Live data from Hacker News

Viewing profile — hobom

hobom

HN member
Joined
Fri, Oct 30, 2020, 10:04 PM UTC
HN karma
155
Public activity
40 items

About hobom

No profile information was provided.

Recent public activity

  1. comment
    Comment #49086904

    You totally ignored the point OP was making! Even if it is a parenting issue, simply declaring that will not make parents parent better. It's the same reason why we ban children fr…

  2. comment
    Comment #49047095

    Counting cache hits towards the token budget is exactly how it should be done for these kind of evals, and at any rate, for cyber evals all frontier models benefit from more tokens…

  3. comment
    Comment #49002903

    China is 6 months behind because of distillation attacks, otherwise the gap would be larger

  4. comment
    Comment #47847093

    They are taking in more than they are spending hosting them. However, the cost for training the next generation of models is not covered.

  5. comment
    Comment #47383822

    Meta was strongly against the Australian social media ban.

  6. comment
    Comment #47188028

    Sorry, the "they" referred to the hyperscalers

  7. comment
    Comment #47186991

    Even more extreme, that might mean they won't be able to offer Claude to non-US companies at all.

  8. comment
    Comment #47096689

    The article didn't suggest that the video mentioned was AI slop, it correctly recognised it as human generated.

  9. comment
    Comment #46863710

    [flagged]

  10. comment
    Comment #46574752

    So the problem IS that people are putting in too much (even if accidentally) ?

  11. comment
    Comment #46077968

    There are several concrete proposals to regulate AI either proposed or passed. The most recent prominent example of a passed law is California SB53, whose summary you can read here…

  12. comment
    Comment #45849800

    Australia's soon-to-take-effect ban affects nine platforms, including Instagram and Facebook, but not HN. These bans often operate on the amount of users a platform has, so HN is u…

  13. comment
    Comment #45849764

    Jonathan Haidt, the most prominent psychologist pushing for restrictions on social media use for children, is also the most prominent proponent of letting kids play and roam more f…

  14. comment
    Comment #45712245

    What a fantastic way to write a post mortem, pedagogically very useful.

  15. comment
    Comment #44554827

    At least in part, according to the article, because the not-yet-on-GLP1 folks are NOT customers since they are often denied coverage in the first place.

  16. comment
    Comment #42742342

    The West told plenty of its companies, through public pressure or laws, that they have to divest from Russia, and they did. Rationally they recognized that selling their assets is …

  17. comment
    Comment #42301018

    As far as I know that was a distinct, previous endeavor. Only afterwards did the Vinklevoi approach him to make the website, an idea which he then allegedly ran with on his own.

  18. comment
    Comment #42278152

    Does 4Chan also have bot BEHAVIOR detection (e.g. unnatural mouse movements)that google captcha has?

  19. comment
    Comment #41488483

    The article provides almost no hard data to back the headline up. They link a report on buy backs, but at least the abstract/summary does not provide support for the claim either. …

  20. comment
    Comment #40076197

    FWIW, I thought he did pretty well and came away more convinced of his theses. For most of Tyler's alternative and let's say colorful explanations of the data, Haidt provided diffe…

  21. comment
    Comment #39585121

    Like the OP, I also greatly respect the work Rachel and you did for AI. Nevertheless, I can't help but think that you are seeing this issue to negatively. "The point of the article…

  22. comment
    Comment #37192416

    It's basically all present in the article

  23. comment
    Comment #34773288

    While true in general, it is really not that necessary to have a representative sample for checking whether your software works. With large enough sample size a convenience sample …

  24. comment
    Comment #34712440

    Comparing against 1990 is wrong though, as we should compare against the start of widespread social media use (2010 at least), and it's also pretty likely that other OECD countries…

  25. comment
    Comment #34712377

    This is not an accurate reading of the Conclusion section. He just sums up what has been observed and does not mention any causal factors. Earlier, he explicitly says that causal f…