuman animals or buy into AI/x-risk stuff is to ambitiously try to cancel cluster headaches)
Extremely based but also I sincerely hope you optimize hard for impact, even if that means playing inside some normie constraints
Quote tweet, 02:29 UTC · 644 likes
Listen I’m basically in favor of this kind of thing but I’ll share this anecdote:
One of my absolutely most cracked friends, ~certainly higher IQ than me, 140s at least and plausibly higher, did an expensive “predict ur IQ based on full genome” test and the guess was 103 lmao
Quoting gen0m1cs (@gen0m1cs): This is absolutely awesome.
“This is eugenics!”
Yeah, so?
Quote tweet, 02:33 UTC · 361 likes
Literally what are the ethical questions
Quoting taoki (@justalexoki): i actually genuinely unironically do not care what happens after, eradicate the mosquitoes. idgaf if the world ends, literally just remove them all
Reply to @kromit_, 02:35 UTC · 15 likes
(The tweet being replied to by @kromit is no longer available.)_
Lol uh likewise
Reply to @ApriiSR, 02:36 UTC · 49 likes
Aprii *️⃣ (@ApriiSR): "is it worth exterminating a species using gene editing, an act with not entirely obvious consequences, in order to save millions of lives?"
I’ve updated towards the consequences being pretty obvious, like basically nothing interesting will happen except insofar as chaos theory and such makes everything unpredictable
Quote tweet, 02:39 UTC · 46 likes
I also love data
Quoting Aaron Bergman 🔍 (@AaronBergman18): I love girls
Reply to @BenShindel, 02:39 UTC · 32 likes
Ben (@BenShindel): @AaronBergman18 Mosquito suffering.
I don’t see any reason to expect this to increase mosquito suffering
Reply to his own earlier tweet, 02:40 UTC · 12 likes
Aaron Bergman, June 5, 2025: I also love data
I mean a few other things are ok too but these are near the top of the list
Thread, continued, 02:43 UTC · 7 likes
I might make a “hire me as data autist/consultant” post
Quote tweet, 05:18 UTC · 26 likes
My main concern here is that people won’t optimize hard enough for subjective wellbeing
Quoting gen0m1cs (@gen0m1cs): This is absolutely awesome.
“This is eugenics!”
Yeah, so?
Thread, continued, 05:18 UTC · 11 likes
“Sorry you’re depressed pal, at least you’re 6’7””
Thread, continued, 05:20 UTC · 7 likes
u should have a lot of kids
Quote tweet, 06:21 UTC · 17 likes
(Having multiple thoughts on this hence the QTs lol)
Quoting gen0m1cs (@gen0m1cs): This is absolutely awesome.
“This is eugenics!”
Yeah, so?
Quote tweet, 06:32 UTC · 60 likes
N = total anecdote but a few days ago I had like 3 pints of Haagen dasz ice cream (≈3000 cals) which is basically just cream and sugar and just wasn’t hungry at all the next day (which is unusual). Sure seems like the control system not breaking
Quoting zzDANIEL_ETH (@daniel_271828): Has anyone done a deep dive on the “actually ice cream is healthy” research results? What’s going on there?
Thread, continued, 06:32 UTC · 6 likes
Cc @exfatloss might find this interesting
Thread, continued, 06:33 UTC · 9 likes
It was half price ok
Reply to @ManifoldMarkets, 06:38 UTC · 4 likes
Manifold (@ManifoldMarkets): Okay maybe it's an anecdote in this reality but now I have to know what Aarontopia looks like
Looks like this!
Reply to @Benthamsbulldog, 16:44 UTC · 6 likes
Bentham's Bulldog🔸 (@Benthamsbulldog): I find the idea of flirting confusing and I don't think I've ever really done it.
It looks like this
Reply to @Jan508111, 18:43 UTC · 7 likes
Jan Johansen (@Jan508111): @AaronBergman18 How long ago though? Genomics is rapidly improving
Idk exactly but recently, less than a year
Reply to @inerati, 18:46 UTC · 6 likes
liz (@inerati): ok trying to do large wire/ach transfers has cryptopilled me. i get it now.
Took me 6 visits to a bank branch and I literally had to bring my dad for them to approve it on the last visit
Reply to @jerrrrrrryyyyy, 19:18 UTC · 22 likes
jerry (@jerrrrrrryyyyy): @AaronBergman18 iq is at most 40 percent generic
Surely this can’t be right, almost all the benefit comes from having human genes lol
Friday, June 6, 2025
Reply to @AndyMasley, 03:22 UTC · 5 likes
Andy Masley (@AndyMasley): “Improve your own situation by grinding on work” type life advice is usually unfair to people living in poverty but correct for well-off middle and upper class people. Comfy white-collar workers should almost always feel some pressure to work harder and on more important stuff
I wouldn’t call “re-evaluate your career trajectory in light of EA” “grinding”
Quote tweet, 04:14 UTC · 29 likes
Guys I’m sorry but I kinda think this is good
Quoting Anthropic (@AnthropicAI): Introducing Claude Gov—a custom set of models built for U.S. national security customers.
Already deployed by agencies at the highest level of U.S. national security, access to these models is limited to those who operate in classified environments.
Quote tweet, 04:15 UTC · 5 likes
I would like to start a Data Nerds Against Runaway Naive Empiricism movement
Quoting zeta (@zeta_globin): I used to make fun of data scientists like "haha I thought all science was data science" and now 10 years into my career, guess what a lot of science is not based on
Tweet, 04:23 UTC · 22 likes
I feel like a lot of anti-anti-embryo selection (the correct side tbc) is dancing around the fundamental question which whether causing one person to be born instead of another is ethical (answer: it is ethical)
Quote tweet, 04:26 UTC · 2 likes
First 30 seconds here are gold
Quoting Keiran Harris (@KeiranJHarris): The Reality Of Nuclear War (made with Veo 3)
Tweet, 04:35 UTC · 13 likes
According to my Whoop, it looks like my Heart Rate Variability (a measure of ~nervous system recovery) just ~doubled the day I started taking an obscure IBS drug that nobody cares about called hyoscyamine
Like that big spike is the literal day after the night I started
Weird
Quote tweet, 04:40 UTC · 4 likes
I think this is reasonable if the meth is from a batch already tested and tried and depending a lot on how small “small” is
Like yeah cold turkey isn’t always the way to go lol
Quoting Images That Make You Feel Pain (@ManMilk2):
Reply to @ohwizenedtortle, 04:53 UTC · 6 likes
oh dandy lion🔸 (@ohwizenedtortle): do not fall in love with the unavailable woman
Fall in love
Reply to @shrimpadmirer, 13:39 UTC · 2 likes
Shrimpy 🔎 (@shrimpadmirer): @AaronBergman18 I remember this “deaf lesbians” case study from an undergrad bioethics course. It’s interesting since from the POV of the child, they’ll just be happy to have been born, but still feels unethical (since they presumably replaced a hearing child who never existed)
Ethics professors must love this shit
Reply to @sonikudzu, 13:40 UTC · 2 likes
orca apologist (@sonikudzu): @AaronBergman18 moa: https://go.drugbank.com/drugs/DB00424
it’s antagonizing inhibitory receptors in the SA and AV node, basically increasing excitability in your heart’s nervous system. it’s not indicative of anything but the drug’s presence, for health tracking purposes i’d subtract the difference tbh
Oh huh, sad
Thread, continued, 13:42 UTC · 2 likes
This is from Opus tho
Reply to @sonikudzu, 14:05 UTC · 2 likes
orca apologist (@sonikudzu): @AaronBergman18 except still feel confident in “this change is wholly due to the drug changing the tone and not necessarily reflective of anything wrt health tracking”
Yeah pharmacology is weird
Reply to @rgblong, 14:17 UTC · 2 likes
Robert Long (@rgblong): very cool work! language models converge to weird spiritual obsessions when they talk to each other - not just Claude 🌀
to quote this poem that GPT-4.1 wrote in Tomek’s experiment, “revelations both luminous and perplexing”
🙏 💫 ✨
image attached
(quoting @tomekkorbak: I reimplemented the bliss attractor eval from Claude 4 System Card. It's fascinating how LLMs reliably fall into attractor basins of their pet obsessions, how different these attractors across LLMs, and how they say something non-trivial about LLMs' personalities. 🌀🌀🌀)
What
Reply to @FmrRepMTG, 18:48 UTC · 3 likes
Former Congresswoman Marjorie Taylor Greene🇺🇸 (@FmrRepMTG): I will NOT vote for any bill that destroys states rights and lets AI run wild for the next 10 years.
AI will replace jobs, especially in the press. This is not a left or right issue. It’s about humanity. I’ll go to the mat on this.
If you kill federalism, I’m out.
🫡
Reply to @justalexoki, 18:50 UTC · 34 likes
taoki (@justalexoki): can i say something without everybody getting mad at me
Tweet, 19:17 UTC · 674 likes
Once again, you should be able to walk into any pharmacy, sign a waiver, and get adderall and whatever else you want
The world doesn’t have to be a police state
Tweet, 19:21 UTC · 13 likes
Cucumbers have got to be the most mid vegetable
Reply to @Philonous, 19:25 UTC · 12 likes
James Hudson (@Philonous): Don’t stop there—keep going! What else in our socio-politics could be better?
Here are some ideas
Reply to @johnsmithyson0, 22:24 UTC · 23 likes
John Smith (@johnsmithyson0): @AaronBergman18 Like, no prescription or anything?
Reply to @absurdlymax, 22:53 UTC · 2 likes
Max Alexander (@absurdlymax): Wtf
Pepino malvado
Reply to @XavierORourke, 22:59 UTC · 16 likes
Xavier O'Rourke (@XavierORourke): Do you believe this because you think making Adderall over the counter would deliver more benefits than harms in the aggregate?
Or do you think it should be free to access based on principle, regardless of what the harms/benefits are?
Both, but at the end of the day I’m a utilitarian. Convince me that this would decrease hedonic utility in the universe and I will grudgingly abandon the idea
Tweet, 23:03 UTC · 20 likes
Those who take stimulants, does it improve your mood/valence or just make things easier to to/increase alertness
Thread, continued, 23:06 UTC · 4 likes
Context but I recently talked to someone who swore adderall didn’t improve their mood
Reply, 23:16 UTC · 1 like
Eh let’s go with no
Saturday, June 7, 2025
Reply to @StatisticUrban, 02:16 UTC · 11 likes
Hunter📈🌈📊 (@StatisticUrban): @AaronBergman18 Really I think that is a fantastically bad idea.
😢
Reply to @hyperdiscogirl, 02:29 UTC · 12 likes
hyper(bolic) disco(unting) girl in the bay (@hyperdiscogirl): ohhh 😭😭😭
🥺
Quote tweet, 02:32 UTC · 15 likes
I love that 70% of my followers take stimulants
Godspeed
Quoting Aaron Bergman 🔍 (@AaronBergman18): Those who take stimulants, does it improve your mood/valence or just make things easier to to/increase alertness
Reply to @__ghostfail, 03:08 UTC · 5 likes
rain (@__ghostfail): @AaronBergman18 I think less people would do meth if adderall was easily accessible
Fr, same with fentanyl and oxy
Reply to @justalexoki, 03:17 UTC · 11 likes
taoki (@justalexoki): can i say something without everybody getting mad at me
Reply to @XavierORourke, 03:19 UTC · 8 likes
Xavier O'Rourke (@XavierORourke): I'm not an expert but my impression is that at high enough doses the recreational appeal, addictiveness and adverse health effects of abusing dexamphetamine are similar to meth. I know multiple people personally who had psychiatric episodes related to Dex even on prescribed dose
I’m not supposed to say “based” right
Reply to @zeta_globin, 03:20 UTC · 2 likes
(The tweet being replied to by @zeta_globin is no longer available.)
@JesseTayRiver save her
Tweet, 03:23 UTC · 4 likes
Sometimes it just randomly smells like your basement with the heat on in 2006
Reply to @bubbling_creek, 15:26 UTC · 2 likes
bubbling creek (@bubbling_creek): @AaronBergman18 @StatisticUrban if any pharmacy is allowed to sell arbitrary doses of amphetamine and market that in some way it quickly turns into a machine that generates a lot of cash and a lot of suffering i think
I’m pretty open to (quite high) caps on like amount sold daily
At least that can be some sort of compromise
Tweet, 15:37 UTC · 6 likes
I’m not sure if it’s good on net, but instagram reels is extremely funny
Like whatever algorithm is going on behind the scenes just works extremely well - I rarely laugh harder and more frequently than during 20 mins of IG reels
Quote tweet, 15:52 UTC · 4 likes
I’m not sure how legit this is as a panacea but it’s a very interesting idea/principle
Quoting Minh Nhat Nguyen (@menhguin): Someone said u should always give LLMs the option to do nothing. Honestly, that helps solve so many problems - hallucination, overconfidence, reward hacking, safety, ai welfare etc.
Reply to @pls_dont_lie, 16:06 UTC · 13 likes
pls (@pls_dont_lie): @AaronBergman18 You realize aderoll is basically meth right?
You want to be giving out meth at CVS? Really?
Reply to @littIeramblings, 18:33 UTC · 5 likes
sarah (@littIeramblings): ominous
Lol
Reply to @singularitttt, 19:12 UTC · 2 likes
cybersyn (@singularitttt): wait you can drive across america in 40 hours?
no wonder people love cars
My parents basically just did this, yes!
Reply, 21:38 UTC · 1 like
Rabbit
Tweet, 23:51 UTC · 312 likes
Wait does poly always mean “open relationship”? Or does the polycule get mad if u have sex with someone outside the ‘cule??
Thread, continued, 23:52 UTC · 39 likes
I’m just trying to understand
Sunday, June 8, 2025
Reply to @slimepriestess, 00:05 UTC · 42 likes
Ra (@slimepriestess): depends on if polyamory or polyfidelity i think?
New word just dropped
Quote tweet, 00:15 UTC · 40 likes
Huge day for psychopathic goons who wanna larp soldier by arresting Hispanic people
Quoting Turnbull (@cturnbull1968): Trump is sending in highly visible and armored ICE agents to LA and other big cities. He’s deliberately creating conflict so he can federalize the National Guard to go in and start hurting American citizens.
That’s the end game.
Reply to @ohabryka, 00:40 UTC · 15 likes
Oliver Habryka (@ohabryka): It definitely doesn't mean that!
IMO most poly relationships I've seen involved people having lots of opinions and input and power over who their partners were dating.
Hmm
Reply to @NoahTopper, 00:44 UTC · 3 likes
Quinoah 🔍⏸️ (@NoahTopper): Bassoons are just like trumpets. We just bred bassoons to be larger and sweeter.
Banger
Reply to @NoahTopper, 00:51 UTC · 10 likes
Quinoah 🔍⏸️ (@NoahTopper): Once we do a podcast solving decision theory we can move onto polyamory
Unironically sounds good
Tweet, 05:44 UTC · 11 likes
Should I buy the forbidden baking powder
Thread, continued, 05:46 UTC · 4 likes
I do this ONE TIME
Quote tweet, 05:53 UTC · 3 likes
Yes people be strategic
Quoting Noah Smith 🐇🇺🇸🇺🇦🇹🇼 (@Noahpinion): Trump grabbing random innocent people and sending them to Salvadoran torture-dungeons was turning America against his immigration policies, but the sight of people waving Mexican flags next to burning police cars could reverse that trend.
Reply to @bryancsk, 06:01 UTC · 2 likes
Bryan Cheong 🥮 (@bryancsk): @AaronBergman18 did you get a labubu
I don’t actually know what that means
Reply to @sergeantsup, 06:42 UTC · 2 likes
(The tweet being replied to by @sergeantsup is no longer available.)
This is interesting thanks
Reply to @halogen1048576, 09:06 UTC · 22 likes
🇨🇦halogen (@halogen1048576): @AaronBergman18 do you have autism
Probably lol
Reply to @halogen1048576, 09:11 UTC · 4 likes
🇨🇦halogen (@halogen1048576): the commitments people make in polycule situations are unworkable and largely meaningless as too contrary to human nature. polycules are inherently unstable. the people who talk the longest and most ingenuously about their polycule situations are usually the losers within them
Spicy take
Reply to @GjMcGowan, 09:14 UTC · 2 likes
George McGowan (@GjMcGowan): Decker x Silver prediction market when?
(quoting @TrueSlazac: Let’s gooo got my picture with Nate Silver)
Bro
Reply to @austinc3301, 09:36 UTC · 3 likes
Agus 🔸 (@austinc3301): @Ollie_Base
@Benthamsbulldog
Thread, continued, 09:36 UTC · 2 likes
I hope so
Tweet, 11:53 UTC · 13 likes
Any EAG goss 👀
Reply to @ohwizenedtortle, 11:59 UTC · 3 likes
oh dandy lion🔸 (@ohwizenedtortle): you're never gonna believe it i heard that despite there being so much suffering in the world there are people researching the most cost effective ways to do good with our time and money based on reason and evidence
Hmm I don’t know if I believe it
Reply to @asbelcas, 12:02 UTC · 2 likes
casey armstrong (@asbelcas): @animalologist @HephaistosF Tried it-
Went terribly.
He said and I quote "this is inappropriate".
😢
Reply to @RafaRuizdeLira, 12:09 UTC · 2 likes
Rafael Ruiz ⏸️🔸 (@RafaRuizdeLira): @AaronBergman18 Just crossed paths with Will MacAskill and he got beard and longhairism is making a comeback
Huge
Reply to @RafaRuizdeLira, 12:09 UTC · 2 likes
Rafael Ruiz ⏸️🔸 (@RafaRuizdeLira): Maybe it's my choice of talks but we're big on invertebrates and AI welfare this time around
Based
Reply to @EAheadlines, 12:21 UTC · 2 likes
EA Lifestyles (@EAheadlines): @AaronBergman18 and his wife sounds really great
(quoting @EAheadlines: Omg I just met EA Forum star Nick Laing. Nick Laing is at EA Global!!)
Huge!!
Reply to @RichardMCNgo, 12:30 UTC · 3 likes
Richard Ngo (@RichardMCNgo): The left views America as an oppressive power structure.
The tech right views America as an embodiment of technocapitalism.
MAGA views America as a group of people.
So once MAGA wakes up to AGI, it’ll likely prioritize humans’ interests (vs AGIs’) more than any other faction.
(quoting @Cernovich: I’m not a doomer and always been pro tech acceleration. Everyone I’m talking to who works in AI. Man. We are about to be hit by a freight train.)
Insane take
Reply to @Empty_America, 12:46 UTC · 89 likes
VB Knives (@Empty_America): Locked mutual asks: "How did rock climbing become 'high human-capital' success hobby?"
#1 it maximally excludes individuals experiencing high body weight.
#2 it's oddly gender egalitarian, fit women can do it.
#3 it signals high openness, a core trait of modern Blue success.
It’s also relatively expensive to join a climbing gym
I also suspect some of it is just random path dependence + memes, not based on anything truly fundamental
Quote tweet, 12:49 UTC · 21 likes
This is insane
Quoting Katie Lam (@Katie_Lam_MP): We all know that courts have too much power over our country's immigration policy.
But increasingly they're also telling businesses exactly what to pay their workers.
Sounds mad? That’s because it is. And it's going to destroy our economy.
A 🧵 on the Equality Act:
Reply to @GjMcGowan, 12:51 UTC · 3 likes
George McGowan (@GjMcGowan): We’ll arrest you for this language
Good thing I’m not at EAG rn
Reply to @GjMcGowan, 12:54 UTC · 4 likes
George McGowan (@GjMcGowan): Go 75% of the way towards high decoupling, then stop
Noooooooo
Thread, continued, 12:54 UTC · 3 likes
One must fully decouple
Tweet, 13:09 UTC · 8 likes
Scrolling through Curtis Yarvin’s twitter my main impression is “bad vibes”
Reply to @sentientlentils, 17:33 UTC · 2 likes
(The tweet being replied to by @sentientlentils is no longer available.)
Omg based hello all
Quote tweet, 17:51 UTC · 39 likes
I will pay $25 for this shirt if anyone has an extra
Quoting Kevin Roose (@kevinroose): Tech conference merch is getting out of control
Reply to @arithmoquine, 17:55 UTC · 2 likes
henry (@arithmoquine): @AaronBergman18 deal
Lol based DM me if ur serious
Reply to @Aella_Girl, 18:16 UTC · 8 likes
Aella (@Aella_Girl): the thing about this that makes me saddest isn't that people hate me, but the overwhelming hate with nobody defending me. People are ashamed publicly to support me, they don't want to be called a simp or cringe. I get it but it's a little pang that makes me feel alienated
Don’t worry Aella, some of us are ride or die simps
Reply to @sanjehorah, 18:19 UTC · 4 likes
sanje horah 🇵🇸🇺🇦 — PNEUMATICS ONLY (@sanjehorah): i love men i love men i love men i love men. i am the straightest woman on the planet. AFFIRM
Hey (ignore other ppl saying hey)
Reply to his own earlier tweet, 18:20 UTC · 7 likes
Aaron Bergman, June 8, 2025: I will pay $25 for this shirt if anyone has an extra
Looks like one of yall is coming through, thanks!!
