Hacker News: IceWreck

New comment by IceWreck in "Running Gemma 4 locally with LM Studio's new headless CLI and Claude Code"

IceWreck — Sun, 05 Apr 2026 22:45:05 +0000

> So while feasible it's only great for batch jobs not interactive usage.

I mean yeah true but depends on how big the model is. The example I gave (Qwen 3.5 35BA3B) was fitting a 35B Q4 K_M (say 20 GB in size) model in 12 GB VRAM. With a 4070Ti + high speed 32 GB DDR5 ram you can easily get 700 token/sec prompt processing and 55-60 token/sec generation which is quite fast.

On the other hand if I try to fit a 120B model in 96 GB of DDR5 + the same 12 GB VRAM I get 2-5 token/sec generation.

New comment by IceWreck in "Running Gemma 4 locally with LM Studio's new headless CLI and Claude Code"

IceWreck — Sun, 05 Apr 2026 20:03:28 +0000

It does if you use an inference engine where you can offload some of the experts from VRAM to CPU RAM. That means I can fit a 35 billion param MoE in let's say 12 GB VRAM GPU + 16 gigs of memory.

New comment by IceWreck in "Components of a Coding Agent"

IceWreck — Sat, 04 Apr 2026 21:22:44 +0000

> This is speculative, but I suspect that if we dropped one of the latest, most capable open-weight LLMs, such as GLM-5, into a similar harness, it could likely perform on par with GPT-5.4 in Codex or Claude Opus 4.6 in Claude Code.

People have been doing that for over a year already? GLM officially recommends plugging into Claude Code https://docs.z.ai/devpack/tool/claude and any model can be plugged into Codex CLI (it's open source and can be set via config file).

New comment by IceWreck in "Claude Code's source code has been leaked via a map file in their NPM registry"

IceWreck — Tue, 31 Mar 2026 18:08:23 +0000

> What Google and OpenAi have open sourced is their Agents SDK, a toolkit, not the secret sauce of how their flagship agents are wired under the hood

And how is that any different? Claude Code is a harness, similar to open source ones like Codex, Gemini CLI, OpenCode etc. Their prompts were already public because you could connect it to your own LLM gateway and see everything. The code was transpiled javascript which is trivial to read with LLMs anyways.

New comment by IceWreck in "Astral to Join OpenAI"

IceWreck — Sat, 21 Mar 2026 16:57:50 +0000

basedpyright has existed for years and now we have pyrefly from meta too. I think ty is also working on one.

New comment by IceWreck in "Just-bash: Bash for Agents"

IceWreck — Thu, 26 Feb 2026 15:18:02 +0000

At this point why not make the agents use a restricted subset of python, typescript or lua or something.

Bash has been unchanged for decades but its not a very nice language.

I know pydantic has been experimenting with https://github.com/pydantic/monty (restricted python) and I think Cloudflare and co were experimenting with giving typescript to agents.

New comment by IceWreck in "zclaw: personal AI assistant in under 888 KB, running on an ESP32"

IceWreck — Sun, 22 Feb 2026 18:48:09 +0000

I've been using https://github.com/sipeed/picoclaw

New comment by IceWreck in "QNX Self-Hosted Developer Desktop"

IceWreck — Sat, 27 Dec 2025 10:33:23 +0000

Blackberry OS 10 was also running QNX under the hook afaik.

New comment by IceWreck in "Nvidia to buy assets from Groq for $20B cash"

IceWreck — Fri, 26 Dec 2025 01:02:44 +0000

This is exactly what Google did with Windsurf and similar to what Meta did with Scale AI. Seems like a rising trend,

New comment by IceWreck in "NIST was 5 μs off UTC after last week's power cut"

IceWreck — Mon, 22 Dec 2025 19:24:18 +0000

We need nanosecond precision for trading - basically timestamping exchange/own/other events and to measure latency.

New comment by IceWreck in "Go is portable, until it isn't"

IceWreck — Sat, 13 Dec 2025 09:12:23 +0000

You're linking to a different version - this is the one that most people use https://github.com/modernc-org/sqlite

New comment by IceWreck in "Tongyi DeepResearch – open-source 30B MoE Model that rivals OpenAI DeepResearch"

IceWreck — Sun, 02 Nov 2025 18:54:00 +0000

LlamaCPP supports offloading some experts in a MoE model to CPU. The results are very good and even weaker GPUs can run larger models at reasonable speeds.

n-cpu-moe in https://github.com/ggml-org/llama.cpp/blob/master/tools/serv...

New comment by IceWreck in "Claude Code 2.0"

IceWreck — Mon, 29 Sep 2025 19:14:22 +0000

I was using aider quite a lot from ~ 7 months ago to ~ 3 months ago. I had to stop because they refuse to implement MCPs and Claude/Codex style agentic workflow just yields better results.

(Ab)using Agentic Coding CLIs for Data Cleaning and Standardisation

IceWreck — Sat, 27 Sep 2025 20:18:10 +0000

Article URL: https://abifog.com/blog/data-standardisation-with-agentic-clis/

Comments URL: https://news.ycombinator.com/item?id=45399003

Points: 2

# Comments: 0

New comment by IceWreck in "Seedbox Lite: A lightweight torrent streaming app with instant playback"

IceWreck — Fri, 29 Aug 2025 16:31:10 +0000

Does it download torrents on your server or web torrent on your browser? - the readme really doesn't say.

Imo downloading on the server is more useful. Web torrent is great but I don't think it's very practical in many places.

New comment by IceWreck in "F-Droid build servers can't build modern Android apps due to outdated CPUs"

IceWreck — Wed, 13 Aug 2025 08:26:58 +0000

Huawei and Honor are seperate app stores?

And Oppo and Vivo too?

In both instances one company owns the other - why have competing app stores?

New comment by IceWreck in "Ollama's new app"

IceWreck — Wed, 30 Jul 2025 22:23:48 +0000

Why not Linux? The UI looks to be some kind chrome based thingy - probably electron - should be easy to port to Linux.

Also is there a link to the source?

New comment by IceWreck in "Qwen3-30B-A3B"

IceWreck — Tue, 29 Jul 2025 23:20:30 +0000

You can already use in it ollama by using the unsloth quants:

``` ollama run hf.co/unsloth/Qwen3-30B-A3B-Instruct-2507-GGUF:Q4_K_M ```

> And what is the best offline model for coding?

That would depend on your hardware.

New comment by IceWreck in "Harper – an open-source alternative to Grammarly"

IceWreck — Fri, 20 Jun 2025 23:10:38 +0000

Slightly controversial compared to other comments here but I haven't used Grammerly at all since LLMs came out. Even a 4B local LLM is good enough to rephrase all forms of text and fix most grammer mistakes.

New comment by IceWreck in "Google Duo will be replaced by Google Meet in Sept 2025"

IceWreck — Sat, 31 May 2025 08:23:31 +0000

I think Allo and YouTube Chat were also around for the same time as Duo.