Hacker News: redrove

New comment by redrove in "MicroVMs: Run isolated sandboxes with full lifecycle control"

redrove — Fri, 26 Jun 2026 16:38:41 +0000

No, it doesn’t seem like it.

New comment by redrove in "U.S. government will decide who gets to use GPT-5.6"

redrove — Fri, 26 Jun 2026 12:12:04 +0000

I’m sorry but I can’t take a European Commission link seriously about training SOTA LLMs.

New comment by redrove in "The worthlessness of Vitamin D is mildly exaggerated"

redrove — Wed, 24 Jun 2026 15:15:27 +0000

Do you have any sources for this? Genuinely interested in reading more.

New comment by redrove in "The European Social Stack"

redrove — Sat, 20 Jun 2026 17:01:27 +0000

No it’s not.

You’re generalizing, DACH != the entire EU.

New comment by redrove in "Local Qwen isn't a worse Opus, it's a different tool"

redrove — Thu, 18 Jun 2026 07:50:03 +0000

> My dream would be a local model that can do, say, 80% of the day to day tasks I need; "how does X Handler connect to Y storage?", "commit that feature, but leave out the bits that relate to billing" etc.

Qwen 3.6 27B can do that today, but setup properly and in a good quant, I run an autoround [0] with weights in int8 and attention heads in f16 on a single RTX 6000 Pro Blackwell Max-Q via vllm with mtp=2 and full context, --max-num-seqs 3, KV in f16, mamba f32.

>It would have 99% reliable tool calling

I managed to score 93/100 in tool-eval-bench [1]. For me this is very good already, at least in the pi coding harness I've never had an issue that wasn't auto-fixed in the next turn(s).

>the ability to go "this task is beyond my skills" and refer to a Big Boy Online Model in a gigantic datacenter somewhere

This is heavy on the harness engineering side I think, but also quite contrary to the nature of LLMs today. If you figure this out I'd love to know.

[0] https://huggingface.co/Minachist/Qwen3.6-27B-INT8-AutoRound/...

[1] https://github.com/SeraphimSerapis/tool-eval-bench

New comment by redrove in "GPT‑NL: a sovereign language model for the Netherlands"

redrove — Wed, 17 Jun 2026 06:43:20 +0000

They’re still debating that, they’ll get back to us soon I’m sure.

New comment by redrove in "Home alone: Remote work, isolation, and mental health"

redrove — Sun, 07 Jun 2026 09:25:14 +0000

I read the entire thing and it seems to me like they started with the conclusion and tried to find proof, like a lot of psychology papers.

Thoroughly unreliable.

New comment by redrove in "MAI-Code-1-Flash"

redrove — Tue, 02 Jun 2026 19:30:16 +0000

It’s about bang for buck. That high a score for 5B params is pretty good, nigh unbelievable a short while ago.

It is my belief that smaller models will get better and better, and even cloud SOTA models will shrink.

Yet another reason the current buildout will feel like the railroads.

New comment by redrove in "Odysseus – self-hosted AI workspace"

redrove — Mon, 01 Jun 2026 13:12:12 +0000

Wow, if you you could've just done docker compose up. I guess we'll never know.

New comment by redrove in "Can we have the day off?"

redrove — Thu, 28 May 2026 18:58:51 +0000

So the theory goes. Can you point out one place where this is working out?

New comment by redrove in "I bypassed AWS API Gateway auth with a trailing slash. Got $12K bounty"

redrove — Tue, 26 May 2026 10:28:14 +0000

Don’t vibe code your auth path folks.

New comment by redrove in "Show HN: Rmux – A programmable terminal multiplexer with a Playwright-style SDK"

redrove — Thu, 21 May 2026 09:43:30 +0000

I’m genuinely interested in what you find to be the problem with that?

It’s a landing page.

New comment by redrove in "Show HN: Mezz, a curl-able WiFi sandbox for IoT pentesting"

redrove — Mon, 18 May 2026 07:35:49 +0000

curl | bash and pentesting go well together

New comment by redrove in "Reverse engineering Android malware from popular Chinese projectors"

redrove — Mon, 18 May 2026 06:18:03 +0000

I’m immediately suspicious of cheap chinese crap like this.

$35 for a projector should cause you to raise at least one eyebrow.

Also, as always with “IoT” type devices, they’re best kept in an isolated VLAN with no internet access.

New comment by redrove in "GPT-5.5 Price Increase: What It Costs"

redrove — Sat, 09 May 2026 18:11:14 +0000

What coding plans are you using that let you use those models specifically and not pay per token? Which provider?

New comment by redrove in "K3k: Kubernetes in Kubernetes"

redrove — Sat, 02 May 2026 05:58:42 +0000

So this is basically vCluster[0] but Rancher branded?

[0] https://github.com/loft-sh/vcluster

New comment by redrove in "Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview"

redrove — Fri, 01 May 2026 14:13:04 +0000

cool!

New comment by redrove in "Ask HN: Does Claude use 'prior' in a Bayesian sense more than English?"

redrove — Fri, 01 May 2026 06:53:26 +0000

Nothing, but he got to plug his vibe coded startup that he advertises in the about section.

New comment by redrove in "Mistral Medium 3.5"

redrove — Wed, 29 Apr 2026 16:41:36 +0000

It’s 128b dense model. Good luck getting more than 3t/s out of a mac. It doesn’t matter if it fits or not.

New comment by redrove in "Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview"

redrove — Mon, 27 Apr 2026 15:23:56 +0000

Yes I'm on ChatGPT Pro (OAuth) and I'm trying to use gpt-5.5-xhigh.

That was the issue, 5.4 works just fine.

Support for service: priority (GPT /fast mode) would also be cool!