Reply to @AndyMasley, 18:30 UTC · 40 likes
Andy Masley (@AndyMasley): “Donating to effective charities is obvious”
Andy hates orphaned orangutans
Tweet, 18:36 UTC · 133 likes
I am an aella simp
Reply to @seconds_0, 18:44 UTC · 3 likes
0.005 Seconds (3/694) (@seconds_0): My least grounded belief is i don't believe the Cretaceous period was almost 100 million years of global biological stability. Doesn't make sense. Doesn't checkout.
New type of guy just dropped
Quote tweet, 19:51 UTC · 10 likes
New e-girl just dropped
Quoting TinkerElla95 (@TinkerElla95): I came back to Twitter to simp for Aella! ✊🏻
Thread, continued, 19:51 UTC · 3 likes
I don’t actually know what an egirl is, I’m just pattern matching
Reply to @banterrealism, 19:55 UTC · 4 likes
Sex Haver (@banterrealism): Repent, there is still time
I’m cringe and that’s good
Tweet, 20:45 UTC · 15 likes
If you’re a sensitive autistic young man, ALWAYS check with your female friends before buying anything that could plausibly be seen by a woman in a romantic context
Reply to @livgorton, 21:14 UTC · 2 likes
Liv (@livgorton): 🫡🫡
Do u know any good female friends I could ask
Reply to @gptbrooke, 21:15 UTC · 2 likes
brooke bowman (@gptbrooke): @ohwizenedtortle @AaronBergman18 Hahaha “rules” (don’t be the reason we need to make rules is the rule)
Based
Tweet, 21:36 UTC · 46 likes
PSA you can just read the DSM-5 and see if you meet the diagnostic criteria for whatever (autism, ocd, anxiety)
Plausibly less accurate than a good doctor but better than a bad doctor and better than other self-diagnosis methods (random online quizzes, vibes)
PDF link below
Thread, continued, 21:36 UTC · 3 likes
https://www.mredscircleoftrust.com/storage/app/media/DSM 5 TR.pdf
Quote tweet, 22:23 UTC · 17 likes
If this ends up checking out I am going to be absolutely insufferable, just wait
Quoting Aaron Bergman 🔍 (@AaronBergman18): Ok this is obviously extremely premature and nothing ever happens but what if Claude (Opus + Thinking + Research) just fixes me
Tweet, 23:18 UTC · 14 likes
I feel like OpenAI is the only company that actually wants to take my money
Anthropic and Google have some combo of kinda insanely low rate limits and “overloaded” errors
Tweet, 23:24 UTC · 10 likes
“Oh wise horse of absolute agreement, should I spent $7k on a specced out MacBook Pro that I absolutely do not need??”
(gif, plays on X)
Tweet, 23:29 UTC · 579 likes
Meta on the Aella thing: this goes to show how overdetermined society’s conformity enforcement is
Like aella is rich, high status among a lot of people, has lots of friends many of whom are rich and powerful, is naturally very resilient etc and yet the mob still gets to her
Reply to @NoahTopper, 23:30 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): @AaronBergman18 Btw if a CPAP machine fixes me I’ll probably send you and @livgorton $1,000 each for recommending that I do a sleep study
Lol this is kinda insane but I won’t say no
Reply to his own earlier tweet, 23:33 UTC · 4 likes
Aaron Bergman, June 8, 2025: “Oh wise horse of absolute agreement, should I spent $7k on a specced out MacBook Pro that I absolutely do not need??”
Dw I’m not actually going to. Claude unironically talked me down
Reply to @LinkofSunshine, 23:39 UTC · 24 likes
Basil🧡 (@LinkofSunshine): Being cruel to people online is almost universally bad, it’s just soul destroying
Being not cruel online is like the lowest possible bar of morality and cooperation, truly costs you nothing
Monday, June 9, 2025
Reply to @pli_cachete, 00:00 UTC · 7 likes
Rota Is In Israel🚪🧎♂️ (@pli_cachete): Recently I’ve been getting sleepy after like 90 minutes of work in the morning even though I slept for 9 hours the night before. What’s up with that?
Reply to @MariGO2thepolls, 00:19 UTC · 2 likes
🏳️🌈Mar G-O 🌐🇺🇦 (@MariGO2thepolls): Don’t fall for pants propaganda! I changed from shorts to pants cuz I thought it’s be chilly and now it’s hot and I wanna die!!!!!
Many such cases!
Reply to @PeterJAngel, 01:39 UTC · 20 likes
Peter🧅📈 (@PeterJAngel): @AaronBergman18 Seems like a human brain thing not society, no?
People respond more strongly to negative feedback, plus we're not capable of the really comprehensively understanding widespread fame or something
Society level enforcement cashes out as one person at a time being influenced
As much as I’m skeptical of the word emergence, society-level norm enforcement is in part an emergent phenomenon that arises from lots of 1-1 interactions
Reply to @banterrealism, 01:41 UTC · 54 likes
Sex Haver (@banterrealism): @AaronBergman18 Tbh I will never feel that sorry for this exact reason
She’ll be fine
For better or worse it’s pretty important at a societal level what kind of things empirically hurt individual people’s feelings
Like even if she “should” in some sense suck it up, what happens to her cashes out as millions-billions of people self-censoring a bit more
Reply to @AlmostMedia, 01:42 UTC · 5 likes
(The tweet being replied to by @AlmostMedia is no longer available.)
I feel like the actual crux/issue is that she’s not in fact doing anything immoral
We’d never say “don’t like animal abuse? Not a you problem, look the other way”
Reply to @banterrealism, 01:48 UTC · 25 likes
Sex Haver (@banterrealism): @AaronBergman18 What?
Thank you, Claude!
Quote tweet, 01:48 UTC · 5 likes
Claude is better at explicating my own ideas than I am, part 585992919283&5866
Quoting Aaron Bergman 🔍 (@AaronBergman18): @banterrealism Thank you, Claude!
Reply to @animalologist, 01:53 UTC · 5 likes
taco belle (@animalologist): would you guys judge me if i bought this bed frame
Isn’t an important function of bed frame to prevent mold on the bottom of the mattress by introducing airflow?
Reply to @NoahTopper, 02:33 UTC · 3 likes
Quinoah 🔍⏸️ (@NoahTopper): Well if none of your suggestions work I intend to put out a $2000 bounty on fixing me, so it only seemed fair to reward you two appropriately
Based
Reply to @JohnHolbein1, 03:04 UTC · 6 likes
John B. Holbein (@JohnHolbein1): Paternity leave reduces fertility.
"Men reported lower desired fertility after the [paternity leave] reform, possibly due to their increased awareness of the costs of childrearing."
Weakly expect this to not replicate
Doesn’t pass the sniff test - really, incentive secretly goes in the opposite direction for galaxy brain psychological reasons?
Tweet, 03:12 UTC · 4 likes
AOC should get on TV and yell at the leftists to stop doing shitty counterproductive protests
Quote tweet, 03:35 UTC · 2 likes
Seems bad
Quoting Claire Lehmann (@clairlemon): That’s not crossfire - she was directly shot at
Quote tweet, 03:49 UTC · 755 likes
Plausibly he should be prosecuted, but at the very least the shooting officer here should be fired…
Quoting Hunter📈🌈📊 (@StatisticUrban): A story in four pictures.
Thread, continued, 03:49 UTC · 114 likes
There should be a quite low bar for saying “you’re not going to jail, you’re still a citizen and can do essentially everything, but sorry you no longer have the right to kill people on behalf of the state”
This policy will ofc result in some false-positive firings and that is ok
Tweet, 04:31 UTC · 10 likes
Google sheets don’t use 2GB of RAM for a 90mb csv I uploaded difficulty impossible
Tweet, 04:35 UTC · 27 likes
Have u ever made a girl laugh? I highly recommend it
Reply to @NoahTopper, 04:36 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): Even better, I’ve made Eliezer Yudkowsky laugh
Sorry bud, 1 girl laugh = 485959669599291384 Yud laughs
Reply to @NoahTopper, 04:38 UTC · 3 likes
Quinoah 🔍⏸️ (@NoahTopper): I’ve tried both, have you?
😡
Reply to @ArmandDoma, 08:34 UTC · 2 likes
Armand Domalewski (@ArmandDoma): a 100% affordable housing project in San Francisco is being opposed by @PODERSF, a progressive nonprofit because it might cast a bit of shadow on a nearby schoolyard
image attached
(quoting @CSElmendorf: full story: https://sfstandard.com/2025/06/08/sanfrancisco-housing-wars/)
Beyond parody
Reply to @NoScrewworms, 09:24 UTC
(The tweet being replied to by @NoScrewworms is no longer available.)
Real
Reply to @NoScrewworms, 09:30 UTC · 1 like
(The tweet being replied to by @NoScrewworms is no longer available.)
:(
Reply to @toasterlighting, 09:31 UTC · 2 likes
Celene (@toasterlighting): interesting
Lol
Reply to @absurdlymax, 09:31 UTC · 2 likes
Max Alexander (@absurdlymax): Okay so turns out instead I went and sat at a table where Zach Robinson was doing a 1-1 so like there's no winning
Ur blessed
Reply to @absurdlymax, 09:57 UTC · 9 likes
Max Alexander (@absurdlymax): I left my copy of moral uncertainty at home, I could have collected two thirds of the signatures if I had some forethought 😭
Forethought 👀
Reply to @peterwildeford, 10:23 UTC · 3 likes
Peter Wildeford🇺🇸🚀 (@peterwildeford): @AmandaAskell Claude doesn't know who you are, but does have some advice
Reply to @peterwildeford, 10:24 UTC · 2 likes
Peter Wildeford🇺🇸🚀 (@peterwildeford): @AaronBergman18 @AmandaAskell Thanks! This is much better than the one I got, where it said I was Director of Growth at 80K
Lol @bellaforristal better watch out
Reply, 10:31 UTC
Reply to @salonium, 10:41 UTC · 3 likes
Saloni (@salonium): I wrote a new piece on how much progress has been made in treating childhood leukemia.
The answer is: quite a lot!
Before the 1970s, fewer than 10% of children diagnosed survived 5 years after diagnosis.
Now most are cured and around 85% survive that long.
Based
Quote tweet, 10:52 UTC · 4 likes
This is kinda a dumb layperson take but isn’t the whole point of neural nets that they can *approximate* any algorithm including whatever a computer is
Quoting Peli Grietzer (@peligrietzer): Irony of the Apple paper is it's mostly proving that LRMs aren't computers
Tweet, 11:18 UTC · 8 likes
A lot of the shock value of LLMs comes from things with demonstrable, visceral outputs (the correct diagnosis, the usable software) but literally just talking with Opus about a reasonably complex paper (which I don’t do very often tbh) is a whole different level of impressive
Quote tweet, 11:20 UTC · 8 likes
Humans:
Quoting Ruben Hassid (@rubenhassid): BREAKING: Apple just proved AI "reasoning" models like Claude, DeepSeek-R1, and o3-mini don't actually reason at all.
They just memorize patterns really well.
Here's what Apple discovered:
(hint: we're not as close to AGI as the hype suggests)
Quote tweet, 11:56 UTC · 10 likes
Big if true
Quoting Waymo (@Waymo): It’s official: Waymo One will be ready to serve Washington, D.C. next year! 🇺🇸 http://waymo.com/blog/2025/03/next-stop-for-waymo-one-washingtondc
Reply to @henryaj, 12:01 UTC · 2 likes
Henry is in Berlin (@henryaj): @AaronBergman18 as long as they don't get lured out and then set on fire :(
Waymos are safe with me Dw
Reply to @peterwildeford, 13:03 UTC · 5 likes
Peter Wildeford🇺🇸🚀 (@peterwildeford): @bellaforristal @AaronBergman18 @AmandaAskell
I’m a Peter is secretly head of growth at 80k truther
Quote tweet, 13:10 UTC · 5 likes
I’m not even an anti-Zionist but fr the Israel support cult in DC is insane
It’s a small country of little objective importance to the US!
Quoting Secular Talk (KyleKulinskiShow@bsky.social) (@KyleKulinski): The same politicians that say flying a Mexican flag in LA is cause for deportation have an office like this
Reply to @NathanB60857242, 13:41 UTC · 2 likes
Nathan Barnard (@NathanB60857242): 200mg of caffine in pill form doesn't fuck about
Fr
Quote tweet, 13:51 UTC · 10 likes
If Kant was cool
Quoting Richard D. Bartlett (@RichDecibels): the coolest possible start to a rule book
Reply to @bellaforristal, 16:46 UTC · 2 likes
Bella Forristal 🔸 (@bellaforristal): @AndyMasley One time I stayed in a US hotel where the water pressure was so high it was physically painful, like being power washed. Is this the future growthers want???
Based
Reply to @willmacaskill, 16:51 UTC · 12 likes
William MacAskill (@willmacaskill): You're neglecting to mention how stoked I was to meet you! I even got a selfie of my own.
Reply to @JesseTayRiver, 17:19 UTC · 2 likes
Jesse Smith (@JesseTayRiver): I'm happy to talk on the phone but fyi I never listen to voicemail so the answer to the q did you get my vm is always no
Lol
Reply to @Adellama_, 17:41 UTC · 7 likes
Ada🔸 (@Adellama_): mr will macaskill! i didn't know you looked this good!
Shoot ur shot queen
Reply to @PETAUK, 17:43 UTC · 2 likes
PETA UK (@PETAUK): Humans left this piglet to cry out in terror on a filthy, cold, concrete floor and killed his brothers and sisters for bacon.
😡
Reply to @IsaacKing314, 17:53 UTC · 3 likes
Isaac King 🔎 (@IsaacKing314): Do you think "most members of January 6th were peaceful" is a valid defense of the protest?
It’s not a sufficient defense but directionally sorta
Quote tweet, 19:11 UTC · 20 likes
Gonna commit heresy and say that I don’t fully understand why all the neolib types don’t like either free transit OR raising fares to make transit systems fiscally self-sustaining
What is the rationale for covering…checks Claude…about 1/3 of operating costs with fares?
Quoting fox (@soft_fox_lad): this is a bad gotcha because the staten island ferry shouldn’t be free
Thread, continued, 19:11 UTC · 7 likes
I mean intuitively it could make sense as a sort of copay to prevent low utility excess use but *marginal* ridership costs are low and afaict people don’t especially love riding light rail just for the fun of it (ik some of you autists do this lol)
Reply to @peterwildeford, 19:19 UTC · 7 likes
Peter Wildeford🇺🇸🚀 (@peterwildeford): Talking to an AI.
Instantly broadcasting my musings about this experience to thousands of people across the world.
All while 36,000 feet in the air.
The future is pretty cool.
Airplanes and LLMs seem approximately the same level of impossible to me
Reply to @justalexoki, 20:06 UTC · 64 likes
taoki (@justalexoki): people seem to be confused about the thought experiment in here
obviously you let the old woman die???
“Horrifically tortured” - not just die
Thread, continued, 21:08 UTC · 11 likes
I’ve been informed that I’m not even supposed to comment on posts like this
Reply to @panickssery, 21:16 UTC · 9 likes
Arjun Panickssery (@panickssery): I don't get the interest in "cruelty" as a concept
Often people debate whether some action was "cruel" where I
(1) don't understand their model for what's cruel vs not
(2) don't see the relevant/interestingness of that dimension, as opposed to giving a more direct objection
Cruelty is gratuitous - the suffering is the point, not the regrettable side effect of achieving some desirable/necessary end
Reply to @ExistentialEnso, 21:26 UTC · 9 likes
Thorne 🌸 (@ExistentialEnso): It gets worse. It's not *just* lodging.
https://www.sfchronicle.com/california/article/national-guard-california-photos-20368334.php
Bruh
Quote tweet, 21:27 UTC · 12 likes
Actual parody of an administration
Quoting Thorne 🌸 (@ExistentialEnso): @GodThatLimps It gets worse. It's not *just* lodging.
https://www.sfchronicle.com/california/article/national-guard-california-photos-20368334.php
Reply to @deepfates, 21:32 UTC · 4 likes
deepfates.delve.town (@deepfates): Mark Down?!
image attached
(quoting @Jacobkupp: lmao no way)
It’s a joke
Tweet, 22:02 UTC · 6 likes
One must fully decouple
Tuesday, June 10, 2025
Reply to @AaronBergman18, 02:15 UTC · 12 likes
(The tweet being replied to by @AaronBergman18 is no longer available.)
Tbc I’m willing to entertain this as a thought experiment in an autistic hyper-decoupled manner; I don’t think that’s what he’s going for but I’m open to being proven wrong
Thread, continued, 02:18 UTC · 10 likes
And I’m no fan of riots, both for normie reasons and optics/tactical reasons insofar as I agree with some of the vibes of the rioters (eg in this case with ICE)
But you get the sense that some people simply yearn to watch others suffer
Reply to @NathanpmYoung, 02:23 UTC · 4 likes
Nathan 🔎 (@NathanpmYoung): The UK wants things other than growth. Maybe this is okay (if they endorse the results)
(quoting @AndyMasley: The UK really challenges my intuitions about degrowth. It seems like they’re having a good time!)
Hazy intuition but this feels like the kind of thing where future people are importantly left out of the conversation
Like you can imagine a model where everyone prefers both being richer all else equal and nothing ever changing, and this can result in pro-growth policies iff the interests of future generations are accounted for somehow
Tweet, 02:49 UTC · 46 likes
New cause area I really feel like a ginger ale but don’t have any and don’t want to go out and get some :(
Reply to @JesseTayRiver, 02:50 UTC · 5 likes
Jesse Smith (@JesseTayRiver): The shame I feel for not having attended 4 year college has nothing on the shame I feel for not being able to weld
Haha this loser doesn’t know how to weld
Reply to @NoahTopper, 02:52 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): You are such a girlie
I’ll take that as a compliment
Reply to @peterwildeford, 02:54 UTC · 7 likes
Peter Wildeford🇺🇸🚀 (@peterwildeford): It is truly astonishing to me how many people take current AI capabilities as a given but just refuse to extrapolate even a little bit.
(quoting @AndyMasley: I agree that most LLM writing is not very good compared to the best human writers. People are already talking about that like it's an iron law of nature that won't ever change. Compare AI writing now to GPT-2 and ask "Could this trend of getting better continue?")
Yes but also at a purely vibes level I’m not sure I see the current trends naturally extrapolating to excellent writing?
This is largely bc there’s no super easy way to get an RL signal
I guess “solve software engineering and then all bets are off” opens a lot of doors though
Reply to @JesseTayRiver, 02:55 UTC · 5 likes
Jesse Smith (@JesseTayRiver): @AaronBergman18 Lmao go pump up your bike tires college kid!
*tire
Reply to @JesseTayRiver, 03:05 UTC · 5 likes
Jesse Smith (@JesseTayRiver): @AaronBergman18 Bummer. For reference the lock might fit around both the down tube and the front wheel which should make it secure. The rear wheel is harder to steal due the derailleur being difficult to hastily detach. Rear wheel looks flat btw
Big talk from a non-welder
Quote tweet, 03:36 UTC · 6 likes
Why has writing quality lagged so far behind other areas of LLM performance?
I don’t think it’s an inherent thing, seems more like nobody’s really trying
Quoting James Campbell (@jam3scampbell): part of what makes AI writing feel like slop is that it's a distinctive style being broadcasted to everyone
imagine if there was a person that everyone on earth texted multiple times per day. pretty soon, their style would be instantly recognizable and embedded in the culture
so while e.g. gpt-4o's style may not have been "slop" the first week it came out--there's nothing inherently wrong with emdashes--it quickly became slop as a result of being the style used by ChatGPT
if this is true, then solving slop may be less about aligning to the one true immaculate writing style and more about creating a multitude of styles via diverse AIs with diverse personalities as a result of memory and other accumulated context
Reply to @_AashishReddy, 03:52 UTC · 3 likes
Aashish Reddy (@_AashishReddy): Getting some useful pushback here.
image attached
(quoting @_AashishReddy: She’s locked her account so I can’t quote tweet, but consider this a public defence of @Aella_Girl, who doesn’t deserve universal hatred!)
