Live data from Hacker News

Viewing profile — SleekEagle

SleekEagle

HN member
Joined
Tue, Nov 30, 2021, 6:03 PM UTC
HN karma
949
Public activity
211 items

About SleekEagle

No profile information was provided.

Recent public activity

  1. story
  2. story
  3. story
  4. story
  5. story
  6. story
  7. story
  8. story
  9. comment
    Comment #37617295

    > moving products or services right to the pain or tolerance threshold When did we start talking about the airline industry? But seriously, the reason it works for the airline indu…

  10. comment
    Comment #37585788

    This reminds me of the case of the 500 mile email: https://www.ibiblio.org/harris/500milemail.html

  11. comment
    Comment #37422687

    You can mail in cash for ivpn too

  12. comment
    Comment #37408950

    A comprehensive discussion of the topic: https://www.assemblyai.com/blog/emergent-abilities-of-large-...

  13. story
  14. comment
    Comment #37405502

    Fun fact: All of Wikipedia (text only, compressed) is like 20 GB

  15. comment
    Comment #37350989

    > Such as? I'm not specifying them, I'm referencing their existence. You can take them not to exist, but in that case good luck explaining the measurable consistency that every hum…

  16. comment
    Comment #37350821

    I have no idea what "true" means if it does not mean "a quality of a statement as being an accurate description of objective reality". While mathematics itself may be a human inven…

  17. comment
    Comment #37350778

    How is it a leap of faith when mathematical models that describe reality demonstrate continued reliance in making predictions? Does this not demonstrate that there exist mathematic…

  18. comment
    Comment #37350591

    If by "objective truth" we mean the qualities of nature that exist irrespective of any individual's perception, then I think the continued reliance of our scientific knowledge in p…

  19. comment
    Comment #37343408

    Is the fact that these systems cannot prove their own consistency actually a feature of the incompleteness theorem? I thought it effectively boiled down to "you can keep either con…

  20. comment
    Comment #37343382

    Not that you were, but I don't quite understand why people get so caught up on this fact. There are objective facts about the nature of reality, and we are all (or at least compete…

  21. comment
    Comment #37237542

    I'm assuming this is the "Cramer-Rao lower bound" Rao? What an interesting life, I knew nothing about him before this.

  22. story
  23. comment
    Comment #37223791

    One use case is in the training of Diffusion Models. In the original formulation, the likelihood-maximizing objective is recast in terms of KL divergences. This is done because the…

  24. story
  25. comment
    Comment #37147069

    Fantastic article - I would love to see an analysis comparing larger quantized models to smaller unquantized models. e.g. is a 14b quantized model better than a 7b unquantized mode…