busfahrer

6 ms·

1.
by busfahrer 25d ago
Can confirm, it's very much Hiro Protagonist. I love the style, and the over-the-topness of it all made more sense once I learned that it was originally planned as a graphic novel, IIRC. Gibson's style is much more post-modern.
2.
by busfahrer 25d ago
> The future is here, it's just not evenly distributed. One of my favourite Gibson quotes.
3.
by busfahrer 1mo ago
I am eyeing one of these specifically for this use case, could you please post roughly what kind of tokens per second numbers you get for text generation for this 27B model? edit: and which quant you are using, please :-)
4.
by busfahrer 1mo ago
Fun fact: The pattern of the torches' flickering is encoded via special strings in the source code: https://jasemagee.com/post/2022-03-27-26-year-old-quake-ligh...
5.
by busfahrer 2mo ago
For anyone who enjoys this sort of thing, I found this postmortem of an Alone in the Dark clone for PICO-8 highly informative: https://www.lexaloffle.com/bbs/?tid=3804
6.
by busfahrer 2mo ago
has anyone been able to get bonsai models to run at a proper speed in a x86 windows cpu-only setting, using llama.cpp?
7.
by busfahrer 2mo ago
For anybody else trying to grok backprop, I found it very helpful to look at karpathy's microgpt, a barebones LLM in 200 lines of python. The autograd itself is just 40 lines, and reading that code and then trying to do a clean room im
8.
by busfahrer 2mo ago
As a kid I played Ken's Labyrinth, a raycaster FPS from around the same time as Wolf3D. Only years later I found out that this was actually the same Ken Silverman that went on to create the Build engine for Duke Nukem 3D etc. Kind of
9.
by busfahrer 2mo ago
I'm on the fence because I have a TSC-X controller and it's unclear if it's supported. Somebody on the forums posted a tool that converts generic joystick axes to keypresses, but not sure how well that works.
10.
by busfahrer 2mo ago
> Or maybe Lisp is so pure that it embodies both Light Side and Dark Side, like a god that spawned the programming universe. This isn't so far off, considering that some people consider the "Lisp in Lisp" bit from the 1.5
11.
by busfahrer 4mo ago
The PS1 architecture is fascinating - it will also reveal why PS1 games have such a unique, recognizable style that games are trying to recreate nowadays.
12.
by busfahrer 4mo ago
To add some context, the author of this article is a science fiction writer, one of his stories was adapted into the movie "Arrival".
13.
by busfahrer 4mo ago
Not a one-liner, but I was impressed by this simple rogue-like written in 10 lines of BASIC: https://bunsen.itch.io/the-snake-temple-by-rax
14.
by busfahrer 4mo ago
Having only dabbled in Lisp and knowing nothing about the plugins TFA mentions, there's also this FWIW: https://github.com/jpalardy/vim-slime
15.
by busfahrer 4mo ago
Thank you, good sir!
16.
by busfahrer 4mo ago
Great, thanks! :-) and to mirror another poster: what kind of prompt parsing (prefill) speed do you get for that model? Also how is the speed for the 27B model?
17.
by busfahrer 4mo ago
I have been contemplating a M5 Pro MBP, but for the life for me I wasn't able to find benchmarks for real-world models, do you happen to know how many tokens per second roughly you get with MoE models like Qwen 3.6 35B/A3B or Gemm
18.
by busfahrer 4mo ago
Seeing as this is HN, I was expecting something on chemical properties of iron etc, but was pleasantly surprised
19.
by busfahrer 4mo ago
OTOH, with the IMSAI-8080 you'd have the satisfaction of using the same machine as David Lightman of Wargames fame
20.
by busfahrer 4mo ago
To add some context, this was made for the PICO-8 fantasy console, which (by design) is very limited, so it's immensely impressive to see something like this working on it. Honestly, even just the 3D logo from the intro is impressive,
21.
by busfahrer 4mo ago
So I'm guessing it's sort of like Fennel, which consciously doesn't completely hide its underlying Lua runtime? https://fennel-lang.org/
22.
by busfahrer 4mo ago
I am considering a M5 Pro (18/20C) Macbook with 64GB of RAM, but I'm having a really hard time finding benchmarks of real world models: Could somebody please provide some tokens-per-second numbers for example for Qwen 3.6 35B&#x2F
23.
by busfahrer 5mo ago
This is the whole point of Karpathy's nanochat which OP refers to, to train a GPT-2 level LLM for under $100, renting an 8xH100 VM.
24.
by busfahrer 5mo ago
This kind of is Postel's law, in a way: https://en.wikipedia.org/wiki/Robustness_principle
25.
by busfahrer 5mo ago
Out of curiosity, what size Qwen did you use, at what quantization?
26.
by busfahrer 5mo ago
I think I remember a Carmack tweet where he mentioned in most cases he only considers it once he reaches three duplicates
27.
by busfahrer 5mo ago
This seems like a nice context to mention Sturgeon's law: > ninety percent of everything is crud https://en.wikipedia.org/wiki/Sturgeon%27s_law
28.
by busfahrer 5mo ago
> Low cost specialized models Can you elaborate on this? Is this something that companies would train themselves?
29.
by busfahrer 5mo ago
I've noticed that subconsciously, I've started to use `bat` for interactive paging, even though that's not its main use case
30.
by busfahrer 5mo ago
Once you get your head around it a bit, doing a new in this circumstance will be second nature, since you will have realized that a `new XYZ` in jj leads to the same underlying git state as a `git checkout XYZ` in git

More ›