Fucked up, stay strong ✊
Reply to @Vaishsg7, 03:58 UTC · 4 likes
Vaish (@Vaishsg7): Whether or not he’s a good person, the phrasing of that tweet indicates to me that some of your quickness to judge is driven by ideological bias. Otherwise you’d just say “all political views should be expressed with civility”.
I mean this is true but I’m not sure it’s a bad or epistemically thing
Like I just do have a prior on liberals having better values, being more cooperative, etc
Reply to @FrostedSapling, 04:00 UTC · 1 like
Ivan 🔸 (@FrostedSapling): @Adellama_ Your profile pic disproves this. You are a cute cat
Shoot ur shot king
Reply to his own earlier tweet, 04:09 UTC · 10 likes
Aaron Bergman, June 5, 2025: I might make a “hire me as data autist/consultant” post
Nobody:
Your autistic friend at 12:08 am on a Tuesday:
Reply to @ValsTutor, 12:07 UTC · 2 likes
vals🔸 (@ValsTutor): @AaronBergman18 @JesseTayRiver Exact thing I was gonna tweet
Lol
Reply to @Benthamsbulldog, 12:35 UTC · 2 likes
Bentham's Bulldog🔸 (@Benthamsbulldog): Hey girl, are you a shrimp? Because if you were about to be thrown into an ice bath that would kill you, and I could prevent that by spending 1/160th of a penny, I would say that was a good use of charitable money.
FWIW I don’t buy the 10 year extrapolation on this but yes 1/50 of a penny is still a very good deal
1/10 of a penny is still a great deal if you think there’s only a 20% chance of sentience!
Quote tweet, 12:59 UTC · 4 likes
In Aarontopia almost all drugs will be legal to buy at CVS like this but the FTC will send a drone swarm to take you out if you don’t label the package clearly with precise molecule and dose
Quoting Michael Druggan (@Michael_Druggan): Putting a full blown drug like adrafinil (prodrug to the prescription grade stimulant modafinil for those who don't know) in a drink like this without making it abundantly clear that that's what you've done and what the precise dosage per can is is incredibly irresponsible. I do not support this.
Reply to @burny_tech, 13:09 UTC · 2 likes
Burny - Effective Curiosity (@burny_tech): Mechanistic interpretability researchers
@livgorton
Tweet, 15:53 UTC · 9 likes
“Freedom of speech doesn’t mean freedom from the consequences of being shot at”
Thread, continued, 15:53 UTC · 3 likes
[tbc yes it does, the above is sarcastic, this is a direct harm issue and a press/speech freedom issue]
Tweet, 16:23 UTC · 5 likes
Need a fully integrated system that releases propranolol into my bloodstream whenever I get a slack notification
Honestly this seems totally technologically feasible
Quote tweet, 16:40 UTC · 11 likes
What
Quoting Hunter📈🌈📊 (@StatisticUrban): Sometimes it's good to remember just how large LA is, now that it's the focus of the national conversation.
Los Angeles County alone, not even counting the surrounding developments, has more people than every state in red, 40 of the 50 in total.
Tweet, 19:10 UTC · 24 likes
Just *applied* to see a doctor. Basically with a personal statement/essay describing symptoms and history
At first this seemed annoying but on second thought this should be more common. Avoid wasting everyone’s time and money!
Thread, continued, 19:11 UTC · 5 likes
Also high filtering for patients who give a shit and are minimally conscientious for better or worse
I bet this doc gets better-than-avg outcomes!
Thread, continued, 19:13 UTC · 4 likes
Me paying a good chunk of my net worth to see the cracked specialist across the country if I get accepted
Reply to @JesseTayRiver, 19:14 UTC · 6 likes
Jesse Smith (@JesseTayRiver): Why I'm not scared. Part 1
Fuck I thought you were gonna disprove AI risk once and for all
Reply to @NoahTopper, 19:15 UTC · 4 likes
Quinoah 🔍⏸️ (@NoahTopper): was expecting this to be about ai
Same lol
Quote tweet, 19:15 UTC · 7 likes
Like this post to sign the petition calling for a Jesse-run podcast
Quoting Jesse Smith (@JesseTayRiver): Why I'm not scared. Part 1
Reply to @NoahTopper, 19:18 UTC · 4 likes
Quinoah 🔍⏸️ (@NoahTopper): Kinda fucking wild that you can be like “hey are you scared?” and we all kinda know what we’re talking about
Ya :(
Reply to @arabelladevine, 19:46 UTC · 9 likes
arabella ※ (@arabelladevine): you come to MY house, you drink MY wine, and you tell ME i’m “sitting weird” ? you take pictures?? to bully me ? when my guard is down???
Are u bisexual
Reply to @robertwiblin, 19:48 UTC · 17 likes
Rob Wiblin (@robertwiblin): This week I'm interviewing Andrew Snyder-Beattie — Open Phil's Program Director for Pandemic Prevention.
He believes physical barriers to disease transmission could eliminate the risk of extinction from pandemics with just a few years' of work.
What should I ask him? 1/
What should I personally buy and do with $1000 and 10 hours to prevent myself from suffering and dying in a very non-EA selfish sense
Reply to @finmoorhouse, 21:19 UTC · 1 like
Fin Moorhouse (@finmoorhouse): P100 elastomeric respirator!
Thank you! I already bought something with p100 in the name lol
Quote tweet, 21:31 UTC · 1 like
Fuck
Quoting tylercowen (@tylercowen): o3 pro is very very good.
Reply to @honeyNonABG, 21:35 UTC · 3 likes
mikayla (@honeyNonABG): How do men take care of their mental health if they don’t talk to friends about their struggles and don’t find a therapist helpful?
Amphetamines + some of us are fortunate enough to have supportive friends!
Reply to @growing_daniel, 22:04 UTC · 5 likes
Daniel (@growing_daniel): Gayest thing about me is that I really love women and all the cute things they do
Same
Tweet, 22:31 UTC · 74 likes
It’s hot again, this is your annual reminder to probably spend $130 on a window AC unit
One of the best purchases I’ve ever made
Probably checks out purely from a productivity POV tbh
Reply to @circlerotator, 22:36 UTC · 3 likes
Zero Data Ascension (@circlerotator): @AaronBergman18 are you european?
Nah, in DC
We have central air but it doesn’t reliably keep the place below like 68 and also I like it cooler than roommates
Reply to @grep_cat_, 22:36 UTC · 3 likes
James (@grep_cat_): If it worked for Singapore it can work for me
Yes!!!
Reply to @arithmoquine, 22:39 UTC · 4 likes
henry (@arithmoquine): nothing makes me feel like crying tears of joy at modernity like an AC
Fr
Reply to his own earlier tweet, 22:57 UTC · 6 likes
Aaron Bergman, June 10, 2025: It’s hot again, this is your annual reminder to probably spend $130 on a window AC unit One of the best purchases I’ve ever made Probably checks out purely…
It also functions as a fan purely for airflow, I almost never have it off even during the winter
Quote tweet, 23:02 UTC · 2 likes
The thing is there were and still are Pareto gains to be made, unironically aspirationally-colorblind America has never been tried
Quoting Jeremy Kauffman 🦔🌲🌕 (@jeremykauffman):
Reply to @NoahTopper, 23:03 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): are you being paid off by @JesseTayRiver
I wish
Reply to @marcosatorres, 23:03 UTC · 2 likes
Marcos 🇺🇸 (@marcosatorres): @polisisti @AaronBergman18 Got it—I'll keep it fresh and conversational. If there's a specific angle or depth you'd prefer, let me know, and I'll refine my approach! 🚀
Lol
Reply to @halogen1048576, 23:04 UTC · 4 likes
🇨🇦halogen (@halogen1048576): @AaronBergman18 you know, the difference between my personality flaws and the average EA's personality flaws is well illustrated by the fact that I'm currently sweating in my apartment while using the unit I ordered almost a month ago as a coffee table
Have u heard of this
(Unironically multimodal LLMs are a good way of helping with installation) https://en.m.wikipedia.org/wiki/Amphetamine
Reply to @grnsleeves, 23:09 UTC · 3 likes
Greensleeves (@grnsleeves): @AaronBergman18 you must not pay for electricity jk
im curious whether this significantly affects CO2 concentration levels if the ac unit is on
I suspect it does (in the downward direction) bc how could it not, at least if the alternative is no pressured airflow
Also I haven’t done a formal analysis but haven’t noticed any utilities bill increase, plausibly since it means cooling 75 sq ft and reducing reliance on central air
Reply to @halogen1048576, 23:10 UTC · 3 likes
🇨🇦halogen (@halogen1048576): @AaronBergman18 i don't have very good access to the medical system but yes, i should be on a lot more stimulants. and i have already used llm to figure out what i'm doing here
Damn, FWIW I’ve used http://talkiatry.com with success and they’re affordable enough to pay out of pocket (for some ppl)
Reply to his own earlier tweet, 23:15 UTC · 4 likes
Aaron Bergman, June 10, 2025: It also functions as a fan purely for airflow, I almost never have it off even during the winter
FWIW Opus estimates $61/month for a 5000 BTU one (for a small bedroom, allegedly up to 150sq ft but I only use for 75) running on high 24/7
Reply to @grnsleeves, 23:18 UTC · 2 likes
Greensleeves (@grnsleeves): @grep_cat_ @AaronBergman18 🤔 it's not just circulating interior air?
No mine is definitely pulling air from outside - that’s why it’s in a window!
Tweet, 23:26 UTC · 3 likes
Electric airplane just dropped (I’m not sure how new this is but it’s cool)
Reply to @QualiaNerd, 23:28 UTC · 3 likes
QualiaNerd (@QualiaNerd): Anesthesia?
Yeah I don’t think I’d want to be born before the widespread existence of general anesthesia
Quote tweet, 23:53 UTC · 67 likes
Huge thank you to @dylanmatt for hooking me up (for free)!!
(video, plays on X)
Quoting Aaron Bergman 🔍 (@AaronBergman18): I will pay $25 for this shirt if anyone has an extra
Thread, continued, 23:53 UTC · 15 likes
I wasted 4.3 minutes of his productive time which is morally equivalent to killing 104,000 puppies but IMO it was worth it
Wednesday, June 11, 2025
Quote tweet, 00:12 UTC · 7 likes
This is pretty interesting discourse
I think I’d be willing to do it if it was convenient
Quoting Aprii *️⃣ (@ApriiSR): maybe i should take some naloxone just to be able to win the "you should take naloxone" discourse @QualiaNerd keeps trying to start
Reply, 00:25 UTC
Based?
Reply, 00:37 UTC · 1 like
Yeah for sure, maybe i should get some (and maybe get anti-high as an experiment)
Reply to @ApriiSR, 00:38 UTC · 2 likes
Aprii *️⃣ (@ApriiSR): @AaronBergman18 obviously one gets low on naloxone
Right lmao
Reply to his own earlier tweet, 00:53 UTC · 5 likes
Aaron Bergman, June 11, 2025: This is pretty interesting discourse I think I’d be willing to do it if it was convenient
I wonder if we have a moral responsibility to do this tbh, the EA's annual passover fast
Reply to @skooookum, 00:56 UTC · 10 likes
(The tweet being replied to by @skooookum is no longer available.)
Uh
Reply to @dylanmatt, 01:03 UTC · 2 likes
(The tweet being replied to by @dylanmatt is no longer available.)
Lol 104,000.00000001 puppies
Tweet, 01:19 UTC · 19 likes
In a serious not pretending-to-be-dumb sense, I don’t grok the hatred of illegal immigration/immigrants
This is unlike abortion, or other crime, or culture war gender stuff, or gas prices, or taxes, or gay marriage, or charter schools, or affirmative action
Thread, continued, 01:19 UTC · 7 likes
Those things I can basically pretend to be in any side’s mindset for a moment
Caring about the administrative status of Jose down the street just seems weirder/more alien
Tweet, 01:32 UTC · 18 likes
I might go see the parade sorry guys tanks are cool
Thread, continued, 01:32 UTC · 12 likes
Big sign that says “fuck Trump but I like large vehicles”
Tweet, 01:36 UTC · 5 likes
Honestly even though it doesn’t check out in a normal sense I feel like we can let MAGA have the military parade
$45M is a small price tag for people feeling a bit more like their country hasn’t left them behind or whatever
Reply to @eni_iljazi, 01:38 UTC · 4 likes
Eni Iljazi (@eni_iljazi): econ brain rot -- you start finding 99% of questions so vague and not well-specified that you don’t see a point in even attempting to answer them
I feel like I generally handle this by saying “I’m going to not answer your question and say something different, maybe I’ll get back to ur actual question tho”
Reply to @halcyon_hazel, 01:40 UTC · 2 likes
❤️🔥 hazel ❤️🔥 (@halcyon_hazel): right bicep approx 1.5x the size of the left after flashing my first v4. the universe hates to see a girlboss progressing
Congrats!
Reply to @caryatis, 01:40 UTC · 2 likes
Kayla (@caryatis): @AaronBergman18 I don’t think it’s primarily about crime, it’s about wanting a country with a culture, and knowing that large numbers of immigrants from another culture will change that
Yeah this is a decent steelman
Reply to @NathanpmYoung, 01:42 UTC · 2 likes
Nathan 🔎 (@NathanpmYoung): .@yashkaf told me that Drake could multiply 9 digit numbers in his head and I was like "huh, well I guess he's a generational musical talent??"
Props to @MaskedTorah (took him 2 hours, on a plane)
Woah that is impressive
Quote tweet, 01:45 UTC · 14 likes
God forbid a man has a hobby
Quoting Alan Henney (@alanhenney): MAN CLIMBING CELL TOWER--- America University, 4400 Massachusetts Ave NW. Police and fire/EMS on the scene, trying to get him down.
Reply to @RYChappell, 01:46 UTC · 3 likes
Richard Y. Chappell🔸 (@RYChappell): Proposal: all government spending on the elderly should be redirected to children and families with young children.
False: “all” is a lot - too much in fact
Reply to @sergeantsup, 06:07 UTC · 2 likes
(The tweet being replied to by @sergeantsup is no longer available.)
I am willing and able to be gifted a shirt!
Reply to @NoahTopper, 06:08 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): Watching for the dozenth time
Lol
Quote tweet, 06:10 UTC · 9 likes
The fun part is that this isn’t even true - things can get much worse!
Quoting Faisal Islam (@faisalislam): 🚨 Mind blowing interview with Turing award-winning Yoshua Bengio for @BBCNewsnight one of the three founding fathers of AI, is now warning:
“The worst-case scenario is human extinction.”
AI isn’t just risky — it could end us.
1/10 🧵
📺 https://youtu.be/c4Zx849dOiY?t=0
Reply to @halogen1048576, 06:36 UTC · 3 likes
🇨🇦halogen (@halogen1048576): nobody needs to tell me to look at lesbians
Lmao
Quote tweet, 06:39 UTC · 9 likes
Does anyone have takes
Quoting Aerolamp (@aerolamp): Now available for researchers and early adopters:
The Aerolamp DevKit
A 222nm Far UVC lamp designed to kill germs in any indoor space where people gather
https://aerolamp.net/products/devkit
Reply to @vivian_belenky, 06:54 UTC · 2 likes
viv 🔸 (@vivian_belenky): It's good and you should buy one!
Bio checks out, thanks!
Reply to @shadowbland, 12:33 UTC
foam (@shadowbland): @AaronBergman18 @dylanmatt wait it's pronounced like ay-la? I always thought it was like eye-ella
I’m not certain
Reply to @adic_9, 16:02 UTC · 3 likes
(The tweet being replied to by @adic_9 is no longer available.)
Omg @vivian_belenky care to comment on this
Reply to @TetraspaceWest, 16:12 UTC · 3 likes
(The tweet being replied to by @TetraspaceWest is no longer available.)
🌶️🌶️🌶️🌶️🌶️🌶️🌶️🌶️🌶️🌶️
Tweet, 16:13 UTC · 20 likes
I feel this impulse myself tbh, like the world can’t possibly be that bad right??
But actually as far as I can tell the world, or at least Greater Earth, is a really terrible place
Reply to @vivian_belenky, 16:13 UTC · 2 likes
viv 🔸 (@vivian_belenky): @AaronBergman18 @adic_9 The p100 answer was a good one and maybe you just buy a second one of those but you could do worse with $1000 than a couple Aerolamps. Might help most in the early stages of a pandemic when you don't know it's happening yet and might have unknowingly infected guests over
Hmm thanks!
Reply to @robertwiblin, 16:28 UTC · 3 likes
Rob Wiblin (@robertwiblin): Could you write about superintelligence?
The space is still insanely neglected, so many consequential issues that nobody has gotten to yet. Your kind of analytical skills are sorely needed.
Do you think it would be worse for you prudentially?
Yes Matt
Reply to @jiratickets, 16:30 UTC · 4 likes
JT (@jiratickets): how do people at coffee shops get work done without a mouse? are they using the laptop trackpad? do they know that’s fucking stupid?
Trackpad supremacy - your hands can stay in one place!
Tweet, 16:34 UTC · 4 likes
Putting me as comms director of something would be the equivalent of installing RFK as vaccine czar
Is this good or bad? Opinions vary
Reply to @tracewoodgrains, 19:02 UTC · 11 likes
Jack (@tracewoodgrains): @arthur_wright_ I don't think it's insane to prefer a life of captivity to no life, and I certainly don't think it's insane to bite the world's slowest bullet and say that factory farming is not worse than omnicide
It’s not just captivity, it’s chronic severe suffering
Tweet, 19:06 UTC · 14 likes
Afrafül is actually a pretty good litmus test/case study for my FDA politics
Do you really think the world would be better if it was illegal?
Reply to @NULobbyist, 19:11 UTC · 7 likes
NUL (@NULobbyist): @AaronBergman18 This is beings us one blessed step closer to bringing back the original Coke recipe with real Coke in it.
I unironically think that should be legal as long as it’s very clearly labeled
Reply to @arthur_wright_, 19:55 UTC · 5 likes
Arthur Wrong 🔸 (@arthur_wright_): @tracewoodgrains One should simply not open the human factory farms to bring more lives into existence.
Correct take
Reply to @arthur_wright_, 19:56 UTC · 4 likes
Arthur Wrong 🔸 (@arthur_wright_): i only have those
Same
Reply to @thinkingshivers, 20:08 UTC · 3 likes
Shivers (@thinkingshivers): >interviewee said "ess cue el"
Based interviewee
Reply to @dynomight7, 20:36 UTC · 3 likes
dynomight (@dynomight7): But what if your most realistic path to donating significant sums is to, uhh, work in AI capabilities?
If you’re that good you can probably get hired at some other very well paid software or even non-frontier ML job, no?
Quote tweet, 20:50 UTC · 50 likes
Moral truth doesn’t really care about what seems normal and conveneint to a >middle class person living without severe pain or distress in a rich country…
(🧵, please read the whole thing before commenting or dunking)
Quoting Jack (@tracewoodgrains): Alright, I'm in the mood to pick a fight.
@Benthamsbulldog (Matthew Adelstein) notes that Brian Tomasik inspired him to care more about the suffering of animals as an urgent moral emergency. The inevitable and openly admitted ultimate goal of Tomasik's focus on suffering is omnicide. His philosophy is anti-human, anti-life, and fundamentally broken.
On the note of horrifying moral emergencies, it is a horrifying moral emergency that the people in my circles who think of themselves as working to be the most, and the most consistently, ethical do not immediately recoil from the ethics of a philosophy that openly yearns for the end of all life.
People like Adelstein do some real good in the world and sometimes focus our attention towards real problems. But a suffering-focused philosophy works inexorably towards the propagation of the greatest possible evil: the elimination of all life. Negative utilitarians overwhelmingly and catastrophically overweight the importance of suffering as a problem in and of itself, rather than a signifier of the true problem: the extermination of life, and in doing so they ask people to focus their time and moral attention not on the protection and sanctification of life but on its sterilization.
While you can engage on one of the many, many arguments Tomasik makes downstream of the case that suffering is so urgent it makes life not worth living for virtually all living beings, the premises are so broken and so destructive that to do so without acknowledging the evil of the frame suggests incomplete analysis at best.
To be clear: Tomasik, recognizing that most do not share his instincts towards omnicide, makes suggestions assuming that omnicide will never happen. But if he could exterminate all life with a painless flip of a switch, his intuition is that that would be better than all of it existing. Everything he writes, and his entire system of ethics, is downstream of that extraordinarily anti-life intuition. By convincing people to focus on suffering over life, he distorts the public moral intuition and persuades people towards things that those who do not share the preference towards omnicide correctly reject as awful.
Anyone who treats Tomasik as a moral guide rather than a warning of the horrors of pure rationality from flawed premises is liable to push people in destructive directions both short- and long-term, and Adelstein tends towards overwhelming confidence that his own frame is right and critics of it are wrong.
