Live data from Hacker News

Viewing profile — throw7

throw7

HN member
Joined
Tue, Apr 23, 2013, 1:57 PM UTC
HN karma
2,756
Public activity
1,205 items

About throw7

No profile information was provided.

Recent public activity

  1. comment
    Comment #49139863

    "...the model may give different advice even when the underlying question is the same." Isn't this the point of LLMs? If not it would be deterministic and that's not "new" and/or "…

  2. comment
    Comment #49014310

    It seems a lot of people that read about their bio on wikipedia feel the same way. In fact, it's hard not to be suspicious of those that don't.

  3. comment
    Comment #49008682

    I heard about one in northern virginia which seems to be the front lines in this debate. It has onsite power generation which seems to be going 24/7 which is positively awful if yo…

  4. comment
    Comment #49008401

    Labels don't help? The Smithsonian begs to differ.

  5. comment
    Comment #49006320

    old reddit is broken... the message/mail notifications don't work/go away.

  6. comment
    Comment #49000787

    Utah was the one I remember... have more states passed laws? https://www.eff.org/deeplinks/2026/04/utahs-new-law-regulati...

  7. comment
    Comment #48845760

    This isn't hacker news... but whatever i guess.

  8. comment
    Comment #48819125

    I just go under and over one more time in the initial step of tying a traditional "shoelace" knot. That gives a little more friction on the initial tightening draw to keep it in pl…

  9. comment
  10. comment
    Comment #48662807

    Just today a website presented me a qrcode captcha. I threw up.

  11. comment
    Comment #48654505

    It's not if nothing works, it's what works when nothing works, i.e. the safe thing must always work. period. (i'm not trying to be pedantic or an ass)

  12. comment
  13. comment
    Comment #48576038

    Just tried it. Still broken. Maps doesn't respond at all now. I get the beep that maps hears my "hey google", but that's it now. > I see. So you intentionally broke the feature, no…

  14. comment
    Comment #48570204

    They've regressed for a long time and there's no signal to consumers that "AI" is anything that "fixes" or brings back what was working. The example I always give is when google ma…

  15. story
  16. comment
    Comment #48424815

    I've always favoured Tessier-Ashpool S.A. as comparable. The creation of AI's and Freeside and Musk's pronatalism slots in nicely.

  17. comment
    Comment #48419822

    Trust is slowly gained and easily lost. The amount of apologia I hear from top-tier developers signals an inflection point downward.

  18. comment
    Comment #48388704

    this is the internet sir.

  19. comment
    Comment #48383977

    Value to Anthropic's shareholders.

  20. comment
    Comment #48379101

    Amazon? I'm sure Macgyver would've had to be a woman. And make her gay.

  21. comment
    Comment #48378449

    Oh, a "Human-Cented" study by AI lover: Julian Nyarko Professor of Law Co-Chair Stanford Law AI Initiative Senior Fellow, Stanford Institute for Human-Cented AI (HAI) LOL!

  22. comment
    Comment #48359123

    [flagged]

  23. comment
  24. comment
    Comment #48303853

    Ehh, more like Rush would've been found dead like Abbott after declaring "I'm a patriot" to internal CIA. What's tantalizing about Bourne is something about who we are and capable …

  25. comment
    Comment #48283148

    Without the badging I wouldn't have guessed Ferrari.