Philpax

35 ms·

1. 721.▲
   by Philpax 2y ago

   Apple usually have iPhones available to purchase within two weeks of announcement or less; it's due to their famous logistics pipeline.
2. 722.▲
   by Philpax 2y ago

   Erlang / BEAM would seem to fit, but I haven't seen much disdain for it from HN posters; quite the opposite, actually.
3. 723.▲
   by Philpax 2y ago

   It is deeply funny that's the case - it happened with Grok-2 as well - but I can't imagine that remaining the case when they (ostensibly) scale to superintelligence. After all, it would be unwise to build a superintelligence that
4. 724.▲
   by Philpax 2y ago

   It's complicated by the ComfyUI data model, which treats strings as immediate values/constants and not variables in their own right. This could ostensibly be fixed/worked around, but I imagine that it would come at a cost to
5. 725.▲
   by Philpax 2y ago

   My Prime subscription lapsed due to my card expiring before updating it, and I can't say I've been missing it much. I miss some content from Prime Video, sure, but knowing I'm not supporting the Bezoian empire as a consumer m
6. 726.▲
   by Philpax 2y ago

   Unfortunately, those humans are developing that super-intelligence. Are you ready to submit to Elon Musk's Grok ASI?
7. 727.▲
   by Philpax 2y ago

   I know the r-word is coming back in vogue, but it was still unpleasant to see it in the middle of an otherwise technical blog post. Ah well. Diffusion LMs are interesting and I'm looking forward to seeing how they develop, but from pla
8. 728.▲
   by Philpax 2y ago

   This is a terribly callous and unproductive response that discourages intellectual development for people who aren't at your level. We all have to start somewhere, and this could very well help someone enter the Julia ecosystem.
9. 729.▲
   by Philpax 2y ago

   This may be useful: https://www.interconnects.ai/p/deepseek-r1-recipe-for-o1 but the tl;dr of the idea is that we can use reinforcement learning on a strong base model (i.e. one that hasn't been fine tuned) to eli
10. 730.▲
    Anthropic's Recommendations to OSTP for the U.S. AI Action Plan(anthropic.com)

    4 points by Philpax 2y ago|0 comments
11. 731.▲
    by Philpax 2y ago

    to be fair, I don't think vibe coding is "precision software" - as described, it's much more slapdash than that
12. 732.▲
    by Philpax 2y ago

    Keen is a fully remote outfit, so he can work wherever. It's pretty likely that his reputation would open that door for him no matter where he goes.
13. 733.▲
    Vibe Coding(en.wikipedia.org)

    12 points by Philpax 2y ago|6 comments
14. 734.▲
    by Philpax 2y ago

    I would encourage you to try out a recent reasoning model like Claude 3.7 Sonnet / o3-mini / R1. No, it's not perfect, but it can very much architect and design things at a higher level than you might think. This is already t
15. 735.▲
    by Philpax 2y ago

    oh come on, it made me chuckle
16. 736.▲
    by Philpax 2y ago

    There's no reason AI couldn't do that too. It will continue moving up the cognitive ladder: there's nothing special about our ability to ask those questions. For example, try asking Claude "what decisions do I have to ma
17. 737.▲
    Anthropic raises Series E at $61.5B post-money valuation(anthropic.com)

    18 points by Philpax 2y ago|0 comments
18. 738.▲
    by Philpax 2y ago

    https://tilde.team/~ben/suckmore/
19. 739.▲
    by Philpax 2y ago

    serde doesn't support CBOR/MP, implementations of those support serde, and those implementations are listed in the table. You might have a point about JfMC++, though.
20. 740.▲
    by Philpax 2y ago

    I don't think weatherproofing them is as much as a lift as making them work autonomously: once you have a proof of concept, it's much easier to evolve it to whatever spec you need. For replication, probably not by themselves, but
21. 741.▲
    by Philpax 2y ago

    https://en.wikipedia.org/wiki/Social_murder
22. 742.▲
    by Philpax 2y ago

    Actually, fair enough. I still disagree with their argument, but this was the wrong tack for me to use.
23. 743.▲
    by Philpax 2y ago

    I used NVIDIA on sway/hyprland on Wayland on NixOS with a variety of driver versions last year and still had awful flickering that made it basically unusable. After much consternation, I fixed it by buying an AMD card, and it's wo
24. 744.▲
    by Philpax 2y ago

    Under the dingnuts regime, Dwarf Fortress will be illegal. Actually, any game with a procedural story? You better believe that's a crime: we can't have a machine generate text a human might enjoy.
25. 745.▲
    by Philpax 2y ago

    Humanoid robots are coming: - Figure: https://youtu.be/Z3yQHYNXPws - Unitree: https://youtu.be/iULi4-qz22I - 1X: https://youtu.be/uVcBa6NXAbk and those are just the ones I know about. I&#x2
26. 746.▲
    by Philpax 2y ago

    > But would you have enough to buy your favorite society and run it into the ground? If you pull it off right, you can get a 2-for-1 deal on the social network and parlay the result of your ketamine bender into a role as Shadow President
27. 747.▲
    by Philpax 2y ago

    And what led him to take his own life?
28. 748.▲
    by Philpax 2y ago

    Just as long as you don't buy NVIDIA for your Linux desktop!
29. 749.▲
    by Philpax 2y ago

    Ladybird and Servo are coming up, but I suppose you're covering that with "working".
30. 750.▲
    by Philpax 2y ago

    If they reach their goal, they won't need a lower class, and they can address that problem however they see fit. A soldier might refuse to fire on their own side, but an autonomous drone won't.