Log inSign up
Ben Chugg
965 posts
user avatar
Ben Chugg
@BennyChugg
these pretzels are making me thirsty
The Old Forest
benchugg.com
Joined December 2020
304
Following
1,028
Followers

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • user avatar
    Ben Chugg
    @BennyChugg
    May 8, 2024
    The theorists are doing well at ICLR
    39K
  • user avatar
    Ben Chugg
    @BennyChugg
    May 9, 2024
    How does one learn to appreciate poetry? Just read it? Read about it? Read reviews of it?
    63K
  • user avatar
    Ben Chugg
    @BennyChugg
    Jun 9, 2024
    the most chaotic places on the internet are by far the homepages of mathematicians. They learned how to use html and css once in the 90s and called it quits michel.talagrand.net hunch.net/~jl/ luc.devroye.org bactra.org theo.inrne.bas.bg/~bozho/index.p…
    25K
  • user avatar
    Ben Chugg
    @BennyChugg
    May 22, 2021
    Imagine how boring and dysfunctional civilization would be if humans all had the same interests. Instead we’re idiosyncratic, interested in different problems, and wrong about different things. And we make progress via positive sum collaborations Eff yes humanity. We’re awesome
  • user avatar
    Ben Chugg
    @BennyChugg
    Oct 2, 2023
    "AI will do science in its head" First they rely on classical empiricism (of Locke, Hume etc) to explain why AI will become all powerful, then switch to classical rationalism (of Descartes etc) to explain how it will do science. Just stick to one mistaken epistemology please
    user avatar
    Liron Shapira
    @liron
    Oct 2, 2023
    Me: Superintelligent AI will do science in its head; human scientists are weak minds who lean on experiments as a crutch My critics: Insane doomer claim! AI can't speed up science that much! @sama: AI could let us do a century's worth of science in 1 year
    00:00
    8.5K
  • user avatar
    Ben Chugg
    @BennyChugg
    May 30, 2023
    "train an LLM to predict the future" I'm going to have an aneurysm.
    user avatar
    Roko 🐉
    @RokoMijic
    May 29, 2023
    I just had the most amazing idea. Imagine you train an LLM with data that is all chronologically labelled. Every book, every scientific paper, every newspaper headline, every primary source from history. Where primary sources are lacking, descriptions of events or archeological
    5.8K
  • user avatar
    Ben Chugg
    @BennyChugg
    Jul 8, 2024
    won my AI capabilities bet with @kohjingyu (who's a little too successful for his own good so this was needed). Before throwing arbitrary p(doom)s around let's remember that LLMs have yet to make a novel scientific contribution
    benchugg.com
    An AI-capabilities bet
    Will a paper written by a language model be accepted to an A-tier conference or journal?
    16K
  • user avatar
    Ben Chugg
    @BennyChugg
    Jul 8, 2022
    Eternally baffled by the position that we’re near the end of science. What about depression, sleep, consciousness, dementia, aging? P vs NP, the multiverse, origin of life, AI? We barely understand any of these things, and progress in any of them will raise even more questions
  • user avatar
    Ben Chugg
    @BennyChugg
    Nov 17, 2021
    PhD theses would have been so much bolder in the pre-socratic era. “On the nature of matter: Literally everything is made of water” “Humans emerged from the mouths of big fish: An empirical investigation” “A logical proof that motion is impossible”
  • user avatar
    Ben Chugg
    @BennyChugg
    Dec 6, 2022
    I find it bizarre when elites colleges tout the success of their alumni. Admissions are so competitive precisely because they select for this kind of person. It would be much more impressive if they admitted students at random and still showed amazing outcomes
  • user avatar
    Ben Chugg
    @BennyChugg
    Jun 12, 2024
    Sure, you could read about AI, trillion dollar compute clusters, geopolitical warfare, and situational awareness. Or you could read about some philosophical issues with p-values
    P-values and counterfactuals
    From benchugg.com
    2.5K
  • user avatar
    Ben Chugg
    @BennyChugg
    Aug 7, 2023
    Replying to @kasratweets
    Wait until you read about the world record beer mile
    2.9K
  • user avatar
    Ben Chugg
    @BennyChugg
    Feb 9, 2022
    “We have not succeeded in answering all our problems. The answers we have found only serve to raise a whole set of new questions. In some ways we feel we are as confused as ever, but we believe we are confused on a higher level and about more important things.”
  • user avatar
    Ben Chugg
    @BennyChugg
    Nov 14, 2022
    Replying to @paulg
    1,500 respondents split into 10 groups, not a huge sample size. Also worth noting that the survey was of job seekers, so naturally skewed towards those with reason to have high regret at the moment