Hacker News Reader

haellsighon Blender's UI and Runtime: Much Faster Than GPUI

So, about a week's worth of vibe coding.

haellsighon AI models may be capable of self-replicating without access to a datacenter

*only against open weight models

haellsighon Bitwarden Dual License Model

You use Keyguard (https://github.com/AChep/keyguard-app).

haellsighon Run Qwen 3.8 Flash Next (125B) on consumer hardware (RTX 4090) at 100T/s

It looks like he uses oh-my-pi, which is a full featured version of pi and has this to-do style out of the box.

haellsighon Topcoat is pushing the boundary of server applications with Rust

I have tried it. It feels pretty good to use, but you have to keep in mind that it's very early software. They currently break the API every other week, and it's still missing a lot of fundamental features. I have made a few basic web apps with it, and on v0.8.1 it was still missing a lot of reactivity which meant you had to use Javascript to bridge the gap.

haellsighon Grok 4.7

Huh, I've had a totally different experience. I've used it extensively, maxing out the 200€ plan on personal projects and it's the best model I've ever used, so easy and pleasant to use. It's great for frontend design and using it in Rust I've had Coming from Opus 5, it's a breath of fresh air.

haellsighon Grok 4.7

I remember the Gemini 2.5 Pro series was the best at text, but failed half of it's tool calls. They really redeemed themselves with the 3.5 Flash series, it's great. I love how fast it is. Now if only they could offer a subscription service for it...

haellsighon Ask HN: How to recover Google auth after phone stolen?

I've had this issue before, and the only thing that saved me was SMS (phone number) backup. I believe Google is slightly easier than Meta to contact by phone so this should help.

haellsighon Oomwoo, an open-source robot vacuum you build yourself

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

haellsighon Fable 5 Is Back

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

haellsighon Apple Foundation Models

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.

haellsighon Chipotlai Max

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.

haellsighon Ternary Bonsai: Top Intelligence at 1.58 Bits

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.

haellsighon Claude Opus 4.7

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

haellsighon Claude Opus 4.7

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...