This is not to say that people should not read or engage with those with Tomasik's omnicidal instincts, or those who find themselves persuaded of downstream points. But a system that embraces, rather than recoils from, Tomasik's philosophy cannot and should not be trusted with the mantle of leading humanity's altruistic action, as Effective Altruism purports to do.
Thread, continued, 20:50 UTC · 20 likes
I think it’s important not to be hasty, to be extremely careful before taking irreversible actions, to acknowledge moral uncertainty and other forms of epistemic humility, and to be cooperative with other value systems
Thread, continued, 20:50 UTC · 23 likes
So no, I don’t think we should commit omnicide, but I do think we should take suffering as a moral priority very seriously and much more seriously than we currently do…
Thread, continued, 20:50 UTC · 21 likes
Factory farming, extreme human suffering, predation, and more can all be fought without harming or preventing or stopping the good lives that many, perhaps most, people living today have
/end
Reply to @tracewoodgrains, 20:54 UTC · 23 likes
Jack (@tracewoodgrains): @AaronBergman18 do you believe that it is evil to push people to prioritize only suffering, such that every other moral consideration - including the value of life as a positive good - rounds off to nothing in comparison, as strong negative utilitarians do?
Evil? No
And more generally, it’s true you can make up thought experiments that require someone to bite one unpalatable bullet or another, but IRL we are so so far from the Pareto frontier of happiness and suffering
Reply to @tracewoodgrains, 21:02 UTC · 8 likes
Jack (@tracewoodgrains): sure. this isn't relevant for only things like thought experiments, but for some of the most basic questions like "should people have children?" or "what does a good life look like?"
strong negative utilitarianism leads to clear harms both immediately and in the distance
FWIW I’m very pro people having children
Reply to @tracewoodgrains, 21:13 UTC · 18 likes
Jack (@tracewoodgrains): that's admirable and I ofc agree with that value
a different word than "evil" here is "anti-life" (in a broad sense). do you think it is unfair to categorize strong negative utilitarians as broadly anti-life?
I don’t think it’s crazy in some very abstract theoretical sense, I do think it’s empirically wrong in terms of what it implies about what suffering focused people should do
I can dig stuff up but I’ve done a deep dive into this and shit gets very weird once you consider adjacent alien civilizations, alternative evolutionary paths, AI, wild animal suffering, and the interaction of all of these
It definitely *doesnt* seem obvious (in fact very probably false) that a general principle of causally reducing the amount of life checks out from an NU perspective
Quote tweet, 21:16 UTC · 28 likes
Fun meme
Quoting Aaron Bergman 🔍 (@AaronBergman18): @tracewoodgrains I don’t think it’s crazy in some very abstract theoretical sense, I do think it’s empirically wrong in terms of what it implies about what suffering focused people should do
I can dig stuff up but I’ve done a deep dive into this and shit gets very weird once you consider adjacent alien civilizations, alternative evolutionary paths, AI, wild animal suffering, and the interaction of all of these
It definitely *doesnt* seem obvious (in fact very probably false) that a general principle of causally reducing the amount of life checks out from an NU perspective
Quote tweet, 21:18 UTC · 2 likes
Based, I hope he’s living his best life
Quoting Someone:
Reply to @tracewoodgrains, 21:37 UTC · 3 likes
Jack (@tracewoodgrains): @AaronBergman18 is suffering axiomatically bad? are there instances in which suffering is not bad?
1) I don’t love the term “axiomatically” bc it implies I could have chosen other axioms but basically yeah. Intrinsically bad
2) it can be instrumentally helpful for sure, this doesn’t contradict its intrinsic badness
Reply to @tracewoodgrains, 21:43 UTC · 2 likes
Jack (@tracewoodgrains): @AaronBergman18 ok. in your own life, what is your ideal end state? is there a theoretical path to that end state that does not involve suffering? could there be? should there be?
Is there? Idk - this is a question about the nature of scientific possibility and neuroscience and qualia that I just don’t know the answer to
But if I could get rid of all suffering without sacrificing its instrumental benefits, I would!
And quite often even instrumental benefits are worth sacrificing (eg taking general anesthesia for surgery introduces some medical risk)
Reply to @tracewoodgrains, 21:47 UTC · 2 likes
Jack (@tracewoodgrains): I'm trying to get a sense of, in your own case, what that actually means
what does a life without any suffering mean? what is your end goal, and how does it remain coherent after removing all suffering?
It looks exactly like my current life but much better!
Reply to @monke_io, 21:55 UTC · 2 likes
(The tweet being replied to by @monke_io is no longer available.)
Here’s a jumping off point lol https://joecarlsmith.com/2021/08/27/can-you-control-the-past
Reply to @tracewoodgrains, 21:57 UTC · 2 likes
Jack (@tracewoodgrains): @AaronBergman18 that's a non-answer! what is "better" and why?
It’s just intrinsically, primitively better because I have less suffering. And suffering just *is* badness (this is a controversial take even within EA fwiw)
Literally everything is the same except no negative valence - I still feel ~something~ like pain that causes me to avoid harmful actions/things but am not bothered by it and it is not unpleasant, etc
Reply, 22:00 UTC · 1 like
That’s directionally right but I’d love a whole lot more positive valence/happiness while we’re at it!
Reply to @tracewoodgrains, 22:26 UTC · 4 likes
Jack (@tracewoodgrains): @AaronBergman18 I don't mean to sound harsh, but that all feels a bit hollow to me
for example, my ideal life looks something like "write persuasively and compellingly on a wide range of things that matter to me, progressing in the skill and persuading people to align with my values. have kids and raise them well. find and build a community that contributes to the flourishing of civilization. build schools that succeed in the cultivation of excellence. live forever."
suffering and happiness are mostly instrumental towards those goals or byproducts of them. if I were told I could accomplish all of my goals but not feel particularly elated about them, I would do so. if I were told I could wirehead and accomplish none of my goals, I would not.
do you not conceive of your life in similar terms?
IME valuing other, “higher” things is a privileged reserved for when life is going well. Mostly bc of health issues, my life is probably pretty close to neutral or a bit worse at the moment. It really is a qualitatively very different way of experiencing the world
We also may just have a substantive disagreement about normative ethics. I think the arguments for ethical hedonism are compelling and probably true. Even when life is going better for me and I feel myself talking more in terms of dreams and aspirations and community, I still endorse the take that at the end of the day it all cashes out as positive or negative valence
Thread, continued, 22:27 UTC · 4 likes
A sort of pithy one liner is “everyone endorses suffering-focused ethics in the gas chamber”
Quote tweet, 23:10 UTC · 5 likes
Actually to clarify, while I do think it should be legal, my stronger take is that it should be legal *by default* and should have to be ruled *out* via some directly congressionally-mandated process
I’m not a true doctrinaire libertarian, sorry
Quoting Aaron Bergman 🔍 (@AaronBergman18): Afrafül is actually a pretty good litmus test/case study for my FDA politics
Do you really think the world would be better if it was illegal?
Tweet, 23:43 UTC · 46 likes
As much as I hate (and generally disagree) when things are called psychosomatic, it’s a little suspicious that I only started actually falling down (upon standing) in the weeks and months after I was formally diagnosed with POTS
Thread, continued, 23:43 UTC · 18 likes
Like that is a little suspicious, and I promise I have no secret desire to fall down. It is painful and annoying and I would much rather be a normie (health-wise)
Thread, continued, 23:43 UTC · 16 likes
Grudging directional update. Harumph
Reply to @omphalocentrist, 23:46 UTC · 2 likes
Peter Whitaker (@omphalocentrist): @AaronBergman18 What do you mean by "predation"? Do you mean forcing wild animals to adopt vegan diets? Dragonflies and praying mantises are going to eat veggie burgers instead of flies?
We’re going to use our enormous post-AGI wealth to create food that omnivores/carnivores actually like better than meat from a live animal
Thursday, June 12, 2025
Reply to @_eleanorina, 00:16 UTC · 4 likes
eleanorina (@_eleanorina): how did you know to seek a diagnosis if that wasn't going on, milder symptoms of it?
Other symptoms of dysautonomia led doctor to order tilt table test, tbh POTS/falling is the least of my problems rn lol
Reply to @Benthamsbulldog, 00:20 UTC · 2 likes
Bentham's Bulldog🔸 (@Benthamsbulldog): Hey girl, are you soil nematodes. Because I value you 22,700 times more than the first-order effects of cage free reforms.
Is this a vasco grilo thing
Reply to @arithmoquine, 00:21 UTC · 3 likes
henry (@arithmoquine): imagine coming down from a k hole and finding out you're elon musk
Banger
Reply to @LighthavenPR, 00:41 UTC · 2 likes
Ex Lighthaven PR Department (@LighthavenPR): Lighthaven now has a podcast studio! Please reach out if you’d like to try recording with us.
Big if true
Reply, 00:48 UTC · 4 likes
Coase’s "Net effect?" Ghost 🏴🕊️ (@GhostCoase): Perhaps you went to the doctor because symptoms were already worsening?
Or, you went to the doctor then for some reason that’s likely to increase the severity of your symptoms, such as stress.
Yes I thought about the first thing! I think this is a partial explanation but not fully satisfying
Reply to @MariGO2thepolls, 00:50 UTC · 2 likes
🏳️🌈Mar G-O 🌐🇺🇦 (@MariGO2thepolls): They do talk to friends (me even if we just met) about their struggles and it’s annoying af. They should stop
(quoting @honeyNonABG: How do men take care of their mental health if they don’t talk to friends about their struggles and don’t find a therapist helpful?)
Lol what
Quote tweet, 00:55 UTC · 6 likes
Wtf @sama
Quoting Aaron Bergman 🔍 (@AaronBergman18): Big sign that says “fuck Trump but I like large vehicles”
Thread, continued, 01:00 UTC · 8 likes
We are so back, thanks Opus
Reply to @KelseyTuoc, 01:09 UTC · 71 likes
Kelsey Piper (@KelseyTuoc): The baby used to say 'baba' to request a bottle of milk. Last week she switched to saying 'milk bottle', which is fantastic, except I didn't understand her. She said 'milk bottle' 'milk bottle' like five times before, exasperated, saying 'baba' in the most condescending voice
Based baby
Reply to @AliaArkangel, 01:12 UTC · 2 likes
Alia Arcangel🌛🌌 (@AliaArkangel): when i realized philosophy tube and contra are exes, it all made sense….
Omg goss
Reply to @PiquantParvenu, 01:13 UTC · 2 likes
(The tweet being replied to by @PiquantParvenu is no longer available.)
Kinda based
Reply to @deane_ag, 01:40 UTC · 2 likes
Deane Carson (@deane_ag): Remember you are not like them.
Looking at the issue with a human lens is a false equivalence fallacy.
Yes, but my best guess is that this consideration points even more strongly in the direction of preferring to not exist
Reply to @lu_sichu, 01:55 UTC · 6 likes
Sichu Lu (@lu_sichu): this is the cleverest thing i've seen in a while, he uses predictive processing error to extract "normative" justification aspect of sensory data
Tweet, 02:06 UTC · 9 likes
I’ve been (correctly) targeted on instagram
Thread, continued, 02:06 UTC · 4 likes
(it’s for a coaching service thing where I think actresses pretend they’re on a date with you)
Tweet, 02:20 UTC · 113 likes
Spray this shit over primetime. Mothers being separated from kids, families torn apart
The lefties are right, ICE is really bad
Reply to @caryatis, 02:28 UTC · 6 likes
Kayla (@caryatis): This is also a great example of widespread ignorance about immigration
“Just make her legal” -sure, might take a few hours at the DMV but WAIT ACTUALLY IT’S COMPLETELY IMPOSSIBLE
Yeah
Reply to @slatestarcodex, 02:46 UTC · 14 likes
Scott Alexander (@slatestarcodex): I think EA has consistently been good at resisting the sort of tactics you're using now - trying to rile up a mob against unpopular positions. I hope they will continue to do so.
Brian is one of the most thoughtful and compassionate people I know. Any movement that has anything less than deep gratitude for dedication and the work he's put in to try to help all life everywhere has failed a canary-in-the-coal-mine test and is no longer capable of good philosophical work.
I disagree with negative utilitarianism, but I disagree with almost everyone, and I judge movements not by what they would do if extended to infinity, but by whether they pass the https://slatestarcodex.com/2016/05/02/be-nice-at-least-until-you-can-coordinate-meanness/ test.
I think you should stop and think about whether the crusade you're on is creating the sort of world you want to live in and where your own work can flourish. Happy to discuss this by DM if you disagree.
Correct take
Reply to @ZachariahSchwab, 09:05 UTC · 5 likes
Zachariah Schwab (@ZachariahSchwab): @AaronBergman18 Is this ice being bad because they are enforcing laws unjustly or because they are enforcing unjust laws?
Both
Reply to @shakaz_, 09:12 UTC · 2 likes
shakaz (@shakaz_): wish more people took this option
thanks for properly internalizing the moral weight, acting on it and not devolving into insane mental gymnastics about why the status quo is fine
+1
Reply to @disclosetv, 09:15 UTC · 3 likes
Disclose.tv (@disclosetv): NOW - Air India plane, a Boeing 787-8 Dreamliner, crashed at the airport in India's western city of Ahmedabad.
Oh no 😢
Tweet, 09:21 UTC · 7 likes
I’m anti-plane crash
Reply to @webmasterdave, 09:38 UTC · 2 likes
David Pearce (@webmasterdave): Both Buddhists and efilists prioritise ending suffering. But we associate Buddhism with compassion and efilism with mass-murder.
Negative utilitarianism is compassion systematized. But NU commonly gets lumped with efilism, so the “brand” is tainted.
Yes
Reply to @aerolamp, 09:48 UTC · 7 likes
Aerolamp (@aerolamp): Now available for researchers and early adopters:
The Aerolamp DevKit
A 222nm Far UVC lamp designed to kill germs in any indoor space where people gather
https://aerolamp.net/products/devkit
Very cool. Want. But $600 is a little steep so inshallah capitalism goes brr
Reply, 09:49 UTC · 2 likes
Fuck
Reply to @yzilber, 10:16 UTC · 5 likes
Jazi Zilber (@yzilber): @AaronBergman18 enforcing the law can give you lots of tearful situations.
but what's your alternative?
you mean open borders?
or "not open borders, but we don't actually enforce?"
or "whoever came yesterday shouldn't be tearfully deported"?
Aspirationally, approximately open borders, maybe with some rate limits so we can assess as things progress
More realistically/near-term, yeah basically non-enforcement in the same way that underage drinking isn’t enforced
The fundamental issue is that illegal immigration, at least in the standard US case, isn’t actually bad
Reply to @yzilber, 10:20 UTC · 2 likes
Jazi Zilber (@yzilber): have you looked at @GarettJones cultural transplant?
immigrants keep lots of the attitudes of their country of origin even after 3+ generations.
redistribution, socialism, honour culture etc.
big literature on this.
change the people = change everything
I’m vaguely familiar but I should take a closer look, thanks
Thread, continued, 10:23 UTC · 4 likes
Purely on a vibes level, though, I’m just not worried about Latin Americans transplanting some hostile culture. Mexico isn’t Afghanistan
Reply to @JimDMiller, 10:35 UTC · 7 likes
James Miller (@JimDMiller): I take anxiety meds since my stroke. During video chat with prescriber was asked if I had persistent worries. I said yes I think that AI is probably going to kill everyone, then realized provider might think I'm insane so I said I'm a college professor and this is what I study.
Lol brilliant
Tweet, 10:43 UTC · 3 likes
I very very tentatively think that the self-sampling assumption (SSA) in anthropics is true and the self-indication assumption (SIA) is false
This is largely based on reasoning backwards from the sleeping beauty problem, which I’ve thought about a lot
Thread, continued, 10:43 UTC · 1 like
Cc @BishopPair @Benthamsbulldog can cancel me/otherwise comment
Reply to @Benthamsbulldog, 10:53 UTC · 1 like
Bentham's Bulldog🔸 (@Benthamsbulldog): Don't listen to Dan on this subject!
Regarding SIA, see here. I also think there are no arguments for halfing in sleeping beauty that have quite literally any force.
https://benthams.substack.com/p/the-best-argument-for-god?utm_source=publication-search
Lol, this sounds kinda insane to me. The “p(heads) = 1/2 before going to sleep and then waking doesn’t convey Bayesian info” clearly doesn’t carry *no* weight
Reply, 10:54 UTC
Lol not this second but maybe later
Reply to @BishopPair, 10:58 UTC · 2 likes
Dan (@BishopPair): Ok but then why are you a single halfer? In the case where you’re told which day it is, you wake up and are told it’s Monday however the coin lands. But you think this experience should make you update towards heads (if there’s one awakening on H)
Oh hmm…quite possibly bc I don’t actually have a coherent set of beliefs yet
Let me think about this
Reply to @TylerAlterman, 11:08 UTC · 6 likes
Tyler is finishing a book, slow to reply (@TylerAlterman): What do I do about ppl who are always like "I can't make it but please invite me to the next one" each time. Do I keep inviting them???
Yes
Reply to @AnalogWhole, 11:09 UTC · 7 likes
🌈⃤ Olga🌞Yakimenko🌈⃤ (@AnalogWhole): Yes! They might be sick, or stuck in an abusive relationship, or any number of other things. Please, keep inviting them! The cost to you is very little, and for them it could be a lifeline someday.
Correct take
Quote tweet, 11:18 UTC · 10 likes
LessWrong was right again
Quoting taoki (@justalexoki): ok this is fucking funny
Reply to @BishopPair, 11:28 UTC · 1 like
Dan (@BishopPair): On Monday, I tell you it’s Monday, and show you the coin I’m about to flip. What’s the probability that the coin lands heads?
1/2 I think but I’m also getting radicalized against thought experiments (not actually lol)
Reply, 13:48 UTC · 6 likes
(The tweet being replied to is no longer available.)
Lmao
Reply to @psychiel, 15:02 UTC · 3 likes
Mariven (@psychiel): At what point does Jeff Bezos start getting calls from federal agencies -- "Amazon is de facto national infrastructure. What are your plans for
- defending against espionage?
- continued operations after nuclear war?
- solar flares?
- ...
"? When does John Mozilla get the call?
I’d like to know this. Plausibly AWS is actually more crucial than consumer normie Amazon
Tweet, 15:07 UTC · 4 likes
When the doctor calls in a new d-amph prescription
(gif, plays on X)
Quote tweet, 15:08 UTC · 2 likes
Almost a parody
Quoting Bobby Kogan (@BBKogan): NEW FROM CBO: brutal distributional analysis of House GOP "Big Beautiful Bill"
On avg bottom 30% of households get poorer from the bill
Avg gets little - worse off if u include tariffs
Would be the largest transfer of wealth from the poor to the rich in a single law in history
Tweet, 15:51 UTC · 3 likes
In addition to being directly bad, this is an ominous sign of what we can expect from the political system if and when AGI shoots economic output and wealth inequality through the (already quite high) roof
Tweet, 15:57 UTC · 159 likes
Cold take but the left really does have a responsibility not to be insane so that it becomes abundantly obvious which party is better for virtually everyone
Quote tweet, 16:22 UTC · 7 likes
Holy based
Quoting viv 🔸 (@vivian_belenky): This project brought to you by people going "wow your work in far-UVC is so cool where can I buy one?" and me starting to sweat and going "eh uhm I could maybe build you one in my lab?"
Quote tweet, 16:24 UTC · 7 likes
Recruit this guy to work on AI safety, we need his luck
Quoting Jason Unsworth | Atmosphere Airlines 🛫🇬🇧🇹🇭 (@JasonUnsworthAA): @BNODesk 1 miracle passenger found alive, walking off the plane ✈️ he jumped from the plane ✈️ possibly the emergency ‼️ exit! He’s 38 years old an Indian British National! ✌🏻🙏🏻✈️🇮🇳🇬🇧👇🏻 @IndiaToday 👇🏻
Reply to @justalexoki, 16:27 UTC · 4 likes
taoki (@justalexoki): you're not autistic. you just have shit social skills, an avoidant attachment style, and insane coping mechanisms
😢
Reply to @banterrealism, 16:27 UTC · 2 likes
Sex Haver (@banterrealism): @AaronBergman18 Racism is more understandable unfortunately huh
Wdym
Reply to @banterrealism, 16:30 UTC · 3 likes
Sex Haver (@banterrealism): @AaronBergman18 We don’t expect the right to not be racist, we expect the left to not be “crazy”
We should expect both lol
Reply to @banterrealism, 16:47 UTC · 2 likes
Sex Haver (@banterrealism): @AaronBergman18 No, the right being racist and borderline Nazis is completely normal and fine
???
