Live data from Hacker News

Viewing profile — obscur

obscur

HN member
Joined
Thu, Mar 24, 2022, 8:09 PM UTC
HN karma
21
Public activity
28 items

About obscur

No profile information was provided.

Recent public activity

  1. comment
    Comment #38206282

    GPT4 is multimodal in the sense that it can take images as input. The person is using a speech to text system such as OpenAIs Whisper and serving screenshots and voice transcripts …

  2. comment
    Comment #38181147

    I don’t know how one would settle the question but I’d venture to say that Theravada is not less close to the Western impression of Buddhism than other parts of Buddhism. Additiona…

  3. comment
    Comment #37829529

    Effective altruists are most commonly donating to specific relatively high-evidence causes in low-income countries or much more speculative existential risk causes. Donations such …

  4. comment
    Comment #37799493

    It uses LoRA which is much smaller than full model

  5. comment
    Comment #37642460

    Depends on what you mean by works out. They are planning on doing huge training runs the coming years so they sure can make use of compute from both Google Cloud and AWS

  6. comment
    Comment #37419242

    How is this related to the topic?

  7. comment
    Comment #37335941

    https://chat.openai.com/share/d9d64cf5-3ad5-4491-97a1-43fa01...

  8. comment
  9. comment
  10. comment
    Comment #37088878

    He’s explicitly saying that it is not embarrassing in this sense and that it is rather that the discrepancy is ‘awkward’

  11. comment
    Comment #36976308

    Seems cool! I know basic Spanish from Duolingo but not enough to understand what it says to after a few rounds so my suggestion is that there is a side-by-side translated version o…

  12. comment
    Comment #35667291

    I assume he means that models will be fine-tuned (or otherwise steered) by developers to have submissive behavior.

  13. comment
    Comment #35667279

    Sounds like that to me too.

  14. comment
    Comment #35407740

    Yes, see also the transformers library containing many transformer models in pytorch

  15. comment
    Comment #35398242

    This is intended to be a way to fairly distribute UBI, so a fee alone does not work in that case.

  16. comment
    Comment #35216929

    Your first sentence is interesting, why do you think that?

  17. comment
    Comment #35199740

    While I agree it is worth noting that this was already tested in some kind of sandbox environment that only ‘simulated’ replication.

  18. comment
    Comment #34195693

    What are some things you have met where it seemed outright evil? (I don’t have any work experience yet)

  19. comment
    Comment #33212802

    I would like a source for the first claim. When I read the preprint for this paper half a year ago it was actually one of their claims that to the best of their knowledge their Pon…

  20. comment
    Comment #32929335

    Measuring performance in rounds of successful Chinese whisper (irony)

  21. comment
    Comment #32541586

    Surely this cannot be an argument for whether something is a legitimate field. Surely there are some scientists somewhere studying the risks of a very risky viruses. At least prima…

  22. comment
    Comment #32540225

    I think I heard of people with brain damage who does not keep an internal state, they feel like they continuously wake up and only remember memories from before the brain damage. T…

  23. comment
    Comment #32437239

    Interesting perspective, where did you read about this?

  24. comment
    Comment #32189555

    I heard from a chemistry researcher that it is also common in that field to give vague descriptions of synthesis procedures in order to allow authors to churn out more papers.

  25. comment
    Comment #31012838

    This is a valid problem and my impression is that researchers are aware of this problem but are just working with what they have got. https://twitter.com/RCarhartHarris/status/1513…