Live data from Hacker News

Viewing profile — dinp

dinp

HN member
Joined
Tue, Apr 02, 2019, 2:10 PM UTC
HN karma
142
Public activity
26 items

About dinp

No profile information was provided.

Recent public activity

  1. comment
    Comment #49045940

    The last time I checked, for the arc agi 3 leaderboard, the models are given a simple prompt and the game input and asked to play the game, no harness/tools. If harnesses were allo…

  2. comment
    Comment #47450862

    > If this benchmark becomes popular, then presumably to avoid such embarrassments synthetic data is eventually added to training sets to make sure even esolangs are somewhat more i…

  3. comment
    Comment #47083388

    Zooming out a little, all the ai companies invested a lot of resources into safety research and guardrails, but none of that prevented a "straightforward" misalignment. I'm not sur…

  4. comment
    Comment #46983218

    I thought I was the only person going crazy by the new default behavior not showing the file names! Please don't expect users to understand your product details and config options …

  5. comment
  6. comment
    Comment #43076425

    Source code: https://github.com/don-dp/simulateagents/ Click on 'Play moves' to watch a replay. I initially planned to run a chess tournament for LLMs but they are not good: beside…

  7. story
  8. comment
    Comment #43075519

    The article mentions, he is going to run the marathon, looking forward to what he can do in that distance. I feel it's only a matter of time until someone breaks the 2 hour barrier…

  9. comment
    Comment #43044015

    I think political news are not encouraged here, exceptions for when interesting discussions are possible. Judging by the quality of comments here and in the linked submissions, it'…

  10. comment
    Comment #42719921

    The impact this organization had was incredible. I doubt they would have been able to do this work if they were based out of any other country, which makes me wonder how the US leg…

  11. comment
    Comment #42471166

    Great work! When I use models like o1, they work better than sonnet and 4o for tasks that require some thinking but the output is often very verbose. Is it possible to get the best…

  12. comment
  13. comment
    Comment #40353814

    I don't understand their api not being intended for individual use [0], are developers supposed to use this subscription only? The haiku model is pretty good for the price + availa…

  14. comment
    Comment #39840839

    The idea of system 1 and system 2 had a profound impact on me. While specific conclusions in the book were reported to be based on low quality data, it doesn't take away from the f…

  15. comment
    Comment #39740402

    Somehow the idea of perpetually paying property taxes and land value taxes doesn't sound appealing to me, especially since businesses already pay taxes. I don't understand the argu…

  16. comment
    Comment #38967473

    https://www.cloudflare.com/en-gb/plans/ Looks like level 3 ddos protection is only available on the enterprise plan, it's not included in the unmetered ddos protection.

  17. comment
    Comment #38022422

    With gpt-4 default, it's January 2022. With gpt-4 + bing it's September 2021. Strange..

  18. comment
    Comment #38022364

    Cut off is still January 2022 for me, maybe it's being rolled out. Finally developers' AI generated code won't be stuck using January 2022 versions. I wonder if they use gpt-4 itse…

  19. comment
    Comment #37986608

    In case the founders read this post: what would you do differently if you could start over and would this idea work today? Slightly OT: everyone feels hiring is broken, can you lis…

  20. comment
  21. comment
    Comment #37157412

    Before the function calling update[0] it was possible for the gpt models to use tools using specific prompts but it was unreliable. Have you tried using function calling to let the…

  22. comment
    Comment #37156996

    https://apiforllm.com/ I've been working on this on and off for a few weeks now. The idea is to let LLMs interact with the outside world using function calling. Eg- I built a simpl…

  23. comment
  24. comment
    Comment #36733889

    You can add reviews under the chrome and firefox extensions to warn other users and then report both extensions (assuming you are confident about your findings). More of a meta com…

  25. comment
    Comment #36655491

    > On the other hand, my first self-funded startup got destroyed by a VC funded venture. They had a worse product but far better marketing and they used every dirty trick in book to…