pmarreck

25 ms·

1. 601.▲
   by pmarreck 1y ago

   Well, that's probably part of the issue- I seem unable to have a good cry, and haven't had one in years frankly, even when my mom died
2. 602.▲
   by pmarreck 1y ago

   I probably do need a good cry since my breath does that repeated-hesitation thing when I breathe in sometimes (not sure if there’s a name for it yet, but you also do it for a while after you’re done crying, I call it “aftershocks” then… but
3. 603.▲
   by pmarreck 1y ago

   ah, that explains why all the GLM quants are MLX models
4. 604.▲
   by pmarreck 1y ago

   Once the market moves on, only the people that love the media will be the best caretakers of it and that often ends up looking like piracy. They don't rob media sales; they secure media legacies. They're hoarders, not consumers.
5. 605.▲
   by pmarreck 1y ago

   Fast is why, after decades doing high-level scripting, I'm now exploring lower-level languages that live closer to the metal...
6. 606.▲
   by pmarreck 1y ago

   > refuses to support MCP Why is that? Seems the way to go to add tooling to any LLM that is tool-capable
7. 607.▲
   by pmarreck 1y ago

   https://boltai.com/
8. 608.▲
   by pmarreck 1y ago

   If you have a Mac, I recommend https://boltai.com/ for that
9. 609.▲
   by pmarreck 1y ago

   do you know why ollama hasn't updated its models in over a month while many fantastic models have been released in that time, most recently GLM 4.5? It's forcing me to use LM Studio which I for whatever reason absolutely do not pr
10. 610.▲
    by pmarreck 1y ago

    [flagged]
11. 611.▲
    by pmarreck 1y ago

    Assuming they keep improving, yes. The one I tried (I responded to the other reply with some output, which was great) is as fast as the cloud ones and nearly as good.
12. 612.▲
    by pmarreck 1y ago

    Pretty impressive. Spit out perfectly-working Asteroids on the first try. https://gist.github.com/pmarreck/db782fdb68053292ca746d6c756... I want to hook it up to Zed next and see how that goes
13. 613.▲
    by pmarreck 1y ago

    I have an M4 Mac with 128GB RAM and I'm currently downloading GLM-4.5-Air-q5-hi-mlx via LM Studio (80GB) and will report back!
14. 614.▲
    by pmarreck 1y ago

    I'll have to have a look at Lean 4 then. The simulation hypothesis has a flaw IMHO- If it is modelable and therefore computable, it may be subject to the halting problem
15. 615.▲
    by pmarreck 1y ago

    Have you looked at Idris 2? I'm 53, impressed that you're still going at it at 69!
16. 616.▲
    by pmarreck 1y ago

    Do most people tend to conclude that concatenative languages are absolutely an intellectual curiosity for any nerd who knows RPN and thought "what if we kept going?", but kind of fail in larger-scale production contexts? (Yes, I&#
17. 617.▲
    by pmarreck 1y ago

    > I’m not aware of any research that backs up this claim These things are literally days old . Claude 4, which is arguably the first truly useful LLM code assistant, released in May. Scientific research proceeds at a glacial pace. I&#x2
18. 618.▲
    by pmarreck 1y ago

    I've found (specifically using Idris 2, but any functional language with some typing should work) LLM's to write better code in these restrictive languages simply because there are significantly fewer ways to screw up that are pos
19. 619.▲
    by pmarreck 1y ago

    Ah. Looks like I'll have to get the Ash book...
20. 620.▲
    by pmarreck 1y ago

    It doesn't prevent the error, but it also won't take down your server when malicious users (or just lots of normal users) start to bang on that input with the problem, and your non-BEAM VM pool starts to run out of available prelo
21. 621.▲
    by pmarreck 1y ago

    LOL. This only makes sense if you can know all the ways your code will fail... which you cannot. Erlang/Elixir's approach is to simply say, "It's gonna fail no matter how many precautions we take, so let's optimize
22. 622.▲
    by pmarreck 1y ago

    So if you're comfortable with Ecto and SQL, you can safely skip over Ash?
23. 623.▲
    by pmarreck 1y ago

    Except that it's not like Rails in the sense that in Rails, you will eventually develop extremely difficult to debug race conditions thanks to mutability that Elixir lacks. Source: Has worked on million-line Ruby on Rails codebase
24. 624.▲
    by pmarreck 1y ago

    It's in line with the erlang philosophy of letting things crash trivially and restart instantly. Due to the universal immutability, starting a new process from a given state in Erlang/Elixir is nearly instantaneous and has extreme
25. 625.▲
    by pmarreck 1y ago

    Zed's still pretty beta on linux. It will improve. I'm glad they are at least targeting it as well.
26. 626.▲
    by pmarreck 1y ago

    It just occurred to me that if HN supported, say, 4-bit mono PNG images with transparency in comments, that would help here without impacting bandwidth too much and might add a classy element
27. 627.▲
    by pmarreck 1y ago

    My current favorite code font is Berkeley Mono https://usgraphics.com/products/berkeley-mono It's not free, but I love it. You can customize some variations too (like how zeroes look; I use the "invisible sla
28. 628.▲
    by pmarreck 1y ago

    This is pretty easy to do if you have Nix https://github.com/pmarreck/yt-transcriber A commandline youtube transcription tool I built that runs on Mac (nix-darwin) which automatically pulls in all the dependencies it n
29. 629.▲
    by pmarreck 1y ago

    TIL about Lehmer codes... and "poker encoding" ;) (I just prefer poker to solitaire...) Someone else mentioned that the orientation of the cards (up or down) and possibly even the front-back facingness of the card (facing up, faci
30. 630.▲
    by pmarreck 1y ago

    https://deckcrypt.github.io/ 45 characters according to the blog post and this demo