@creatine_cycle Jayden: I’m going to the sperm race
Anthropic employee: That’s a biosecurity issue. Authorities have been alerted
English
Dr. ELL
4.5K posts
Dr. ELL
@DrellLabs
Into fluid dynamics, medtech, AI in fundamental research and startups. PhD in Biomedical Fluid Dynamics. Writing Angel Checks, SF/London DMs open.
Katılım Kasım 2025
247 Takip Edilen161 Takipçiler
BG3 is a great one. Played it once with my gf (one of the few games with decent split screen) and once on my own. Recently 100% expedition 33 but no ng+.
I’m a big warhammer fan, though more 40k than fantasy/sigmar. I did really enjoy shogun and three kingdoms total war.
Currently going through ghost of tsushima for the first time (I’m really late to the game I know)
English
@DrellLabs BG3 has eaten a ton of my time, multiple honour runs complete. I alternate between that and Warhammer Total War 3 at the moment! Also enjoying the new Star Fox. HBU?
English
I feel like this but for videogames. I still haven’t played Elden Ring or Witcher 3.
Liza Libes@pensandpoison
Every time I feel too lazy to read, I remind myself that I might die before I read all of the great books on the planet—and then I get reading.
English
I'm so tired of biosecurity bullshit. The best way to fight black hat hackers are white hat hackers. Restricting information never works, just ask Barbara Streisand
Sabrina Halper@SabrinaHalper
“You should’ve been worried yesterday.” - @jacobkimmel on the risk of bioweapons, & Anthropic’s approach to limiting Fable’s responses: -You should’ve been terrified of bioweapons yesterday. They’re one of the highest-leverage ways to inflict massive damage, and we’re still woefully underprepared. -Making a bioweapon requires three things: knowing what to make, knowing how to make it, and actually pulling it off in a lab. -Steps 1 & 2 are increasingly accessible. The real bottleneck is step 3: physical reality. Lab work is hard. Most experiments fail. Materials are tightly regulated and difficult to obtain. -AI barely changes that physical layer. Claude can’t ship you research chemicals or make your experiments work. -The scary story is that AI lets a lone smart misfit do it. But the overlap between people capable of clearing the physical hurdles and people who couldn’t already find information on dangerous pathogens is probably pretty small. -Realistically, the highest-consequence bioweapon efforts still require something closer to nation-state resources. -People also worry AI could help design novel pathogens. That’s worth taking seriously. But there are already plenty of naturally occurring pathogens we don’t have immunity to. -Bottom line: I’d rather be too cautious than not. But AI didn’t suddenly create this risk - it has always been there.
English
@NeelNanda5 @andyarditi Interested in the successful papers that were 100% AI written
English
An unexpected benefit of having run 3 mech interp workshops: we have a great dataset for analysing the rise of LLM slop in submissions!
@andyarditi investigated how much AI slop we let in, how things have changed since 2024, and more
Our review process isn't entirely noise!
English
@creatine_cycle If you could plug in some neural implant that lets an AI run your body for say an hour and do your workouts for you, would you?
English
@mask_3dcg @diol2n This is not normal in most developed countries, everyone pays the same price.
English
@diol2n いやこれ 海外 行ったら割と普通じゃない
観光客用の金額と地元の生活してる人たちの金額変えてるでしょう。
私の感覚が変なのかな
日本語
يوتيوبر كوري يدعى سوهون كشف فضيحة كبيرة في بعض المطاعم اليابانية 🇯🇵 لم يتوقعها السياح الاجانب من بلد مثل اليابان:
عندما ذهب لمطعم سوشي في اليابان مع صديقته اعطته النادلة مينو باللغة الانجليزية فطلبت منيو اللغة اليابانية وهنا ترددت النادلة كثيرا قبل اعطائهم وسألتهم هل تفهمون اللغة اليابانية؟ واتضح ان المنيو مختلف كثيرا في الاسعار
القائمة اليابانية أرخص بكثير ويصل لنصف المبلغ والمطعم يتبع نظام منيو A (للمحليين) ومنيو B (للسياح) واعتبروها سرقة كبيرة للاجانب وان هذا بدأ ينتشر في مدن كبيرة مثل طوكيو واوسكا وكيتو وحتى المدن الصغيرة
وبمجرد سماع لغتك يتم اعطائك المنيو المناسب واي لغة غير يابانية يتم تسليمك منيو B وهو استغلال لعدم معرفة السياح الاسعار الحقيقة
العربية
@GergelyOrosz I feel bad that half of those highlighted sentences are things I would've wrote in the past.
English
The amount of AI writing on OpenAI's docs makes me sick.
Filler sentences for nothing. Mannerisms and phrases humans would not write. Has a human even read this?
And all of this will change how other docs are written, and how we all talk - for the worse IMO.
🤮
English
@cremieuxrecueil Is this unnecessary for someone with smooth but “older” skin? Like I don’t have acne (and never have) but I’m just older now so the skin is thinner so to speak
English
So the two things were:
1. Generally useful. More people should do this: low-dose isotretinoin. It's what celebrities do, and it's 99% of the workhorse here.