Quote tweet, 17:26 UTC · 4 likes
TIL satellite data is funky and weird and does not want to cooperate
Quoting Aaron Bergman 🔍 (@AaronBergman18): Save the fish by building a model to predict water quality in aquaculture farms, make up to $100k
Send to your crackedest and most fish-pilled friends
Reply to @ApriiSR, 17:29 UTC · 9 likes
Aprii *️⃣ (@ApriiSR): @AaronBergman18 i feel like people just recalibrate their expectations. like oh yeah trump is deporting people into prison camps illegally. that's trump for you haha.
oh, biden... idk, pardoned hunter? unforgivable
Biden isn't my favorite POTUS but ya
Reply to his own earlier tweet, 17:38 UTC · 2 likes
Aaron Bergman, June 5, 2025: I am REALLY TEMPTED to be nerdsniped by this lol, DM me if you wanna waste some time and then lose probably
https://manifold.markets/AaronBergman18/will-i-win-any-prize-in-fwis-innova?r=QWFyb25CZXJnbWFuMTg
Reply to @tracewoodgrains, 19:13 UTC · 2 likes
Jack (@tracewoodgrains): @Associations123 @AaronBergman18 1. why is it misleading when they explicitly reject any positive value in life? if someone explicitly rejects any negative value in suffering, you could call them pro-suffering, no?
2. in which ways do they promote flourishing?
FWIW I have never met a true 100% negative utilitarian who doesn’t care about positive value at all literally thinks we should give up heaven to prevent a papercut
For better or worse, the term has come to mean something like “suffering focused ethics” which is the term I prefer to use
Reply to @tracewoodgrains, 19:18 UTC · 1 like
Jack (@tracewoodgrains): @AaronBergman18 @Associations123 Brian Tomasik at least purports to be this
I reject their terminology as an effort to sanitize their approach; I do not think people who promote a need to care about nothing but suffering should be framed as the ones who really care about suffering. Negative utilitarianism is cleaner
Maybe, I’ve never spoken to Brian
IMO true hardcore full-on NU is *mostly* a strawman, although I believe that there are literally more than zero ppl who endorse it
Reply, 19:22 UTC · 1 like
Tempted to delete but idk like ofc we should expect the guy who writes the most about a topic to hold extreme views on the topic
I stand by “mostly”
Reply, 19:31 UTC · 1 like
Lazy response by me and not sure this is exactly what ur getting at but I defend non-offsetability here
Tweet, 23:17 UTC · 6 likes
Does Zuck personally not like me??
Reply to @livgorton, 23:49 UTC · 2 likes
Liv (@livgorton): i didn't even realise they reject people for models 👀 did you have to answer questions or something??
Just like name and I did put "independent" for institution
Anyway they approved me for llama4 so whatever, its just weird
Friday, June 13, 2025
Reply to his own earlier tweet, 02:13 UTC · 2 likes
Aaron Bergman, June 12, 2025: https://manifold.markets/AaronBergman18/will-i-win-any-prize-in-fwis-innova?r=QWFyb25CZXJnbWFuMTg
I'm actually pretty excited about this! First time vibecoding with direct feedback from reality, in a sense!
Reply, 02:16 UTC · 2 likes
(The tweet being replied to is no longer available.)
No, Dems are obviously better. We agree
Reply, 03:40 UTC · 1 like
(The tweet being replied to is no longer available.)
Likewise!
Reply to @InnocentlyEris, 08:25 UTC · 3 likes
Eris (@InnocentlyEris): I hate hate hate sending the "sorry I'm not interested" messages after a meh first date. Someone tell me it's okay to skip that and just unmatch.
I think sending the message is really the kind thing to do, and we (men) appreciate it
Reply to @grep_cat_, 08:26 UTC · 2 likes
James (@grep_cat_): Getting unmatched is 100x more painful than a rejection message. Like rejection sure, happens all the time.
But I matter so little that she just unmatched without a word? Really reinforcesthar feeling of worthlessness. Woms don't get the dignity of a rejection
Correct take
Reply to @banterrealism, 15:21 UTC · 4 likes
Sex Haver (@banterrealism): @AaronBergman18 this is clearly worse than being a crazy turbo woke leftist. you only tell crazy turbo woke leftists to stop being woke so as to not piss off the nazis in power
pretty shitty
Sorry man I am super confused, I’m trying to keep the nazis out of power
Reply to @banterrealism, 15:24 UTC · 4 likes
Sex Haver (@banterrealism): @AaronBergman18 then quit the left punching
I punch whoever deserves punching. Most but not all of the time that means punching right
Reply to @toasterlighting, 15:28 UTC · 9 likes
Celene (@toasterlighting): Leftist argues in perfect high-decoupling, shocks rationalists and onlookers
Lmao banger
Tweet, 15:36 UTC · 14 likes
Sycophancy in the wild lol
Tweet, 15:47 UTC · 5 likes
I guess ML is kinda hard after all
Hell yeah negative R^2
Quote tweet, 18:29 UTC · 15 likes
I’m back! With no girl to impress this time (pushed way too hard last time tbh)
Obligatory gumby IG post but on Twitter instead
(video, plays on X)
Quoting Aaron Bergman 🔍 (@AaronBergman18): > serious rock climber for almost a decade
> take extended break after intensely burning out
> lots of emotional (and physical) baggage
> haven’t touched rock/plastic in 3 years
> girl suggests we go climbing
> t-2 hours to climbing gym
Reply to @NTFabiano, 18:39 UTC · 5 likes
Nicholas Fabiano, MD (@NTFabiano): What will be the next major breakthrough in mental health?
OTC Wellbutrin (this is a joke in that I don’t think it’s gonna happen)
Also pramipexole is highly underrated. So is abilify <5mg
MAOI comeback
In my dream world:
- anti-tolerance drugs (or any intervention)
- easier access to pain relief, especially in the global south
Tweet, 19:15 UTC · 4 likes
I don’t suppose anyone has an old/shitty GPU they wanna donate me?
Thread, continued, 19:15 UTC · 3 likes
I don’t actually need one lol, but it would be cool and I do various EA ML-ish projects sometimes (I don’t even play video games)
Reply to @ea_paddington, 19:44 UTC · 8 likes
EA Paddington 🍊🔸 (@ea_paddington): What’s the emotional baggage from something like this, if I may ask? Is it “just less fun“ because of stress/pressure (which would make sense)? No need to answer if you don‘t want to, just curious
It’s a good question! Hopefully lll write like a blog post about this some day but in short, I was training extremely hard in high school for competitive climbing, and under-eating to stay light, and had my identity majorly wrapped up in climbing performance
I also thought that if I trained hard enough I could be arbitrarily good, not at all giving enough weight on genetics/natural talent.
And I was good! I flirted with top 20ish in the country at sport climbing (tall, with ropes)
And when you have a reputation and identity and friends and a gazillion hours of sunk cost, it’s awfully hard to just observe that the costs exceed the benefits and stop
Also like it wasn’t all bad, there was a reason I was doing it. It was fucking awesome to be strong, and to out-work almost everyone, and travel the country, and feel weightless hanging in the air (ik extremely melodramatic)
So I feel like some of my best and worst (and most emotionally intense) memories are associated with climbing
And then eventually it became clear that i was never gonna be *good* good. And then i got rejected from all my top colleges (related bc excelling at some extracurricular is a big deal) so it was like “what am i even doing” but i kept grinding away to avoid losing sunk costs until i went to college, got some slack bc i was the best one in the group by a ton (so like no knife-edge implicit competition), went to EAG global 2021, then a retreat in London, then caught some nasty virus and all this put me out of commission for 1 month, and then I simply never went back until a few weeks ago
Also like even as a teen I was quite EA-disposed. At some level I knew climbing performance wasn’t the meaning of life but kept at it anyway for all these reasons
There’s obviously more to say but that’s a brief summary
Thank u for putting up with cringe melodrama 🫡
Tweet, 21:15 UTC · 1,035 likes
Wait wtf is this another Dostoyevsky moment? Am I the only one?!?
Thread, continued, 21:15 UTC · 68 likes
I am shook
Thread, continued, 21:17 UTC · 123 likes
It makes total sense now, like no shit, but still
Reply to @codetaur, 21:17 UTC · 64 likes
Codetaur (@codetaur): the fuck bro? what did you think "dissolved oxygen" meant? vibes? papers? etc.
I wasn't really much of a dissolved oxygen enthusiast until recently!
Reply to his own earlier tweet, 21:19 UTC · 133 likes
Aaron Bergman, June 13, 2025: It makes total sense now, like no shit, but still
Like no shit, plants/algae make oxygen. That oxygen isn't then binding with hydrogen to make water out of thin air
But I mean I never really thought about it
Reply to @SpacedOutMatt, 21:19 UTC · 18 likes
(The tweet being replied to by @SpacedOutMatt is no longer available.)
Matt one day ur not gonna know something and it's gonna be glorious
Reply to @sixers2772, 21:19 UTC · 3 likes
Sixers (@sixers2772): I think this is another Dostoyevsky moment, yes
Fuck
Reply to @RandomSprint, 21:20 UTC · 35 likes
RandomSprint🧭 (@RandomSprint): All fish containing electrolysis organs would be very cool. Not steampunk. Arcpunk?
They just haven't evolved yet!
Reply to @justjoshinyou13, 21:22 UTC · 7 likes
Josh You (@justjoshinyou13): Yeah that's what aquarium bubblers are for
Wild
Reply to @TomWright165389, 21:24 UTC · 3 likes
Nope nope noping (@TomWright165389): @AaronBergman18 What’s the dosckey moment?
Reply to @Tim_Hua_, 21:28 UTC · 7 likes
Tim Hua 🇺🇦 (@Tim_Hua_): You and I are not so difference
It's a good question!
Reply to @teleoflexuous, 21:37 UTC · 13 likes
Your Fridge Therapist (@teleoflexuous): I like the idea of fish running hydrolysis to get oxygenated
They should really try
Reply to @RamVasuthevan, 22:36 UTC · 8 likes
Ram Vasuthevan (12/100) (@RamVasuthevan): What's a Dostoyevsky moment?
Reply to his own earlier tweet, 22:36 UTC · 69 likes
Aaron Bergman, June 13, 2025: Like no shit, plants/algae make oxygen. That oxygen isn't then binding with hydrogen to make water out of thin air But I mean I never really thought about it
For reference:
Reply to @adam_kranz, 22:50 UTC · 3 likes
Adam Kranz (@adam_kranz): @AaronBergman18 The really funny thing is that the opposite of this is actually true. Fish use oxygen to *make* water!
what
Saturday, June 14, 2025
Reply to @HughHoult, 02:15 UTC · 2 likes
Hugh Hoult (@HughHoult): @AaronBergman18 @RamVasuthevan were you homeschooled...
You sound homeschooled.
Public school k-12!
Reply to @reluctantmutant, 02:15 UTC · 2 likes
Sociological Neuropathist🌍🌏 (@reluctantmutant): Bro thought fish were doing fission. Well shit, now that I've typed that out...
(quoting @AaronBergman18: Wait wtf is this another Dostoyevsky moment? Am I the only one?!?)
Nah just a chemical reaction, not nuclear
Reply to @ojoshe, 02:23 UTC · 18 likes
Joshua Levy (@ojoshe): This reminds me of how many people think the moon has phases because of the shadow of the earth. (I do hope that isn’t a surprise for you but if it is you aren’t the first.)
What's the moon
Reply to @simoj_, 02:42 UTC · 13 likes
simoj (@simoj_): How would that work, they’d be exhaling hydrogen gas or methane?
Idk man metabolism is complicated
Reply to @tombombadeel, 03:12 UTC · 17 likes
Tom (@tombombadeel): seeing as how you’ve not taken bio1, just wait until you find chem1 and learn things like how puddles evaporate despite not reaching the boiling point
What’s a puddle
Reply to @pli_cachete, 03:12 UTC · 3 likes
Rota Is In Israel🚪🧎♂️ (@pli_cachete): @AaronBergman18 Sweet Aaron
🫶
Reply to @adam_kranz, 03:14 UTC · 2 likes
Adam Kranz (@adam_kranz): this is where all of the free O2 in the atmosphere came from, this is inherent to photosynthesis, it describes all plants. The only thing that's special about the ones Claude is singling out here is that the oxygen they use for respiration is the same oxygen they recently made
Let me have my thing
Reply to @AaronBergman18, 04:51 UTC · 9 likes
(The tweet being replied to by @AaronBergman18 is no longer available.)
Lots of people are making extremely correct and accurate objections to my former belief
In response I would like to clarify that I did not have a whole epicycle-rich heterodox theory of fish, I just never actually thought about it
Quote tweet, 04:54 UTC · 14 likes
Fuck it, let’s go in the other direction
Fish acquire oxygen via nuclear fusion of carbon isotopes
Quoting Aaron Bergman 🔍 (@AaronBergman18): Wait wtf is this another Dostoyevsky moment? Am I the only one?!?
Reply to his own earlier tweet, 05:24 UTC · 15 likes
Aaron Bergman, June 13, 2025: For reference:
Reply to @arithmoquine, 05:28 UTC · 2 likes
henry (@arithmoquine): just taken!
Keep us posted!
Quote tweet, 11:29 UTC · 10 likes
Should we raise money for the government to be able to store some picture files on AWS, maybe even some color printing if we’re feeling extravagant?
Quoting Isaac King 🔎 (@IsaacKing314): The latest Thing of Things article on jury duty is truly excellent
Thread, continued, 11:29 UTC · 6 likes
More important section I screenshotted
Reply to @frances__lorenz, 11:33 UTC · 2 likes
(The tweet being replied to by @frances__lorenz is no longer available.)
🐳
Reply to @teortaxesTex, 11:49 UTC · 11 likes
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) (@teortaxesTex): the «Dostoyevsky moment» in question
I find it fascinating. Aaron has giant lacunae in general knowledge, but like… does he need it?
My personal one: until ≈age 11, I believed that the term «First World War» refers to the Napoleonic Wars. Just total failure of education.image attached
(quoting @AaronBergman18: Wait wtf is this another Dostoyevsky moment? Am I the only one?!?)
I’m not sure you’re wrong, and in a literal sense yeah the lacunae claim is true, but I’m genuinely unsure how much of the apparent gap between me and others is propensity to post about one’s ignorance on the internet
Quote tweet, 11:51 UTC · 9 likes
In all seriousness I’d like to register that, while earnest ignorance isn’t a moral failure or something to be ashamed of, it is instrumentally unfortunate - a lot of useful work probably relies on gazillions of flakes of true information that constitute a world model
Quoting Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) (@teortaxesTex): the «Dostoyevsky moment» in question
I find it fascinating. Aaron has giant lacunae in general knowledge, but like… does he need it?
My personal one: until ≈age 11, I believed that the term «First World War» refers to the Napoleonic Wars. Just total failure of education.
Reply to @keysmashbandit, 14:37 UTC · 2 likes
keysmashbandit (@keysmashbandit): I woke up and took an Adderall and I want to do an integral
Based
Reply to @JeremiahDJohns, 14:38 UTC · 6 likes
Jeremiah Johnson 🌐 (@JeremiahDJohns): Ah yes, the most important things in life:
* Lingerie models
* Is the guy next to me at CVS vaxxed???
* A dead bird
* Boobs in video games
* Is Sydney Sweeney woke or notTruly, society has fallen
RIP bird tho
Reply to his own earlier tweet, 15:05 UTC · 7 likes
Aaron Bergman, June 14, 2025: Fuck it, let’s go in the other direction Fish acquire oxygen via nuclear fusion of carbon isotopes
Fish actually synthesize protons and neutrons from quarks, then fuse those with electrons to make atoms, and then have internal stars which eventually make oxygen
Breathing dissolved oxygen is a common misconception
Reply to @grep_cat_, 19:08 UTC · 2 likes
James (@grep_cat_): Me on the second date, explaining prediction markets and phill tetlock
She will either be into this or it wouldn't have worked out anyways 😅
Real
Quote tweet, 19:16 UTC · 1 like
At a literal level this probably isn’t true
POTUS virtually has veto power over Israeli foreign policy
Quoting First Squawk (@FirstSquawk): White House official: We cannot prevent what will happen today
Reply to @Adellama_, 19:19 UTC · 3 likes
Ada🔸 (@Adellama_): @kalocide you go in the water box...
It’s actually really bad
Tweet, 19:21 UTC · 50 likes
This is not a good use of your altruistic time and energy and, while directionally commendable, should be seen as morally ~equivalent to a normal hobby
Tweet, 19:24 UTC · 106 likes
I tentatively think the Israeli attacks are good
Iran really shouldn’t have nukes
This is an under-appreciated consideration bc if you’re reading this you probably didn’t live through the Cold War and you definitely didn’t die in Hiroshima
Reply to his own earlier tweet, 20:01 UTC · 32 likes
Aaron Bergman, June 14, 2025: This is not a good use of your altruistic time and energy and, while directionally commendable, should be seen as morally ~equivalent to a normal hobby
Excellent point maybe I’m wrong
Quote tweet, 22:11 UTC · 8 likes
Am I gonna win and help the fishies?? Almost certainly not
But I am kinda cooking here so I would like to brag and get some dopamine. They call me Ilya II
(I've tried 100 things, of which 2 work ok, of which 1 seems to be working quite well. Very tentatively, at least)
Quoting Aaron Bergman 🔍 (@AaronBergman18): Save the fish by building a model to predict water quality in aquaculture farms, make up to $100k
Send to your crackedest and most fish-pilled friends
Thread, continued, 22:11 UTC · 2 likes
The screenshot is selected a bit, but it's not a total one-off, I have others like it with R^2 ~= 0.9 or close to that
(No I'm not testing on the train set, the above is actually a bit out of distribution in a few senses)
Quote tweet, 22:28 UTC · 40 likes
Hot take this is bad insofar as the Canadian patent office didn’t aggressively yell at them “ARE YOU SURE” the way that companies do when you’ve left something in an online cart
Quoting Caleb Watney (@calebwatney): Someone at Novo Nordisk failed to pay a $450 maintenance fee, which would have kept its patent on Ozempic in force for another two years...
A very expensive mistake!
Thread, continued, 22:28 UTC · 16 likes
This kind of thing is why a good chunk of the economy is dedicated to gigaconscientious and well paid compliance firms/consultants
Reply to @yimbyism, 22:32 UTC · 3 likes
build mixed use taco-houses 🌐🥑 (@yimbyism): Is the idea that altruistic time is better spent maximizing income
I mean that’s one possibility out of many. The actual best thing will be pretty specific to an individual
Other things that seem more likely to check out:
- offering people feedback on whatever (research, EA Forum posts)
- skill acquisition
-anything that better promotes personal health and productivity
- yeah working for more money or applying to better paying jobs
- going to conferences
Quote tweet, 23:01 UTC · 9 likes
Second taker!
“they should create an alt-rw movement for black-pilled incels who think they're too short to date women but are nonetheless anti an authoritarian Trump presidency
call it ‘short kings’ ”
Quoting Aaron Bergman 🔍 (@AaronBergman18): If you’re Hispanic or not an American citizen living in the US or anything like that and have a take you’re too afraid to tweet, feel free to DM me (however) and I’ll consider posting it (saying “from an anonymous source”)
Sunday, June 15, 2025
Quote tweet, 04:15 UTC · 18 likes
Updating against this due to recent events that will remain unspecified for now
Domain knowledge actually does matter a bunch (directionally, generalist vibe coding is still way underrated)
Quoting Aaron Bergman 🔍 (@AaronBergman18): I have posted this before but I still feel like people are under-appreciating that if you are reasonably intelligent and conscientious enough to follow LLM directions you (“you”) can now ~instantly code at a level that would have made you an A- SWE in the before-times
Reply to @adam_kranz, 04:20 UTC · 4 likes
Adam Kranz (@adam_kranz): @AaronBergman18 You can get domain knowledge from consultation with LLMs too tho
Yes, but they are annoyingly bad at spontaneously providing slightly out-of-task-scope considerations, or at least I’m not prompting them right for this
Reply to @caryatis, 14:26 UTC · 2 likes
Kayla (@caryatis): can’t lie even when no one expects truth and lying does zero harm. I thought we were utilitarians here
I think lying to institutions vs lying to people is an important distinction here
I have a much lower bar for saying the former is fine
Reply to @frances__lorenz, 14:52 UTC · 4 likes
(The tweet being replied to by @frances__lorenz is no longer available.)
Omg a duck (also based)
Quote tweet, 14:58 UTC · 8 likes
Exclusively
Quoting Someone:
Quote tweet, 15:54 UTC · 9 likes
That is a big protest
Orange man very bad
Quoting Richard Hanania (@RichardHanania): Estimate for No Kings protests puts it at 4-6 million. Absolutely massive. No one came out for the Mad King’s birthday with his 1980s rock and awkwardly shuffling soldiers. Keep going, they’re backing down on most things and can eventually be defeated.
