Tal Yarkoni

@talyarkoni

academic dilettante; psychological apologist; research associate professor, UT-Austin. I like ice cream, Python, and research methods. probably in that order.

Vrijeme pridruživanja: listopad 2009.

Tweetovi

Blokirali ste korisnika/cu @talyarkoni

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

  1. 30. sij

    galaxy brain take: they already are

    Prikaži ovu nit
    Poništi
  2. 30. sij

    an unscrupulous but computationally competent researcher could probably make a fortune running an undercover "how to fake data without getting caught" course. just saying.

    Prikaži ovu nit
    Poništi
  3. 25. sij

    also interesting to see how far the goalposts have already shifted. nativists used to argue you wouldn't even be able do syntax without a lot of innate structure; they seem strangely quiet on that point these days.

    Prikaži ovu nit
    Poništi
  4. 25. sij

    a few years from now, when GPT-9 is happily writing novels, I fully expect to see writing opinion pieces complaining that it isn't *real* intelligence until it displays context-appropriate emotions or fine motor control

    Prikaži ovu nit
    Poništi
  5. 24. sij

    On behalf of the petition signatories—mostly current or past APS members—I'd like to thank the APS board for doing the right thing and issuing this statement. We appreciate it and hope to continue to help APS advance its mandate to "give psychology away" to the world.

    Prikaži ovu nit
    Poništi
  6. 24. sij

    The APS Board has issued a statement recognizing the problems with the December 18 White House letter and clarifying its position. For context, see the APS statement and linked petition to APS signed by nearly 500 researchers.

    Prikaži ovu nit
    Poništi
  7. 24. sij

    a similar thing happens with confounds. in most of the social sciences, the aggregate impact of unmodeled confounds massively swamps the postulated causal relationship of interest. but assume a 50-50 prior on any statement, and now the onus is on the critic to prove you wrong.

    Prikaži ovu nit
    Poništi
  8. 24. sij

    I think a similar kind of reasoning tacitly underlies many scientists' reverence for null hypothesis testing: an experiment has to favor one of the hypotheses, so once you've formulated H0 and H1, you can stop thinking about their prior plausibility.

    Prikaži ovu nit
    Poništi
  9. 23. sij

    fly less, give more: new blog post in which I argue that flying less is good, but there are also perfectly viable (and potentially much more effective) alternatives

    Poništi
  10. 23. sij

    Me: holy mother of unread emails this morning GitHub, x40: ONE OF YOUR DEPENDENCIES HAS A SECURIT- Me: go away GitHub GitHub: OKAY GitHub, later: GOOD NEWS, A THREAD YOU COMMENTED ON 8 MONTHS AGO WAS JUST UPDATED WITH AN EMOJI

    Poništi
  11. 23. sij

    retweeting for general audience because I'm curious: is there such a thing? or other organizations in the ballpark?

    Poništi
  12. proslijedio/la je Tweet

    I have a favor to ask. For 20+ years I've been working on a dream: to make all science funded by US taxpayers freely available to all. We are on the verge of achieving this. But we need to show that people care. So please, if you can, sign this letter:

    Poništi
  13. 23. sij

    I'm on board with being honest with ourselves, but the criticism of offsets and the like is unwarranted. if you want to maximize your positive impact on the environment, giving to climate-focused charities is a hell of a lot more effective than skipping a few conferences.

    Poništi
  14. 22. sij

    but these are very minor quibbles. overall, I thought both the content and tone of the discussion are very faithful to the paper (and my goals in writing it).

    Prikaži ovu nit
    Poništi
  15. 22. sij

    (3) I'm not critical at all of large, multi-site consortia; I think they're arguably the best way to do psychology. what I think is largely a waste of everyone's time is doing the _same_ narrow experimental procedure dozens of times over while varying only the research site.

    Prikaži ovu nit
    Poništi
  16. 22. sij

    (2) I think "conceptual replications" (as classically conceived) are a travesty, and am certainly not advocating for more of them (what I advocate for is building more _variability_ into studies). but I think perfectly channels my response here.

    Prikaži ovu nit
    Poništi
  17. 22. sij

    very minor quibbles: (1) I think the ~5-minute discussion on induction is a red herring; no scientific approach (including the so-called "deductive" method) can solve THE problem of induction. my paper isn't about that. it's about ensuring that we use words in a sensible way.

    Prikaži ovu nit
    Poništi
  18. 22. sij

    excellent, enjoyable discussion of my recent preprint on generalization in psychology (). if you don't want to read 20 dense pages, you could listen to 40 minutes of this podcast and come away with a pretty accurate summary. thanks , !

    Prikaži ovu nit
    Poništi
  19. 22. sij

    is it too late for Schiff to be the Democratic nominee?

    Poništi
  20. 21. sij

    I don't think I'll be winning many elections in 2020

    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:

    ·