2. Useful in my case. I had redness, seemingly from demodex, and ivermectin cream (+ the isotret) ended that.
Crémieux@cremieuxrecueil
Both annoyed and grateful that all it took to get clear, smooth skin was a pair of twinks telling me two simple things to do that fixed everything. Annoyed that I didn't know earlier; grateful it's so easy.
English
@Ed_Realist @RenuMukherjee1 They’re children of immigrants but they are themselves citizens
English
@RenuMukherjee1 Almost all of them are immigrants, and it's worth asking why the taxpayers should be funding an exclusive school system for non-citizens.
English
The Asian students who attend NYC’s specialized high schools aren’t a monolith. They come from a variety of subgroups, and many are either low-income or working-class. These schools are a tool of social mobility. To suggest that they aren’t, or that they lack diversity, is a lie.
freopp.org/oppblog/in-def…
Asad🗽🍎@AsadFromNYC
"In addition to the three Black students, Stuyvesant's freshman class includes three Native American students, 21 Latino students, 39 multiracial students, 133 white students and 534 Asian students." gothamist.com/news/just-3-bl…
English
Part of what makes Stuy such a bad example in these debates is that anyone that has gone there knows Stuy is just not built for anything other than the kids that qualify via the entrance exam. It's a giant, highly competitive public school where your grades are calculated at end of semester on a hundred point scale to the hundredth decimal (I still remember comparing my 96.36 one semester to someone's 96.83... argh!! Must try harder.). There are incredible teachers, and incompetent teachers protected by the teacher's union. There is very little support for the students. It is sink or swim, and if you can't swim no one hears the ripples as you sink under the water. In other words, Stuy is too late. The work needs to come earlier, or you risk changing Stuy and making it a shadow of what it is today for the kids that not only figure out how to swim there, but become torpedoes.
Daniel Friedman@DanFriedman81
Stuy is New York’s most prestigious high school because it admits the highest-performing students as determined entirely by scores on an objective, unbiased test. Seats at Stuy are not allocated by race. They are not given out through a subjective process that can be influenced by politics or favoritism or money. The kids who got in did the best, and the kids who did not do the best cannot get in, no matter how much rich parents or woke politicians want to corrupt the fair and objective admissions competition. People like Lincoln Restler think Stuy is prestigious for reasons other than the fact that it has a rigorous objective admissions standard. He thinks the prestige these kids earned through their excellence can be redistributed to the students who are not the best that he would prefer to pretend are the best. But these students are not prestigious because they go to Stuy; Stuy is prestigious because these students go there. And if different students — inferior students— went there, then Stuy would no longer be the school that is a magnet for the smartest kids in New York and its prestige and reputation would dissipate rapidly, just like the University of California system’s has. Or maybe Restler knows this, and he just wants to destroy Stuy, because he wants to prevent these students’ talent from being measured and identified, so they can be scattered into low-performing schools where they will languish instead of having their talent cultivated. Regardless, his ideology of racial grievance and destruction is a monstrous evil masquerading as “social justice,” and it will not be tolerated.
English
Dr. ELL retweetledi
Broke: “Pause AI because rogue LLMs might kill us all!”
Woke: “AI will never be a danger! Accelerate at all costs!”
Bespoke: Although the modern “AI doomer” movement may be a kind of collective delusion, building the civilizational capacity to be able to conditionally pause global AI development is Actually A Good Thing
In the possibly-distant future, when we discover powerful new learning architectures beyond LLMs, we’re gonna want to have that option
Michaël Trazzi@MichaelTrazzi
On our way to Anthropic!
English
@dioscuri @Afinetheorem @asymmetricinfo Rare to see references of Confucian philosophy and values beyond generalized statements about East Asian cultures!
GIF
English
@DrellLabs @Afinetheorem @asymmetricinfo Yes, excellent spot, this was explicitly in my mind writing this tweet.
English
And this is good! And also good for the rich to dress and talk in a way that is no different than the middle class. And for servers to be socially of the same respect as those serve. The more one meets these norms, the better.
Matthew Yglesias@mattyglesias
The true spirit of American democracy is that the norm is for people to paint an exaggeratedly humble picture of their class and family background.
English
@natolambert Do you use it for video gen/video editing? With the right connectors ofcourse
English
Claude Fable is another big step in being able to make nice lectures based on existing educational content. Much better than Opus.
GPT 5.6 is still very far off here. Is a good example of where Claude Code being a bit easier to work across different knowledge work tasks.
English
English
@Afinetheorem @asymmetricinfo Hmm, open to being convinced on this, but what’s the argument? I think if you have a good system for producing elites, those elites SHOULD be models of ethical, epistemic, and aesthetic virtue, and having them act like non-elites will be a downgrade.
English
I think there is some revealed-preference evidence that academics are smart: quantitative PhDs at good schools, even say top-20, tend to have a fairly easy time landing very high-paying industry jobs at in-demand firms (tech, quant, etc.). Easier than the median top-20 undergrad
English
@realchillben Varies a lot between European countries but Northern/Western Europe generally (Scandinavia, Germany, France etc) very cliquey and hard to break into a friend group.
English
@sammysintech Probably a FOMO and scarcity mindset when it comes to places available for events, people over commit to secure places even with low actual probability of attending.
English