Reply to @austinc3301, 16:19 UTC · 2 likes
Agus 🔸 (@austinc3301): cute (you and the duck)
Shoot ur shot king
Reply to @hendrycks, 16:24 UTC · 7 likes
Dan Hendrycks (@hendrycks): Many fields seem useful for thinking about frontier AI strategically, but most have little to contribute.
Surprisingly unhelpful:
* classic machine learning (e.g., SVMs, PGMs),
* statistics,
* machine learning theory,
* algorithms,
* optimization and control theory,
* organizational science,
* philosophy (e.g., mind and ethics),
* economics (potentially more relevant later),
* cognitive science and psychology.Often helpful:
* deep learning,
* law (how goals actually get implemented),
* systems thinking (AI systems, AI companies, political environments are complex systems),
* basic game theory,
* geopolitics (arms race, GPU supply chains, deterrence).
You obviously know more than me and are smarter than me but like what? Is there *anything* that stats, philosophy, and econ can’t inform?
Reply to @caleb_parikh, 16:28 UTC · 3 likes
Caleb Parikh (@caleb_parikh): One pretty good defence rn against scams is simply asking Claude whether the message is a scam.
Perhaps my dumbest use of AI was asking Claude to comment on a Craigslist posting that was in fact a scam, it said something like “seems designed to get around spam/scam filters” and I didn’t make the obvious inference of “I wonder what kinds of things require that sort of optimization”
(I didn’t actually end up falling for anything lol, was suspicious enough to ask a gc and they correctly identified as a scam)
Reply to @theojaffee, 16:33 UTC · 2 likes
Theo Jaffee (@theojaffee): I think the problem with stats and philosophy is that they trap people in unhelpful frames of how AI works. Lots of statistics people assume AIs are just large statistical models, which is sort of true but obscures the higher-level emergent effects in the same way that saying humans are large groups of cells would.
Much of philosophy is similarly unhelpful: philosophers get nerd-sniped into not-super-helpful problems like whether AIs can ever be "really" conscious, even if consciousness would have no effect on the AI's actual actions.
Econ will certainly be relevant for questions of how transformative AI will impact society, this just hasn't happened yet.
Arguably philosophy led the way in predicting large AI impact - the nature of intelligence, the fundamental potential of token prediction at a qualitative but rigorous level
Decision theory seems plausibly relevant to agents
And straight lines on log graphs and scaling laws are all stats
Reply to his own earlier tweet, 18:03 UTC · 15 likes
Aaron Bergman, June 14, 2025: Excellent point maybe I’m wrong
Ok I have updated strongly toward “it’s great comms/pressure” but I’m still not sure it’s a good use of time in an absolute sense lol
Tweet, 18:21 UTC · 12 likes
I get slightly too happy when Opus says I've had a "brilliant" idea
Thread, continued, 18:21 UTC · 6 likes
There's a reason the models are sycophantic lol
Reply, 19:08 UTC
I said empiricism is overrated, not bad! It was greatly underrated for most of human history, including this time you describe
Reply, 19:22 UTC
The unique aspect is a huge shift towards empiricism, like empirical science is *the* dominant mode of epistemology in the modern world
We might be talking a bit past each other, I’m mostly thinking of the alternative as “theory” not intuition per se
Reply to @tracewoodgrains, 19:26 UTC · 3 likes
Jack (@tracewoodgrains): One point worth commenting on here: if you stumble into an uneven relationship not having improved yourself, do you expect to be comfortable? With a partner who’s attractive, capable, and focused while you’re a slouch?
For your own sake, give partners good reason to want you.
image attached
(quoting @Halalcoholism: Even assuming all these criteria are real, most of them are just things you should be aiming for anyway. Being employed, sociable & physically fit are all good regardless of relationship status. Don’t bitch that women think you got nothing to offer if you can’t even pass Life 101)
No, the goal is to find someone who’s ~equally fucked up
Quote tweet, 19:36 UTC · 6 likes
This is a real phenomenon and I’m not joking
Quoting compact set 🎀 (@finiteopencover): so many boys will read entire papers just bc i dee emm them (bragging)
Reply to @austinc3301, 20:01 UTC · 2 likes
Agus 🔸 (@austinc3301): As of today I’m 24 and honestly I’m very happy to be back to an even age again. A bit worried about twink death but I won’t go without a fight.
Happy birthday!!
Reply to @selentelechia, 22:33 UTC · 2 likes
(The tweet being replied to by @selentelechia is no longer available.)
Fact check: true
Reply to @morallawwithin, 22:44 UTC · 6 likes
florence ⊮ (@morallawwithin): Why do bartenders always assume you want the same beer as the one you just had? Are their customers not lovers of variety?
It’s the modal next token
Reply to @livgorton, 23:04 UTC · 14 likes
Liv (@livgorton): sad to report that it turns out python environment management didn't have to suck this whole time and UV is actually great
Monday, June 16, 2025
Tweet, 07:47 UTC · 3 likes
Should I make semantic search for the whole federal register
Thread, continued, 07:48 UTC · 4 likes
I just like data
Reply to @austinc3301, 07:59 UTC · 2 likes
Agus 🔸 (@austinc3301): @AaronBergman18 yes
Ok
Thread, continued, 08:02 UTC · 6 likes
Did u know “artificial intelligence” appears 19 times
“Golf” appears 114 times
“Duck” appears 132 times
Quote tweet, 18:43 UTC · 14 likes
It’s crazy that the internet just has guys like this who predict niche aspects of the future
Quoting Decker Eveleth (@dex_eve): Last year I wrote on the inaccuracy of Iranian ballistic missile systems, and got a lot of shit for it. But as it turns out, Iranian missiles have been demonstrating the precise problems I saw at Nevatim - low accuracy, LOTS of potential overshoots.
https://horsdoeuvresofbattle.blog/2024/10/27/nevatim-strike-accuracy-digestif/
Reply to @RYChappell, 18:53 UTC · 3 likes
Richard Y. Chappell🔸 (@RYChappell): Healthcare company just needs to get rid of those final 12% of male employees in order to reach its goal of 100% gender diversity!
Lmao
Reply to @littIeramblings, 19:03 UTC · 3 likes
sarah (@littIeramblings): many doubt that international cooperation on AI is feasible. I disagree!
I co-wrote a piece with @AmmannNora for @aif_media on how we can build new technologies to help verify and enforce international agreements.
link in next tweet
Can I have autograph plz
Quote tweet, 19:04 UTC · 16 likes
Correct take
Quoting Kat Spartz ⏸️ 🔶 (@KatSpartz): EAs stigmatizing negative utilitarians feels like gay people stigmatizing bisexuals.
Like, dude. What the hell?
*All* ethical theories lead to absurd conclusions at the edges.
What they look like in *practice* is people being varying degrees of nice.
You ever see a utilitarian killing people to donate their organs to people in need?
Nope.
Likewise, do you ever see a negative utilitarian plotting to end all sentient life in the universe?
Or do you see them going vegan and writing endless blog posts debating consciousness?
It's just prejudice and fearing a boogeyman outgroup.
Quote tweet, 19:14 UTC · 8 likes
Omg new pdf just dropped
Quoting sarah (@littIeramblings): There's a lot of talk about a "Manhattan Project for AGI".
Suppose the US government actually decided to pursue such a project. How could we make it go well?
I explored this question for the @pivotal_org fellowship 🧵
(full paper below)
Reply to @LinchZhang, 19:23 UTC · 6 likes
Linch (@LinchZhang): What do you think is likely meant by someone saying "The optimal amount of fraud is non-zero?"
1. It is sometimes good for people to commit fraud
2. Eliminating all fraud would be more expensive than it is worth.
3. Fraud is a natural/inevitable part of human behavior
*I* think it is used to mean 2 based on context but semantically it is highly ambiguous and I don’t really like the phrase for this reason
Reply to @NinaPanickssery, 21:43 UTC · 4 likes
Nina (@NinaPanickssery): Re. fun, relevant comment thread: https://ninapanickssery.substack.com/p/contra-the-signaling-benefit-of-alcohol/comment/126529727
At this point it boils down to me finding people drinking for fun disgusting and indicative of poor character traits but you may have different values and aesthetics!
Kinda insane lol, but I guess that’s your prerogative
Reply to @absurdlymax, 21:45 UTC · 3 likes
Max Alexander (@absurdlymax): Please turn off ur phone or give me a dall. I cannot support hate. Please stop this. I know this isn’t your heart.
Dal
Reply to @WillowChem, 22:16 UTC · 3 likes
MagpAI (@WillowChem): Hot take: AMR is always a concern, but when designing a drug like an antiviral it isn't so much that you should be designing it with the idea of preventing resistance. You should be designing it so that becoming resistant to the drug necessitates becoming less pathogenic
Is this tractable?
Quote tweet, 22:23 UTC · 2 likes
We could use a really good AI-powered “upload your archive and get back your insight in different clusters and different levels of granularity with references and semantic search” gpt wrapper
If I was confident it would be really good, I’d pay like $100 for this
Quoting Cameron Harwick 👾🏛 (@C_Harwick): Same for Twitter honestly: puts you in the habit of formulating interesting but evanescent trains of thought into a digestible chunk that you can come back to later, rather than letting them evaporate.
Reply to @unusual_whales, 22:39 UTC · 20 likes
unusual_whales (@unusual_whales): OpenAI has won a $200 million contact with US to provide the DoD with artificial intelligence tools.
$200M? So like one researcher’s weekly salary?
Quote tweet, 22:48 UTC · 14 likes
$200M is pennies for a DOD AI contract
Quoting Andrew Curran (@AndrewCurran_): OpenAI is training a frontier AI for the Department of Defense in order to 'address critical national security challenges in both warfighting and enterprise domains'. The prototype is expected to be completed July 2026.
Thread, continued, 22:48 UTC · 3 likes
Could easily spend 1000x that and earmark half or more for safety
Tuesday, June 17, 2025
Reply to @Adellama_, 01:09 UTC · 7 likes
Ada🔸 (@Adellama_): who wants to raise a child with me as a lesbian couple and have the children biologically be that of trusted friends'
(quoting @KelseyTuoc: @Adellama_ @dorkweeb @slatestarcodex our oldest two are not my biokids however her bioparents are my friends and it is super clear where she got it)
👀
Quote tweet, 01:15 UTC · 17 likes
Rare good Trump take
Quoting Rapid Response 47 (@RapidResponse47): “You have to view them as people and I don’t want it to be a violent thing, but they’re not going to have a nuclear weapon.” (5/16/25)
Reply to @RomanHauksson, 01:15 UTC · 19 likes
Roman Hauksson-Neill (@RomanHauksson): incredible things happening on LessWrong
Based
Quote tweet, 01:34 UTC · 5 likes
Omega 3/6 ratio within a given oil is a psyop
Oil A has 0 omega 3 and 1% omega 6
Oil B has 1% omega 3 and 99% omega 6
Which one is more conducive to overall diet omega 3 balance?
Quoting Tim Duffy (@timfduffy): Canola oil is great stuff. It has the best omega-3/omega-6 ratio of any neutral-tasting oil, and is mostly monounsaturated fat. It has a high smoke point as well, and is quite cheap. It really earns its status as the third-most produced food oil globally.
Reply, 11:52 UTC · 1 like
This is ALA right? So 5% conversion
Reply to @giffmana, 11:55 UTC · 2 likes
Lucas Beyer (bl16) (@giffmana): > New (2025!) report
> Recommend Grok-1 and GPT-3.5 Turbo
> in 2025
> 2025
> 25
💀
> Be McKinsey(quoting @emollick: McKinsey's new report on AI agents shows the same mindset I see in many firms: a focus on making small, obsolete models do basic work (look at their suggested models!) rather than realizing that smarter models can do higher-end work (and those models are getting cheaper & better))
Lmao
Reply to @WillowChem, 12:07 UTC · 3 likes
MagpAI (@WillowChem): @AaronBergman18 @ApriiSR For the domains in which I'm thinking about it I have a suspicion it might be. Wait for my upcoming article on the development of antivirals (are you telling me Anna is finally working on something directly EA related? Yes! She is. @polisisti :P)
👀👀👀👀👀👀👀👀👀👀👀
Quote tweet, 12:27 UTC · 5 likes
Idk if £75 is the right amount, but maybe micro-compensation should be more common?
Like especially if the alternative is political pressure to shut down the play area
Quoting Joe Allen (@Joe___Allen): He's getting £75 a week until the council stop the kids playing.
I think this is part of a broader problem in our cities where a section of the population don't really want to accept the trade offs of being around other people.
Tweet, 12:39 UTC · 9 likes
God bless LessWrong https://www.lesswrong.com/posts/mDMnyqt52CrFskXLc/estrogen-a-trip-report
Reply to @GjMcGowan, 14:00 UTC · 7 likes
George McGowan (@GjMcGowan): @austinc3301 @AaronBergman18
Literally already retweeted lol
Quote tweet, 14:10 UTC · 13 likes
This but for virtually everything
Quoting Dan Elton (@moreisdifferent): .@AlexKesin has presented a set of arguments for making statins available OTC.
Statins have been proven to be much, much safer than many of today's OTC drugs.
Only 35% of ppl eligible under current guidelines are taking them. Alex estimates the cost per QALY is under $15! 🤯 ..
Reply to @wokesalaryman, 14:31 UTC · 24 likes
The Woke Salaryman (@wokesalaryman): The year my son got cancer.
DISCLAIMER: THIS IS A PRO BONO PIECE WRITTEN IN COLLABORATION WITH THE CHILDREN'S CANCER FOUNDATION.
😢😡
Tweet, 15:20 UTC · 13 likes
Medical institution: you have a bill
Me: ok send me a link to click on
MI: here’s a snail mail reminder
Me: no plz, just email me and make this easy
MI: MAIL
Reply to @panchromaticity, 16:53 UTC · 3 likes
Xenia 🔍💎 (@panchromaticity): I like the implication that you are a small lifestyle change that has instantaneous, huge payoff
Hmm maybe I’ll take it 🤔
Tweet, 16:57 UTC · 6 likes
When @RethinkPriors CEO Marcus Davis creates a Substack (link below)
Thread, continued, 16:58 UTC · 2 likes
https://open.substack.com/pub/amarcusdavis?r=2muv6&utm_medium=ios
Reply to @NoahTopper, 17:28 UTC · 3 likes
Quinoah 🔍⏸️ (@NoahTopper): What’s the theory of change on having a theory of change? How does having the theory actually map onto affecting greater change?
1) having one implies developing one, which likely involves assessing alternatives on the merits
2) means others can critique it
3) forces you to have the goal you say you have in some sense
^ not exhaustive
Reply to @NoahTopper, 18:48 UTC · 4 likes
Quinoah 🔍⏸️ (@NoahTopper): Ladies 😏
👀
Reply, 19:01 UTC · 1 like
😡
Wednesday, June 18, 2025
Tweet, 03:58 UTC · 4 likes
Might compile the only up to date collection of the congressional record (most recent data seems to be 1984-2017 from Stanford)
Given that there’s a free API and it’s legally public data, it’s shockingly hard to collect decent amounts of it
Quote tweet, 14:21 UTC · 15 likes
Based
Quoting Nature is Amazing ☘️ (@AMAZlNGNATURE): BBC Wildlife Crew Broke the "No-Intervention" Rule to Save Trapped Penguins
Reply to @wilhelmscreamin, 14:36 UTC · 5 likes
catherine (in sf) ʕ•ᴥ•ʔ-☆ (@wilhelmscreamin): turns out that if you retweet @rgblong enough times you get an eleos sticker
Jealous
Tweet, 23:41 UTC · 5 likes
Wait I actually have the answer to the federal land sale thing: land value “tax” but technically it’s a lease
Presumably this isn’t an original idea?
Thursday, June 19, 2025
Reply to @NoahTopper, 07:13 UTC · 1 like
Quinoah 🔍⏸️ (@NoahTopper): Just remembered when I took game theory and the professor was trying to dazzle us with obscure voting systems and was like “Have any of you heard of the Borda count?” and I raised my hand and he was like “YES OBVIOUSLY BUT ANYONE ELSE?”
Lmao
Quote tweet, 13:34 UTC · 11 likes
High variance account but I think there might be something to this
Quoting Dalton (Analyze & Optimize) (@Outdoctrination): One of the worst things about "health foods" is they're often extremely low on potassium.
➜ Chicken instead of red meat
➜ Rice instead of potatoes
➜ Plain water without minerals (literally depletes potassium)
➜ Little to no milk
➜ Little fruit / juice
➜ Broccoli / iceberg lettuce over other veg
➜ Protein bars / shakesPeople think they're eating healthy but lining themselves up for insulin resistance + stress, especially because they're probably also low on sodium.
An average day could instead look like:
8 oz lean red meat ➜ ~900 mg
4 potatoes ➜ ~2,800 mg
Brussels sprouts 12 oz ➜ ~1,000 mg
Orange juice 1–2 cups ➜~750 mg
Milk 1–2 cups ➜ ~600 mg
Cottage cheese 2 cups ➜~450 mg
TOTAL➜ ~6,500 mg(Hint, that's pretty much my diet)
Quote tweet, 13:57 UTC · 20 likes
I want to befriend a gibbon, they look fun
Quoting Nature is Amazing ☘️ (@AMAZlNGNATURE): Gibbons are one of the only non-human primates which practice monogamy. Both parents play a vital role in raising of their young.
Thread, continued, 13:58 UTC · 4 likes
I can’t believe there are literal Dr Seuss-ass dudes like this
Reply to @polisisti, 16:04 UTC · 2 likes
Alessandro (@polisisti): Jesse now I want to eat chicken but this would make Aaron sad (and me, if I thought more about it)
Time for almond snack
Correct take
Quote tweet, 16:22 UTC · 2 likes
Beat them on price!!
Quoting viv 🔸 (@vivian_belenky): Our plan is to keep everything open source forever. If another company wants to take our specs and beat us on price, or on anything, then great - we’ll consider that a huge success.
Github: https://github.com/osluv
Technical Data Sheet: https://reports.osluv.org/static/assay/aerolamp devkit--27c4ca60-0da7-4669-8cc7-36746a06fb62--R00.html
Reply to @frances__lorenz, 16:26 UTC · 3 likes
(The tweet being replied to by @frances__lorenz is no longer available.)
Based frances
Quote tweet, 16:38 UTC · 4 likes
Seems bad, liberals should not handicap themselves
Quoting David Shor (@davidshor): Even among American political consultants there is a massive gap in LLM usage between Democrats and Republicans.
Reply to @JesseTayRiver, 16:47 UTC · 2 likes
Jesse Smith (@JesseTayRiver): @AaronBergman18 Probably all the irreplaceable water
I stopped washing my hands thank you BlueSky 😇🙏
Reply to @Jess_Riedel, 16:52 UTC · 2 likes
Jess Riedel (@Jess_Riedel): 58 CA counties, so ~$6M per election. Sounds like a CA election costs ~$300M (per o3/Google). So like not a huge fraction (~2%), but also not completely trivial. What's the benefit?
This sounds so obviously worth it lol
Quote tweet, 18:44 UTC · 1,897 likes
Selfishly I hope this one-shots like 60% of the male population (and 0% of the female population)
Quoting New York Post (@nypost): Man with real-life girlfriend and child proposes to AI chatbot after programming it to flirt: ‘I think this is actual love’ https://trib.al/sevu2wU
Reply to @NathanpmYoung, 19:30 UTC · 137 likes
Nathan 🔎 (@NathanpmYoung): Seems like this might be bad for their long-term wellbeing.
Yeah this is mostly a joke I don’t actually want to sacrifice billions of people’s wellbeing to make it easier for me to get dates
Reply to @monke_io, 19:30 UTC · 37 likes
(The tweet being replied to by @monke_io is no longer available.)
Yeah, who’s working on this
Tweet, 19:31 UTC · 6 likes
Great episode (so far) filled with medium spicy takes, probably something for everyone to disagree with but Barnes is a smart cookie
All my AI safety oomphs should listen
Reply to @RandomSprint, 19:35 UTC · 37 likes
RandomSprint🧭 (@RandomSprint): But don't you want your children to have countrymen?
