Hacker News: haellsigh

New comment by haellsigh in "Oomwoo, an open-source robot vacuum you build yourself"

haellsigh — Thu, 02 Jul 2026 07:10:41 +0000

For that, there's https://valetudo.cloud/ that supports a number of models from different brands.

New comment by haellsigh in "Fable 5 Is Back"

haellsigh — Wed, 01 Jul 2026 21:25:50 +0000

Well, it looks like they just did reset the usage!

New comment by haellsigh in "Apple Foundation Models"

haellsigh — Mon, 15 Jun 2026 11:44:29 +0000

I agree with what you're saying. I have a Claude plan for work and I prefer using Claude more than any other LLM I've tried. Having recently tried the Codex 100€ plan with GPT-5.5 in high/xhigh, I don't think it's worse that the Opus models, just different.

I've noticed that depending on how you talk to it, you get wildly different outputs. This seems to happen less with Opus: it mostly understand what I want. GPT is often a bit too literal.

Just my two cents.

New comment by haellsigh in "Chipotlai Max"

haellsigh — Tue, 02 Jun 2026 10:02:09 +0000

I added it in my oh-my-pi configuration before (it's OpenAI compatible), but Llama 3 8B is just absolutely unusable for anything coding related. It is very fast and the latency is very good however.

New comment by haellsigh in "Ternary Bonsai: Top Intelligence at 1.58 Bits"

haellsigh — Tue, 21 Apr 2026 08:10:37 +0000

Fyi, I believe `--flash-attn on` doesn't do anything, you should instead use `--flash-attn 1`. I'm getting ~150t/s on a RTX 3080 10GB as well with f16 cache type.

New comment by haellsigh in "Claude Opus 4.7"

haellsigh — Thu, 16 Apr 2026 16:34:21 +0000

If that's true, then we're following the timeline of https://ai-2027.com/

New comment by haellsigh in "Claude Opus 4.7"

haellsigh — Thu, 16 Apr 2026 16:14:34 +0000

I got Opus 4.7 working on oh-my-pi with this commit if it interests you: https://github.com/azais-corentin/oh-my-pi/commit/6a74456f0b...