misha

@mishayagudin

Closely related to EA and rationality communities in Moscow & Russia.

Vrijeme pridruživanja: kolovoz 2014.

Tweetovi

Blokirali ste korisnika/cu @mishayagudin

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @mishayagudin

  1. 30. pro 2019.
    Poništi
  2. 15. srp 2019.

    I am in Oxford visiting FHI this week. And will stay in Oxford/London until the 23rd. Let's get coffee!

    Poništi
  3. proslijedio/la je Tweet
    6. srp 2019.

    Morgan Kelly suggests that some of the literature on long-run persistence might confuse ‘deep origins’ with spatial autocorrelation.

    Poništi
  4. proslijedio/la je Tweet
    10. sij 2019.

    Metrics can be used well - they often aren't. And many of the issues created by their misuse can be addressed by better design. In case you haven't seen me mention it yet, here's my new paper on doing metric design more thoughtfully:

    Prikaži ovu nit
    Poništi
  5. 2. ožu 2019.

    I even had a prompt :)

    Prikaži ovu nit
    Poništi
  6. 2. ožu 2019.

    I was using the most dangerous writing app to write a piece about AI and compute, which of course discussed the GPT-2. 3 minutes into writing I realized, that time pressure makes me feel I am a language model...

    Prikaži ovu nit
    Poništi
  7. 15. velj 2019.
    Prikaži ovu nit
    Poništi
  8. 15. velj 2019.

    I think 's choose not to release the big GPT-2 partly to (a) Influence the norms of AI capabilities community. We need to consider ↓. (b) Frame AI Policy in the right way. Fake news, Kremlin's bots, and deepfakes are the issues everyone is aware of.

    Tweet je nedostupan.
    Prikaži ovu nit
    Poništi
  9. 13. sij 2019.
    Poništi
  10. proslijedio/la je Tweet

    1/ What's wrong with Moral Foundations Theory? (And how to get moral psychology right) <a long thread>

    Prikaži ovu nit
    Poništi
  11. 16. pro 2018.

    Cube composer is fun: short and easy

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    17. lis 2018.
    Odgovor korisniku/ci

    Reminds me of: > A test patient who does not need treatment is sent to 180 dentists to receive recommendations... > 27% of dentists recommend a filling to someone with no dental problems. 13 teeth suggested.

    Poništi
  13. 27. stu 2018.
    Poništi
  14. 26. stu 2018.
    Poništi
  15. proslijedio/la je Tweet

    The results of a massive replication project in psychology are published today in Advances in Methods and Practices in Psychological Science. . News release here:

    Prikaži ovu nit
    Poništi
  16. 15. stu 2018.

    Strong epistemic norms and intellectual honesty are important in the domains of extremely high-uncertainty like long-term future, where robust empirical research is almost impossible.

    Poništi
  17. 14. stu 2018.

    Jelly no Puzzle by is also great!

    Prikaži ovu nit
    Poništi
  18. 7. stu 2018.
    Prikaži ovu nit
    Poništi
  19. 7. stu 2018.

    This reminds me of self-interlocking polyhedral studied by a friend of mine. Such self-blocking structures could be used to create strong and flexible composite materials with high impact resistance because cracks cannot propagate from one block to the next. ↓ pics & links.

    Prikaži ovu nit
    Poništi
  20. 21. lis 2018.
    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·