girvo

12 ms·

301.
by girvo 6mo ago
> I don’t really see Claude Design fixing the fundamental Storybook hell of designing reusable and pretty components and layouts FWIW Claude Code is decent at scaffolding those out if you have a good set of examples for it to work from
302.
by girvo 6mo ago
True but Z.ai also just silently raised the price, and the entire Chinese frontier set is having to make profit now... hence Alibaba killing the Lite plan and not letting people sign up to their Pro one either; and why MiniMax has their non
303.
by girvo 6mo ago
I have an Asus GX10 that I run Qwen3.5 122B A10B on, and I use it for coding through the Pi coding agent (and my own); I have to put more work in to ensure that the model verifies what it does, but if you do so its quite capable. It makes
304.
by girvo 6mo ago
Now you’re at the mercy of YouTube etc’s compression, and having to decompress it for a worse output
305.
by girvo 6mo ago
Very well written? It’s a bunch of AI generated stuff around an interesting point. It repeats its points over and over again, meanders. It’s an interesting thesis, it’s not well written or well told
306.
by girvo 6mo ago
It’s absolutely a signal. As is the constant repeating of points. It’s AI slop for sure Which is a shame coz the premise is interesting.
307.
by girvo 6mo ago
With how powerful these companies are and the enormous impact on your real world life them capriciously deciding to cancel your account can have, governments need to step in. Good idea to write to my federal member here to be honest, I thin
308.
by girvo 6mo ago
Antigravity CLI or the Gemini one? When I tried the latter about 2 months ago it was shockingly bad, though I was a free user. I assume its better if you're a paying customer?
309.
by girvo 6mo ago
I mean the big AI companies are totally keen around it, wouldn’t surprise me?
310.
by girvo 6mo ago
Genuine question, why did you stop using it! Edit: ah, scrolled down where you answered, thanks
311.
by girvo 6mo ago
Yeah lack of peripheral drivers upstream for all the little things on the board, plus (AIUI) ARM doesn't have the same self-describing hardware discovery mechanisms that x86 computers have. Basically, standardisation. They're clos
312.
by girvo 6mo ago
My favourite was, Opus 4.6 last night (to be fair peak IST time, late afternoon my time), the first prompt with a small context: jams a copy-pasted function in between a bunch of import statements, doesn't even wire up it's own fu
313.
by girvo 6mo ago
It's being explored right now for speculative decoding in the local-LLM space, which I think is quite interesting as a use-case https://www.emergentmind.com/topics/dflash-block-diffusion-f...
314.
by girvo 6mo ago
Which makes sense: a robot can’t be responsible for anything, a human can be.
315.
by girvo 6mo ago
This entire article reads like AI slop anyway. I also recommend anyone with a GB10 device to go try out the spark-vllm-docker setup, and check the Nvidia GB10 forums for the recently released optimised Qwen 3.5 122B A10B setup: 50tk/s
316.
by girvo 6mo ago
It’s pretty decent for natural language -> query language tasks But also you don’t need SOTA frontier models for that!
317.
by girvo 6mo ago
FWIW I’m in love with my Asus GX10 and have been learning CUDA on it while playing with vllm and such. Qwen3.5 122B A10 at ~50tps is quite neat. But do beware, it’s weird hardware and not really Blackwell. We are only just starting to squ
318.
by girvo 6mo ago
I’d happily use anything else, but it (with CF Tunnels and its DDoS and caching systems) is what lets me self host on my little home server on today’s internet. Would gladly move to some other system (or systems)
319.
by girvo 6mo ago
The Surface RT that I tried absolutely was, and everything else I'm seeing to make sure I'm not misremembering shows that Windows RT was still locked down. So you should expand on what you mean. Windows S Mode shows that Microsoft
320.
by girvo 6mo ago
For sure, and I'm glad they backed off from it. I'm also glad they did it because of how it pushed Valve into making Steam OS so good. But Microsoft really did want to go down the same path, and I do not trust them not to try it
321.
by girvo 6mo ago
Is anyone attempting to implement Oberon on LLVM IR? Sounds like a fun project
322.
by girvo 6mo ago
That’s not for lack of trying, though: remember Windows RT?
323.
by girvo 6mo ago
Yeah I saw. Such a shame, I’m playing with a Q4 version of 3.5 122B-A10B on my Asus GX10, it’s kind of nuts how great a model you can run at home (with limitations of course)
324.
by girvo 6mo ago
Typically by deeply misunderstanding what those refactoring tools actually do
325.
by girvo 6mo ago
I am! GLM-5 (which is all I have access to on it, not the newer GLM-5.1) is slightly better for the coding tasks I'm using them for, in terms of being more accurate slightly more often. Both are very good, and very close to one another
326.
by girvo 6mo ago
GGUFs are out too, well done Unsloth as usual! https://huggingface.co/unsloth/MiniMax-M2.7-GGUF I've been using M2.7 through the Alibaba coding plan for a bit now, and am quite impressed with it's coding abil
327.
by girvo 6mo ago
https://huggingface.co/unsloth/MiniMax-M2.7-GGUF GGUF is out actually I believe!
328.
by girvo 6mo ago
Hah, of course it’s written in Ocaml! As a learning exercise, I wrote my own little SSG in ocaml, and man I forgot how nice a language it is. Tooling is still a bit rough but lots better than it used to be My little blog (in my profile) is
329.
by girvo 6mo ago
I’m aware, that’s exactly what my grandfathers (rest in peace grandpa, I miss you) IE window looked and felt like in the early 2010s!
330.
by girvo 7mo ago
My reasoning is because once upon a time, I was using Macromedia Fireworks, and PNGs gave far far better results than JPGs did at the time, at least in terms of output quality. Nearly certainly because I didn't understand JPG compres

More ›