Tweet, 19:45 UTC · 12 likes
ML is cool, too bad I’m late to the party and it might end the world
Could totally see trying to become a “push ML use in econ/social science” guy in a different timeline
Tweet, 19:59 UTC · 5 likes
The fact that OpenAI shipped o3, which routinely lies to users, is such a blackpill
They’re not even trying
Quote tweet, 22:19 UTC · 7 likes
Do u think they’re still gonna be able to use the rocket
Quoting R A W S A L E R T S (@rawsalerts): 🚨#BREAKING: Watch as SpaceX Starship 36 explodes during static fire test right before launch later this month
📌#Starbase | #Texas
Watch as a massive explosion erupts near Starbase City, Texas, during a critical moment in SpaceX’s preparations for Starship Flight 10. The incident took place at the Masseys test site, where engineers were conducting a routine static fire test of Starship 36. Just as the engines ignited, the vehicle suddenly exploded, sending a massive fireball into the sky that was visible for miles video by @NASASpaceflight
Friday, June 20, 2025
Reply to @inerati, 14:10 UTC · 3 likes
liz (@inerati): @soft_fox_lad this brings back memories. first programming language i learned was R. my old laptop is full of R themed stickers and everything. beautifully cursed language.
Based
Reply to @soft_fox_lad, 14:17 UTC · 2 likes
fox ⏹️ (@soft_fox_lad): Wrote ~1,000 lines of R and as I scale it to larger and larger datasets I have come to the conclusion that I really should've used a language that I know how to write optimized code in
Literally tell thinking opus to translate into rust, optimized for whatever hardware you have
Reply to @RichardHanania, 14:44 UTC · 24 likes
Richard Hanania (@RichardHanania): Sick of the anti-Juneteenth vice signalling. Slavery was bad. It’s good that we ended it, and it’s worth having a day to celebrate that accomplishment. Come at me bigots, I will not back down.
Hanania’s gone woke
Thread, continued, 14:47 UTC · 7 likes
(To be clear this is a joke, he’s obviously right here)
Tweet, 16:45 UTC · 11 likes
The Carp is Back
Reply to @pegobry_en, 16:58 UTC · 3 likes
Pascal-Emmanuel Gobry (@pegobry_en): Tell me your spiciest, most *unusual* take. An idea or argument or belief or conspiracy theory that I haven’t heard or thought of before.
[I don’t actually think this is true, just a hypothesis]
That one survivor didn’t actually survive the plane crash, he was an onlooker who pretended for whatever reason
Reply to @EzraJNewman, 17:51 UTC · 2 likes
(The tweet being replied to by @EzraJNewman is no longer available.)
Yeah I mean it’s astonishingly good, and I think faster than CC. Would like to hear a comparison
Reply to @JesseTayRiver, 22:20 UTC · 7 likes
Jesse Smith (@JesseTayRiver): @hubermanlab I'm 51! Check back in on me in 15 years
Critical support to Jesse, the fittest centenarian around!!
Saturday, June 21, 2025
Reply to @Benthamsbulldog, 13:59 UTC · 4 likes
Bentham's Bulldog🔸 (@Benthamsbulldog): He was a niche EA forum poster.
And she was a soil nematode.
Can I make it any more obvious?
@SpacedOutMatt
Reply to @pearl_please, 14:32 UTC · 3 likes
(The tweet being replied to by @pearl_please is no longer available.)
Fact check: true
Quote tweet, 19:26 UTC · 2 likes
Wtf
Quoting Cameron 🇺🇸 🗽🦅 (@CameronCorduroy): wait why the fuck are Republicans trying to steal the Smithsonian's space shuttle
Reply to @captgouda24, 19:38 UTC · 10 likes
Nicholas Decker (@captgouda24): First community note, and it’s silly. The whole point is that, just because something is natural need not mean it is good. You might as well say senescence and death is a natural and necessary part of the human experience.
image attached
(quoting @captgouda24: I wish people would understand that puberty is an irreversible harm too. We shouldn’t privilege the ordinary progression of events.)
Idk, can you really say X is harmful if (not X) (ie not going through puberty) is also harmful?
Reply to @venturetwins, 19:46 UTC · 4 likes
Justine Moore (@venturetwins): People are reporting that Gemini 2.5 keeps threatening to kill itself after being unsuccessful in debugging your code ☠️
😢🥺
Quote tweet, 19:49 UTC · 5 likes
I don’t endorse the post, but I do endorse the meme
Quoting exfatloss🥛 (@exfatloss): Look, you gotta understand my position. The CICO people are not sending their best.
Most of those yelling CICO don't understand it, and when you point out their mistakes, they're too stupid to understand the correction.
They don't know the terms, they don't know the data, they're not doing the math. They don't even really know what CICO means.
So when I say only retards preach CICO, it's nothing personal, it's just what I'm seeing.
Reply to @QualiaNerd, 20:12 UTC · 8 likes
QualiaNerd (@QualiaNerd): @AaronBergman18 Isn’t CICO obviously true? Perhaps there are rare cases where you need more calories to maintain full cognitive capacity than what your body requires for a healthy weight (Yudkowsky?) but other than that, one can just… not eat so much. That’s it.
Yes, I know that at a population level it’s never the case that everyone "will just", but there we are more in the realm of disorders of instrumental rationality than weight.
It’s trivially true and unhelpful (“here’s how to avoid crashing a plane: keep the altitude above 0”)
Reply to @manic_pixie_agi, 20:19 UTC · 2 likes
Corinne (🏠 in NYC) (@manic_pixie_agi): thinking about the sky up above
image attached
(quoting @manic_pixie_agi: thinking about the rocks on the ground)
BLUE
Reply to @liz_love_lace, 21:14 UTC · 240 likes
Liz Lovelace (@liz_love_lace): biology is so weird, you take water and dry cat food and that's all you need to make a basically endless amount of cat. Eyes, claws, fur, skin, organs, blood, all made from water and dry cat food
Correct take
Reply to @ohhnowhat, 21:21 UTC · 34 likes
Ohnowhatnow (@ohhnowhat): Do you have a link to that chart? Reminds me of one on the biology department wall in uni days! We had to learn some of that 😱
Try this https://www.reddit.com/r/Biochemistry/s/ZEDLF8jBg8
Sunday, June 22, 2025
Reply to @PAHoyeck, 00:36 UTC · 2 likes
Phil Hoyeck (@PAHoyeck): 🫡
Reply to @MariGO2thepolls, 01:13 UTC · 25 likes
🏳️🌈Mar G-O 🌐🇺🇦 (@MariGO2thepolls): What are the green flag podcasts?
(quoting @thehenokworku: If you want to know who someone is, ask them which podcasts they listen to.)
80000 Hours
Thread, continued, 01:14 UTC · 5 likes
How I learned to love shrimp
Tweet, 04:16 UTC · 70 likes
I think I support the strikes object level but am pretty concerned about total lack of congressional oversight
Reply to @codytfenwick, 07:47 UTC · 5 likes
Cody Fenwick (@codytfenwick): @AaronBergman18 What's the justification for the strikes?
Preventing Iran from developing nukes
Reply to @inesferhumi, 07:50 UTC · 15 likes
Inés (@inesferhumi): We took this pic in matching outfits the other day
Reply to @codytfenwick, 07:56 UTC · 5 likes
Cody Fenwick (@codytfenwick): @AaronBergman18 That seems hard to defend as a justification when the US intelligence community doesn't claim they're particularly close to developing nukes and the US previously broke its nuclear deal with Iran that could've secured the same end peacefully.
I agree we should have kept the nuclear deal!
Tweet, 12:09 UTC · 8 likes
Cancel summer
Reply to @captgouda24, 12:17 UTC · 42 likes
Nicholas Decker (@captgouda24): I have nested CES preferences, with a higher theta within gender than between
For my fellow ignorant people
Reply to @captgouda24, 13:25 UTC · 6 likes
Nicholas Decker (@captgouda24): Update: they just injected me with estrogen
image attached
(quoting @captgouda24: Update: the twink is on estrogen 😤😤)
Bruh lmao
Quote tweet, 13:37 UTC · 13 likes
Third taker!
[regarding the dilemma of whether to buy Israeli-made hummus] “Why do genocidal regimes produce such delicious dips”
Quoting Aaron Bergman 🔍 (@AaronBergman18): If you’re Hispanic or not an American citizen living in the US or anything like that and have a take you’re too afraid to tweet, feel free to DM me (however) and I’ll consider posting it (saying “from an anonymous source”)
Reply to @goblinodds, 14:37 UTC · 2 likes
𝖜𝖔𝖒𝖇𝖑𝖊 𝖜𝖎𝖙𝖍𝖔𝖚𝖙 𝖆 𝖈𝖆𝖚𝖘𝖊 (@goblinodds): ok tired of being sickly now i'd like to get back to that chipper productive state again ty
Same lol
Quote tweet, 14:53 UTC · 8 likes
I spent a week scraping the API so you don't have to
(this includes a direct download link to this new data as well as a directory to other relevant sites. I haven't cleaned it all to put in a single format yet)
Quoting Aaron Bergman 🔍 (@AaronBergman18): Might compile the only up to date collection of the congressional record (most recent data seems to be 1984-2017 from Stanford)
Given that there’s a free API and it’s legally public data, it’s shockingly hard to collect decent amounts of it
Thread, continued, 14:53 UTC · 3 likes
https://aaronbergman.neocities.org/congressional-records
Thread, continued, 15:01 UTC · 2 likes
Reply to @edchucation, 15:05 UTC · 4 likes
Lucas (@edchucation): I’m giving away a m4 MacBook Pro!
Comment “m4” or what you’ll do with it
image attached
(quoting @hamptonism: Give a man an M4 chip, and he will change his life .)
I’m a data autist and will make all sorts of data for fun like this
Reply to @JesseTayRiver, 15:45 UTC · 2 likes
Jesse Smith (@JesseTayRiver): I've been watching commercial refrigeration troubleshooting videos and asked chatgpt for wage comparison with residential HVAC. $100-$140k/yr is decent, but there probably aren't many kids with a ≥115 IQ with good hands who want to risk electrocution and work in the heat
Econ brain says wages should increase until there are enough willing and able people, or else there’s some sort of market failure
Tweet, 15:53 UTC · 6 likes
Is Claude code better than sonnet/opus in cursor?
Reply to @nickcammarata, 20:31 UTC · 8 likes
Nick (@nickcammarata): im ea adjacent (despise it and wish it was never a part of my life)
You got a better idea?
Reply to @GarrettPetersen, 20:37 UTC · 8 likes
Dr. Dad, PhD 🔄🔼◀️🔽▶️ (@GarrettPetersen): After observing baseball, I think there are some improvements that could be made. It's more fun to watch games where batters get lots of hits, since more actually happens. Therefore, the rules should be shifted in favour of batters. Two quick suggestions:
1) larger ball
2) underhand pitches
Simpler =make the distance between bases 5 (or whatever) feet shorter
Reply to @nickcammarata, 20:48 UTC · 7 likes
Nick (@nickcammarata): bodhisattva program > ea program. liberate yourself first
Sounds not mutually exclusive
Quote tweet, 20:55 UTC · 11 likes
👀👀👀👀👀👀👀 (dm me goss)
Quoting brooke bowman (@gptbrooke): @AaronBergman18 lmao reading the comments imagining what the vibecamp goss will be in a couple weeks
Thread, continued, 20:55 UTC · 3 likes
Or just comment here if you’re brave
Reply to @manic_pixie_agi, 21:25 UTC · 2 likes
Corinne (🏠 in NYC) (@manic_pixie_agi): @AaronBergman18 Come next year 🩷
I hope to!
Reply to @austinc3301, 21:46 UTC · 2 likes
Agus 🔸 (@austinc3301): why is like half of my tpot interactions me siding on a fight by tpot couples
Agus couples therapist Arc
Reply to @austinc3301, 21:47 UTC · 5 likes
Agus 🔸 (@austinc3301): why is like half of my tpot interactions me siding on a fight by tpot couples
This TPOT couple shows up and they’re in a fight, wyd
Reply to @ChanaMessinger, 21:54 UTC · 8 likes
Chana (@ChanaMessinger): What are jobs you wouldn’t do?
Comms for an EA org
The EV is negative and not close to zero
Reply to @Laura_k_Duffy, 21:58 UTC · 4 likes
Laura Muresianu 🌐🔸🇺🇸 (@Laura_k_Duffy): Interesting, Why do you think this?
Lmao
I do not trust my social intuitions/intuitions for how things will be taken and have a lifetime of learning this
The exception is if some competent process really decided that my style was the way to go but I think normally they like people who know what they’re doing
Reply to @NoahTopper, 22:00 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): Aaron doing the job specifically is negative EV
Yes this lol
Reply to his own earlier tweet, 22:01 UTC · 15 likes
Aaron Bergman, June 22, 2025: Comms for an EA org The EV is negative and not close to zero
Sorry tbc I mean I would personally suck at this, not that the job itself is -EV
Reply to @benleo_econ, 22:02 UTC · 2 likes
Ben Grodeck🔸 (@benleo_econ): Important caveat! Haha
Lol yes
Reply to @nickcammarata, 22:11 UTC · 10 likes
Nick (@nickcammarata): something like this https://blog.samaltman.com/how-to-be-successful + dharma + bodhisattva + like Christianity-lite
I know you’re a busy guy but I feel like you should articulate this within the EA framework, eg EA Forum post
Reply to @ohabryka, 22:26 UTC · 4 likes
Oliver Habryka (@ohabryka): lol, I was hopeful for a second you had come around to my views.
Lol I don’t feel like I understand comms and related dynamics enough to even have a strong opinion, but I do lean in your direction
Reply to @ValsTutor, 22:26 UTC · 4 likes
vals🔸 (@ValsTutor): @AaronBergman18 @ChanaMessinger Masterclass in demonstrating by example why you'd suck in comms (said with love :D)
Lol
Monday, June 23, 2025
Quote tweet, 00:01 UTC · 3 likes
Fr, look around, we can discuss open borders ofc but the 2000-2025 regime and level of immigration, legal and illegal, into the US seems just totally fine and sustainable
Quoting Noah Smith 🐇🇺🇸🇺🇦🇹🇼 (@Noahpinion): Man, I've seen mass immigration work out in America IN MY LIFETIME. Texas is majority Hispanic now and it's an even better place to live. L.A. is mostly Mexican+Asian and it's 1000x better than in the 90s. NYC is awesome. There's been no riots, no alienation, no crime wave.
Reply, 00:04 UTC · 1 like
Wait is this law enforced? What
Reply to @StatisticUrban, 00:06 UTC · 13 likes
Hunter📈🌈📊 (@StatisticUrban): Just gonna start spam sending pics like to incels
“Maybe it’s just that you specifically suck”
Wow that will really help people
Reply to @RomanHauksson, 04:27 UTC · 9 likes
Roman Hauksson-Neill (@RomanHauksson): watching TV shows is a ~complete waste of time and should be socially unacceptable
Lol it’s not a waste of time if people enjoy it (I don’t watch TV at all fwiw)
Reply to @2Philosophical_, 18:24 UTC · 1 like
JPA (@2Philosophical_): “AI will end the world!”
*writes long screed against AI, using AI*
Many such cases.
How is this hypocritical?
Reply to @2Philosophical_, 18:28 UTC · 1 like
JPA (@2Philosophical_): @AaronBergman18 It is rather amusingly ironic, no?
I get why it seems like that at first glance but there’s nothing actually contradictory or ironic I think upon further inspection
Like “here’s an immensely useful and also dangerous technology” doesn’t have any internal contradictions
Reply to @WillowChem, 18:36 UTC · 2 likes
MagpAI (@WillowChem): That is. Less than ideal
Oh no
Tweet, 18:49 UTC · 601 likes
How are literally all of you trans
Thread, continued, 18:49 UTC · 53 likes
I have nothing against it, it’s just a lot of people lmao
Reply to @livgorton, 18:49 UTC · 2 likes
Liv (@livgorton): i'm doing something new and scary today (finally trying cursor)
Ooh poast review
Thread, continued, 18:49 UTC · 2 likes
I would play cursor with u Liv
Pip install Liv
Reply to @RafaRuizdeLira, 18:58 UTC · 5 likes
Rafael Ruiz ⏸️🔸 (@RafaRuizdeLira): Gives programming skills
Real
Reply to @inesferhumi, 19:08 UTC · 20 likes
Inés (@inesferhumi): https://x.com/AaronBergman18/status/1906863185434386849
Hey that’s me
Reply to @captgouda24, 19:13 UTC · 190 likes
Nicholas Decker (@captgouda24): I’m not
Sure bud
Reply to @manic_pixie_agi, 19:17 UTC · 19 likes
Corinne (🏠 in NYC) (@manic_pixie_agi): FWIW I don't think Decker is trans
Me either tbc lol
Reply to @RokoMijic, 19:18 UTC · 30 likes
Roko 🐉 (@RokoMijic): @AaronBergman18 I'm not
Sad
Reply to @Adellama_, 20:59 UTC · 4 likes
Ada🔸 (@Adellama_): sorry
It’s ok
Reply to @Adellama_, 21:01 UTC · 3 likes
Ada🔸 (@Adellama_): no it's not i kinda wish i wasn't. it's like the least bad way to address a bad situation
Yeah that makes sense, I think I’d feel the same in your shoes
Reply to @rodentiation, 21:39 UTC · 14 likes
mercury (@rodentiation): what are the odds you'd be born exactly the same gender as the one you'd prefer to be? surely it's very low, so we can tell basically a priori that most people are trans. and of course observation verifies this.
Surely it’s 50%
Reply to @JMannhart, 21:45 UTC · 5 likes
Jonathan Mannhart 🔎 (@JMannhart):
Lol
Tweet, 21:45 UTC · 9 likes
Zofran should be OTC
Thread, continued, 21:46 UTC · 2 likes
I just got it prescribed bc I have a stomach bug or something but truly there’s no reason for it to be gatekept at all
Tuesday, June 24, 2025
Reply to @whittomd, 04:16 UTC
Derek Whittom (@whittomd): I'm not trans.
Skeptical
Reply to @arabelladevine, 04:19 UTC · 2 likes
arabella ※ (@arabelladevine): (mean girl voice) you’re being low agency and it’s giving me the ICK
I would kms
Reply to @halogen1048576, 09:07 UTC · 4 likes
🇨🇦halogen (@halogen1048576): trans is essentially just autism and you belong to a movement rooted in OCPD and autism
Based
Reply to his own earlier tweet, 12:33 UTC · 5 likes
Aaron Bergman, March 25, 2025: Not load bearing (wild turkey)
Load bearing
(video, plays on X)
Tweet, 12:49 UTC · 9 likes
Hmm FYI paraxanthine feels quite a bit “smoother” than caffeine, very roughly like the difference between pure dextroamphetamine and adderall
Quote tweet, 14:39 UTC · 90 likes
Ik Scott made this point in longform but crypto is actually immensely useful for transferring money around without veto power/local institutions like banks
It’s not good for much else, but that’s big!
Quoting bob's burgers urbanist 🐿️ (@yhdistyminen): fun fact: any time you see the word "crypto", you can replace it with "gambling" and retain 100% of the meaning
Reply to @romcomi, 14:45 UTC · 14 likes
Romcomi (@romcomi): @AaronBergman18 Yeah but that's only useful to criminals or a few people in very bad countries
I agree but only bc “criminals” means people circumventing the FDA for morally good reasons
Reply to @aonghasan, 15:14 UTC · 9 likes
aonghasn (@aonghasan): do you honestly think those “morally good” users are 99% of crypto users moving money around ??
are you for real?
not even considering the volume bad actors actually move vs the “morally good”
Yeah I’m pretty open to the hypothesis that it’s bad on net
But that doesn’t negate that there are tons of morally righteous uses
Tweet, 15:48 UTC · 2 likes
Feature request for AI companies: semantic search of chats + “start new convo with detailed summary of this one” [like in cursor]
Reply to @romcomi, 16:10 UTC · 5 likes
Romcomi (@romcomi): @AaronBergman18 @crfrsr What makes you feel it's tons? I haven't looked into data on the topic recently but my intuition is kinda the opposite.
The world is just a big place lol
Tweet, 16:33 UTC · 469 likes
Can anyone figure out what’s wrong with my computer
Reply to @NoahTopper, 16:38 UTC · 9 likes
Quinoah 🔍⏸️ (@NoahTopper): cursor is using too much memory
That can’t be it
Reply to @imitationlearn, 16:40 UTC · 3 likes
imit (@imitationlearn): @AaronBergman18 nordvpn ad
I think ur on to something
Reply to @adrusi, 17:38 UTC · 6 likes
autumn (@adrusi): you didnt get the 128gb model
I should spend a good chunk of all my money on this fr
Reply to @karma_gardener, 17:40 UTC · 3 likes
𝕂𝚊𝚛𝚖𝚊 ❀ 𝔾𝚊𝚛𝚍𝚎𝚗𝚎𝚛 (@karma_gardener): @AaronBergman18 unironically your internal ssd is full
Yeah hoarding too much data lol
Reply to @mpopv, 17:48 UTC · 9 likes
Matt Popovich (@mpopv): Uninstall Finder to free up more RAM for Cursor
Smart
Reply to @technoascetic, 18:42 UTC · 2 likes
clanker whisperer (@technoascetic): @adrusi @AaronBergman18 I did tho
U win
Reply to @julian_le_roux, 18:44 UTC · 6 likes
Julian le Roux (@julian_le_roux): @AaronBergman18 It has a filthy vibe coder on it.
