Live data from Hacker News

Viewing profile — ninjahawk1

ninjahawk1

HN member
Joined
Sat, Mar 28, 2026, 4:59 PM UTC
HN karma
445
Public activity
130 items

About ninjahawk1

https://nathanlangley.dev/

Recent public activity

  1. comment
    Comment #49165318

    Current highest ranking model on the GPQA-Dumb benchmark, where the lower the score the higher the score. I encourage you to take a look at the benchmarks, boasting as low a score …

  2. story
  3. comment
    Comment #49164544

    This is true for output but as well for learning, if you speak to an LLM trying to get it to give you a certain answer, it’ll find a way to tell you you’re right. If you’re truth s…

  4. comment
    Comment #49146788

    When I was in high school at a private school, I was surrounded by students who were religious but didn’t have much curiosity. I got into a debate with basically my whole class abo…

  5. comment
    Comment #49136382

    seeing an emoji in a title on HN really surprised me, I didn’t even know you could do that. Shows how much emotion we use on here lol.

  6. story
  7. comment
    Comment #49113415

    80% less for Luna is absolutely crazy, in my opinion we may reach a point in the next year where powerful models on the API could potentially be cheaper than subscriptions. Compute…

  8. comment
    Comment #49105830

    Openclaw is great from when I’d used it for a couple months, the key difference is that for Openclaw you’ll have to actually schedule the cron jobs or automated tasks yourself, so …

  9. comment
    Comment #49105762

    Good point, another example would be for when I was training my own small LM a while back, the target was about 170M parameters and was trained on 2B tokens worth of movie subtitle…

  10. story
  11. story
  12. comment
    Comment #49105500

    Exactly what my thoughts were when I first heard about Openclaw, that’s the exact idea of Orb that you pointed out. Letting you make less decisions, right now AI gives you answers …

  13. comment
    Comment #49105405

    Happily, current AI is interacted with in a reactive loop. I open the Claude app, CLI, whatever, say my prompt, get an output. I personally wanted an AI that was able to reach out …

  14. comment
    Comment #49104732

    I can’t speak for other startups, but I applied to the most recent YC batch with my idea for making AI proactive instead of reactive, and pre-being selected I’ve published a paper …

  15. comment
    Comment #49089433

    If you can generate traffic to your website while having a deep email list as well of people who are alerted when you write, you essentially make yourself unable to be censored or …

  16. story
  17. story
  18. comment
    Comment #49079521

    He could’ve just signed the letter. We all know that this isn’t some higher ground stance, they’re anti-competitive since they’re currently #1. I’ve been seeing this for a while. T…

  19. story
  20. comment
    Comment #49064908

    Them: you can’t see what’s on the inside of a black hole Me: zooms in

  21. story
  22. comment
    Comment #49049525

    Very nice. Now let’s see Paul Allen’s buttons

  23. comment
    Comment #49036734

    Sounds like the answer may be simpler than it seems. You could just be tired. You said you’ve been balancing a lot for years and years. Take some time to just focus on yourself. Me…

  24. comment
    Comment #49027755

    I’m very confused on what this is, my initial thought was “oh nice, open source router.” I go to the website…and it’s a sign up. I expected a repo. Otherwise how do I use it? As a …

  25. comment
    Comment #49023429

    The U.S. does NOT have copyright protections on the data that AI models were and are trained on, it was mass theft. Now all of a sudden, the Chinese are stealing from the thieves a…