Live data from Hacker News

Viewing profile — sdrg822

sdrg822

HN member
Joined
Mon, May 06, 2019, 10:34 PM UTC
HN karma
83
Public activity
30 items

About sdrg822

No profile information was provided.

Recent public activity

  1. comment
    Comment #43789119

    “Model Control Plane” - quite early to already re-use a popular acronym

  2. comment
    Comment #43460987

    Congrats on the launch!

  3. comment
    Comment #41286247

    +1 it wouldn’t be terribly useful if it were only caching the tokenizer output.

  4. comment
    Comment #41219526

    Where is the lie in this repo?

  5. comment
    Comment #41206516

    I think if you actually meet the people you'd realize they're pretty earnest and candid of these things' limitations, though it may not show in tweets and hackernews posts. This re…

  6. comment
    Comment #40303573

    But indexing *is* training. It's just not using end-to-end gradient descent.

  7. comment
    Comment #39693188

    Dang they use Transformer-XL from 2019 haha - didn't realize people still used that / XLNet-like architectures

  8. comment
    Comment #39518054

    It is not.

  9. comment
    Comment #39469959

    Congrats on the launch!

  10. comment
    Comment #38378424

    Cool. Important note: """ One may ask whether the conditionality introduced by the use of CMM does not make FFFs incompatible with the processes and hardware already in place for d…

  11. comment
    Comment #35174059

    It’s only a matter of time

  12. comment
    Comment #34558537

    Given that GPTZero was an undergrad’s side project it’s not that surprising?

  13. comment
    Comment #34448180

    Attention is Turing complete

  14. comment
    Comment #34391718

    For things like BERT where you just want to extract an embedding, the naive way you reach full utilization at inference time is that you : - run tokenization of inputs on CPU - sor…

  15. comment
    Comment #34112710

    Yeah this is purely a risk decision for a prototype not an actual technical limitation.

  16. comment
  17. comment
    Comment #33416275

    Really well done

  18. comment
    Comment #33416268

    Which humans?

  19. comment
    Comment #28213108

    Good points!

  20. comment
    Comment #28212801

    Just a slight comment - "&" was indeed called "and" (not "per se"), but in reading the alphabet, it was confusing to say "and and." To clear up confusion, one could say "and per se…

  21. comment
    Comment #28170959

    Seasoning isn't the problem... Turmeric and cumin added to your pulses aren't going to give you metabolic diseases...

  22. comment
    Comment #26032436

    Agreed, the "but" in the bolded text[1] further confuses, but the article highlights some great progress. >> STAR, has responded to 748 incidents." >> about 3 percent of calls for …

  23. comment
    Comment #24744009

    > "While it remains to be seen whether transplantation from very young donors can restore cognitive function in aged recipients, the findings demonstrate that age-related shifts in…

  24. comment
    Comment #24657224

    Automated summarization, content moderation/censorship, and advertising. It could make publishing more profitable, but will it make enable higher quality content? It feels more lik…

  25. comment
    Comment #24572251

    Abbreviation for a popular blogger in the financial independence / retire early domain. See: https://www.mrmoneymustache.com