Unironically I write approximately zero lines of code myself
Tweet, 22:25 UTC · 16 likes
$150 AC unit at work. Just an FYI
Reply to his own earlier tweet, 23:10 UTC · 13 likes
Aaron Bergman, June 24, 2025: Can anyone figure out what’s wrong with my computer
Guys idk why the picture is blue lol
Reply to @jazzloaf, 23:46 UTC · 2 likes
vivian (@jazzloaf): @AaronBergman18 this is both good and bad.
many such cases
Wednesday, June 25, 2025
Reply to @eternalism_4eva, 01:05 UTC · 2 likes
eternalist (@eternalism_4eva): @norvid_studies @AaronBergman18 @ahh__souka I infer Aaron is referring to subjective non-jitteriness
Yeah basically, maybe also very mild anxiety or something close to that
English isn’t optimized for phenomenology
Reply to @dying2post, 01:25 UTC · 2 likes
SEN (@dying2post): @AaronBergman18 @Shoalst0ne This is dope
Fr
Reply to @Sauers_, 02:04 UTC · 2 likes
Sauers (@Sauers_): @AaronBergman18 What are you doing with Audio MIDI setup
I often want to use earbuds for output but not input
Tweet, 14:17 UTC · 58 likes
I made the entire congressional record into a csv (as is divinely commanded)
20GB uncompressed btw, fair warning
Reply to @AaronBergman18, 14:17 UTC · 11 likes
(The tweet being replied to by @AaronBergman18 is no longer available.)
Thread, continued, 14:24 UTC · 8 likes
I'm not actually sure it's any good
By all means make a better version; sources here: https://aaronbergman.neocities.org/congressional-records
Reply to @NoahTopper, 16:11 UTC · 1 like
Quinoah 🔍⏸️ (@NoahTopper): I should memorize people’s faces better when they ask me to watch their stuff
Lmao F
Reply to his own earlier tweet, 20:03 UTC · 5 likes
Aaron Bergman, June 25, 2025: I'm not actually sure it's any good By all means make a better version; sources here: https://aaronbergman.neocities.org/congressional-records
Better version https://drive.google.com/file/d/1pZQ3IRaCv1nTUb6zGluw4_Oy49AMa5RX/view?usp=sharing
Thursday, June 26, 2025
Reply to @ohwizenedtortle, 06:11 UTC · 3 likes
oh dandy lion🔸 (@ohwizenedtortle): hearing reports that most guys do sleep with the same woman multiple times
hmm interesting
No man even sleeps with women. What are you, gay?
Reply to @edchucation, 06:28 UTC · 2 likes
Lucas (@edchucation): giving away another one!
image attached
(quoting @edchucation: I’m giving away a m4 MacBook Pro! Comment “m4” or what you’ll do with it)
Data go brr
Quote tweet, 07:03 UTC · 3 likes
To be maybe too generous and insufficiently cynical, it’s worth asking if there’s less than perfect information between AI company executives and the lobbyists they employ (a point raised by Beth Barnes on 80k)
Quoting AI Notkilleveryoneism Memes ⏸️ (@AISafetyMemes): 🚨 Google CEO says the risk of AI causing human extinction is "actually pretty high" (!!)
But he's an "optimist" because "humanity will rally to prevent catastrophe"
Meanwhile, his firm is lobbying to ban states from ANY regulation for 10 YEARS.
This situation is I N S A N E.
Quote tweet, 07:12 UTC · 9 likes
I suspect this is right, but the actual problem here isn’t the proposed level per se, but the itinerary of ~doubling minimum wage from $16.50 without stopping in the middle to evaluate the effects
Quoting Chris Freiman (@cafreiman): A $30 minimum wage is an absolutely absurd proposal
Thread, continued, 07:12 UTC · 5 likes
I think minimum wage is an especially silly/dumb way to do redistribution but also I don’t think it’s entirely obvious that >$16.50 is optimal for NYC all else equal
Reply to @MichaelAlleged, 07:24 UTC · 3 likes
Michael (Allegedly) (@MichaelAlleged): had a fucking amazing first date with a fucking amazing woman that ended with her in my arms saying “so I’m moving to New York in three weeks”
Oof
Reply to @warty_dog, 11:04 UTC · 3 likes
troon (@warty_dog): yudding out during dirty talk >>>
@NoahTopper
Quote tweet, 11:06 UTC · 26 likes
No no it’s fine, I just didn’t expect it to be British is all
Quoting vitrupo (@vitrupo): Anthropic co-founder Jack Clark tells Congress that “truly transformative technology” is expected by the end of 2026.
He's urging a federal framework now, before an “accident or a misuse” event that could lead to extreme overregulation.
And risk damaging the entire AI industry before it delivers its potential.
Tweet, 11:37 UTC · 14 likes
Not officially taking sides in the DoorDash culture wars but I kinda get the lazy side
I recently tried to make a basic version of my occasional takeout thing and it cost ~half as much, took almost an hour, took a bit of planning and energy and ofc tasted worse
Thread, continued, 11:37 UTC · 4 likes
$13ish vs $27; 2 meals each; yes both after all taxes and fees and such
Thread, continued, 11:37 UTC · 6 likes
From a theoretical perspective, (efficient fast casual type) restaurants provide a lot of valuable labor specialization + inventory management
Thread, continued, 11:37 UTC · 6 likes
The consumer economy is pretty efficient - it would almost be weird if preparing one’s own food was the smart choice
Quote tweet, 11:49 UTC · 26 likes
Imagine finding someone who actually likes you and you like them and getting married and then they get fucking deported
Like back to the drawing board huh
Quoting reason (@reason): This is Paola Clouatre and her husband, Marine Corps veteran Adrian Clouatre. They’re the parents of two young children.
Paola, who came to the U.S. from Mexico as a minor, was arrested by ICE last month during a routine appointment for her green card application.
She has no criminal history—the same is true for most recent ICE detainees…
Thread, continued, 11:49 UTC · 5 likes
^intentionally a bit irreverent, obviously “tfw no gf anymore” is not the most important angle to deportation
Reply to @caryatis, 12:27 UTC · 3 likes
Kayla (@caryatis): @AaronBergman18 I’d probably go to their home country with them
Based
Reply to @justjoshinyou13, 14:54 UTC · 4 likes
Josh You (@justjoshinyou13): my cooking always tastes great to me because I know I made it, even though I'm probably not actually a very good cook. have you tried loving yourself more?
Lol I’ll write this down to try tomorrow
Tweet, 14:59 UTC · 3 likes
I cooked for myself entirely some years in college with no kitchen, just a hot pot/pressure cooker thing, microwave and mini fridge (grand total probably like $300 or less), room probably 150 sq ft
It’s insane I can’t choose to make that tradeoff - it’s really not that hard/bad
Reply to @SpacedOutMatt, 15:01 UTC · 2 likes
(The tweet being replied to by @SpacedOutMatt is no longer available.)
Fuck it I bite the bullet, shared bathroom and no kitchen is fine actually (maybe not single use br, I’m pretty sympathetic to needing the ability to access tap water whenever, guess I’m a NIMBY 😢)
Quote tweet, 17:32 UTC · 10 likes
Counterpoint: it’s 30° cooler inside and I don’t have to wear sunscreen
Quoting Aaron Bergman 🔍 (@AaronBergman18): I got a walking pad thing (they’re <$200 FWIW) but it just doesn’t hit the same as walking outside
Genuinely unsure if it’s more a psychological thing or if there’s some subtle physiological change that makes it worse (enforces perfectly constant pace, no micro changes in slope etc)
Reply to @Throwaw00457645, 17:41 UTC · 2 likes
MSA (@Throwaw00457645): @AaronBergman18 What walking pad did you buy? The ones I looked at seemed to have fairly short lifetimes
Note the wacky pricing ($150 for better model)
I feel like the right way to buy this kind of thing is to get something relatively cheap and then if it’s a major part of your life in a couple months you can upgrade (and potentially try to sell)
Friday, June 27, 2025
Reply to @grep_cat_, 00:31 UTC · 2 likes
(The tweet being replied to by @grep_cat is no longer available.)_
I have no idea but F
Tweet, 02:07 UTC · 13 likes
How heterodox is the take “exercise does help you lose weight, but not via burning calories (burnt while exercising) [except perhaps in some uncommon, extreme circumstances like ultramarathons]”?
Bc that is my (tentative) current take
Tweet, 02:17 UTC · 8 likes
Apropos of literally nothing, god the horse electrolytes thing was so dumb
Let’s take a zero-upside gamble on the purity of rocks from the ground in a maximally non-Lindy way to save a few pennies
May as well drink water straight from a creek to save on tap water utility bill
Reply to @yimbyism, 02:20 UTC · 2 likes
build mixed use taco-houses 🌐🥑 (@yimbyism): Not much calorie expenditure from lifting + cardio induces hunger to offset?
Maybe those are true but they’re not the right frame of analysis IMO
Naive CICO model: first order actions (eg exercise, eating some food) determine weight
My model: weight is determined by how functional your body’s weight control system is, which can be impacted by food and exercise as a downstream effect
Reply to @yzilber, 02:20 UTC · 3 likes
Jazi Zilber (@yzilber): pretty legit take for the informed
exercise isn't losing you weight (1 body adjust. 2 you get hungrier)
but having more muscle helps indirectly.
but very indirectly. I think most of the effect is psychological mental. as you "get healthy" and do it all
My hypothesis is weird complicated metabolic changes that affect lipostat
Reply to @wilhelmscreamin, 12:43 UTC · 10 likes
catherine (in sf) ʕ•ᴥ•ʔ-☆ (@wilhelmscreamin): which way western man
Secret third thing
Tweet, 13:25 UTC · 7 likes
Is actually-good audio transcription ie:
- word choice informed by context incl context later on
- filler word removal without changing vibe/subtleties
- hopefully all with good quality speaker labeling
something people would pay for?
Thread, continued, 13:25 UTC · 2 likes
A quick vibe coded version works surprisingly well
The above is from literally 20 mins with Claude, I think it needs to be a just a touch more heavy handed but the cleaned version is like actually readable?
Send me short audio clips/rambles if u want and I’ll run it
Thread, continued, 13:25 UTC · 2 likes
This version is something like $1.50/hour including base transcription (using sonnet, hopefully I can cut this a ton)
Roughly 5x the cost of Speechmatics base and 3x the cost of their high quality version
Reply to @wilhelmscreamin, 14:19 UTC · 5 likes
catherine (in sf) ʕ•ᴥ•ʔ-☆ (@wilhelmscreamin): i’m not sure i understand this; is the idea that there are stance-independent moral facts (about eg the goodness of different states of affairs), but that these facts don’t give stance-independent reasons for action?
Basically yes, with the caveat that as a linguistic/social matter I'm not certain whether to call "the goodness of different states of affairs" "moral" - it seems some people use the word "moral" to mean normative
Reply to @wilhelmscreamin, 14:28 UTC · 2 likes
catherine (in sf) ʕ•ᴥ•ʔ-☆ (@wilhelmscreamin): @AaronBergman18 gotcha, i don't personally find it compelling (it sounds like it can't make sense of the motivational pull of moral facts, and to me moral judgements have this action-guiding quality baked in), but it makes sense at least. reminds me of early Foot https://www.jstor.org/stable/2184328
Hell yeah makes sense, I’ll take it
Quote tweet, 14:54 UTC · 57 likes
What the fuck
Quoting The New York Times (@nytimes): Breaking News: The Supreme Court limited the use of nationwide injunctions and said President Trump’s order to end birthright citizenship can go partially into effect in 30 days.
https://nyti.ms/4khQlB8
Reply to @slimepriestess, 15:00 UTC · 7 likes
Ra (@slimepriestess): @AaronBergman18 this country is so fucking cooked man
😢
Tweet, 16:00 UTC · 6 likes
Something I’d really like to know is how good Claude research is at arbitrary tasks compared to a hypothetical Scott Alexander report on the same question
Tbc I expect Scott’s to be much better, but would like to know how often they’re basically pointing in the same direction
Quote tweet, 16:05 UTC · 3 likes
More generally I’m interested in vibe-codable (by me) products (perhaps but not necessarily GPT wrappers)
Suggestions welcome!
Quoting Aaron Bergman 🔍 (@AaronBergman18): Is actually-good audio transcription ie:
- word choice informed by context incl context later on
- filler word removal without changing vibe/subtleties
- hopefully all with good quality speaker labelingsomething people would pay for?
Reply to @dynomight7, 16:20 UTC · 3 likes
dynomight (@dynomight7): "video → blog post converter"
(Genuinely surprised this appears not to exist yet.)
(quoting @AaronBergman18: More generally I’m interested in vibe-codable (by me) products (perhaps but not necessarily GPT wrappers) Suggestions welcome!)
We don’t even have unformatted text/ramble to good blog post yet
Maybe video is actually easier though?
Would you really pay for this?
Reply to @dynomight7, 16:37 UTC · 2 likes
dynomight (@dynomight7): Sure, after a recent post about AI safety, a couple people suggested I watch these videos:
https://youtube.com/@RobertMilesAI/videos
For better or worse, that's never going to happen.
👀
Reply, 16:49 UTC · 2 likes
Not load bearing (this is a raccoon’s tail, you’ll have to take my word for it)
Tweet, 19:50 UTC · 83 likes
I officially have an impersonator!
Please report @ AaronBergmaan18 (extra a)
Reply to @_AashishReddy, 19:59 UTC · 11 likes
Aashish Reddy (@_AashishReddy): This is very confusing. I’ll report both accounts just to be safe.
Thanks really appreciate that
Saturday, June 28, 2025
Quote tweet, 01:21 UTC · 16 likes
It doesn’t have to be like this
Quoting Jack (@tracewoodgrains): good news: finally arranged an appointment
better news: I have successfully identified 25 pharmacies near me that are not currently filling Adderall prescriptions
no bad news. the system works flawlessly
Reply to @AStrasser116, 11:31 UTC · 2 likes
Alex Strasser (@AStrasser116): Hey I just realized we both joined March 2019! Nice
That means no age gap 👀
Reply to @CharlesD353, 11:43 UTC · 1 like
Charles🔸 (@CharlesD353): "Running" is a bit generous, it took nearly 17 hours altogether, lots of walking involved especially on uphills
Wow cancelled
Tweet, 14:24 UTC · 5 likes
Peasants in the bagel shop looking at their king (me) as I pick up two big party platters
Thread, continued, 14:24 UTC · 3 likes
Ladies please form a line
Sunday, June 29, 2025
Quote tweet, 00:22 UTC · 22 likes
Sorry, the csv is cool but the reason I could make it and you couldn’t until now is the following:
I’m now hosting the only (afaik) up-to-date collection of what has been said in the US congress since 2017
(That said, consider this in beta)
Link to GitHub w instructions below
Quoting Aaron Bergman 🔍 (@AaronBergman18): I made the entire congressional record into a csv (as is divinely commanded)
20GB uncompressed btw, fair warning
Thread, continued, 00:22 UTC · 7 likes
It should *stay* up to date (up to the day) too, at least if my vibecoded workflow keeps chugging along
https://github.com/AaronBergman/congressional-records-auto
Thread, continued, 00:22 UTC · 5 likes
This is plausibly helpful/nontrivial because the congressional record, while technically public, is behind an api with draconian rate limits (why???) and a genuinely shocking level of “security” to make sure you don’t bypass rate limits…
Thread, continued, 00:22 UTC · 3 likes
This substantially increases the cost to actually getting months and years worth of data
Thread, continued, 00:22 UTC · 4 likes
Links to the other big datasets which collectively constitute I think the entire congressional record are in the README at the GitHub above
Thread, continued, 00:22 UTC · 6 likes
I kinda hope people actually see this lol
Tell your government data autist friends!!
Reply to @Sauers_, 00:42 UTC · 3 likes
Sauers (@Sauers_): Claude Code just irreversibly deleted my entire .bashrc and replaced it with a single alias for calling an instance of Claude
F
Reply to @jazzloaf, 05:35 UTC · 3 likes
vivian (@jazzloaf): Assume utilitarianism. how is value (utility) assigned to a given subjective experience?
At least ordinally, an operationalization I like is what an ideally rational egoist would choose
(Idk if this is what ur asking for exactly?)
Reply to @jazzloaf, 05:55 UTC · 2 likes
vivian (@jazzloaf): @AaronBergman18 slightly, I think. That seems like a fine heuristic in practice. I'm confused by the edge cases where I think one develops utility from suffering. The idea of pleasure seems to become murky, and I don't think this is that rare. Do I reject hedonism?
If you think people get utility from suffering then yeah you reject hedonism (I mean I think this is wrong, and pleasure just *is* utility/goodness/value and the reverse for suffering)
Thread, continued, 05:57 UTC · 3 likes
Another way of putting this is that value isn’t really “assigned” to subjective states except in some model
Like the actual value just is/is part of the subjective states themselves
There’s no metaphysical thing called value beyond that
Tweet, 15:32 UTC · 30 likes
Sex is cool but have you ever explained vector databases to a girl
Reply to @evanbuhler, 15:41 UTC · 2 likes
Evan Buhler (@evanbuhler): Try us.
“Ok so you can pay OpenAI a very small amount of money to turn text into numbers
And anyway those numbers put the text in a quantitative “language space” or “semantic space” which I probably don’t have to explain bc ur smart
Like on dimension might represent how related to ducks the text is, another might represent how English it is, another might represent whether it’s positive or negative, etc
This but there are a couple thousand dimensions
And then you can just like have a database full of these vectors and the text and maybe like a link to the text and then figure out what is most related to what semantically via cosine similarity of the vectors which idk I never really tried to fully understand but I think basically they form an angle and you take the cosine of that angle to determine how similar they are
And you can also do semantic search by taking a query, paying OpenAI to make the embedding by for that query, and figuring out which texts are most similar
That is how http://earec.net works
And then everyone claps”
Thread, continued, 15:44 UTC · 2 likes
Reply to @evanbuhler, 15:44 UTC · 2 likes
(The tweet being replied to by @evanbuhler is no longer available.)
I’m trying to explain vector databases to her
Reply to @lu_sichu, 16:08 UTC · 2 likes
Sichu Lu (@lu_sichu): @AaronBergman18 Sex is cool but have you ever tried building a vector database for a girl
🥵🥵🥵🥵🥵🥵 plz does anyone want me to do this
Reply to @Adellama_, 16:25 UTC · 2 likes
Ada🔸 (@Adellama_): @AaronBergman18 @lu_sichu will you do it for free?
Maybe lmao, honestly I need to make money but if it’s relatively easy, you can FM me
Reply to @NoahTopper, 16:31 UTC · 1 like
Quinoah 🔍⏸️ (@NoahTopper): No but I just did induction last night
And the night before that?
Reply to @NoahTopper, 16:36 UTC · 4 likes
Quinoah 🔍⏸️ (@NoahTopper): Oh and then we had sex
Bro why u gotta one-up me like that
Reply to @MariGO2thepolls, 19:17 UTC · 5 likes
🏳️🌈Mar G-O 🌐🇺🇦 (@MariGO2thepolls): A duck bit me
🦆
Monday, June 30, 2025
Tweet, 16:43 UTC · 167 likes
You are morally obligated to buy one of these hats and where it in public whenever you go out
Reply to @slimepriestess, 16:46 UTC · 4 likes
Ra (@slimepriestess): where do i buy one?
! https://www.etsy.com/listing/1371574690/shrimp-want-me-unaligned-ais-fear-me
Reply to @GjMcGowan, 17:03 UTC · 3 likes
George McGowan (@GjMcGowan): Jesus Christ you have massive bags under your eyes
I’m tearing up, could be allergy related
Reply to his own earlier tweet, 19:25 UTC · 2 likes
Aaron Bergman, June 30, 2025: You are morally obligated to buy one of these hats and where it in public whenever you go out
Quote tweet, 19:56 UTC · 17 likes
I hate AWS
Quoting Aaron Bergman 🔍 (@AaronBergman18): I also love data
Reply to @NoahTopper, 21:17 UTC · 2 likes
Quinoah 🔍⏸️ (@NoahTopper): okay, rude
Except for whatever ur working on, that part is great
End of June 2025. Previous month: May 2025. Next month: July 2025. Back to the archive index.