Worth noting this is not only good on benchmarks, but significantly more efficient at inference https://x.com/_thomasip/status/1995489087386771851

Wow, thanks for the info. I'm planning on testing this on my M4 Max w/ 36 GB today.

edit:

So looking here https://ollama.com/library/gpt-oss/tags it seems ollama doesn't even provide the MXFP4 variants, much less hide them.

Is the best way to run these variants via llama.cpp or...?

Still loyal to aider. It just fits my style better, as a very fine tool. I have my workflow and scripts around it, switch freely between gpt-5/sonnet (a bit of gemini-2.5-pro too) and enjoying life.

I wish it was maintained by a larger team though. It has a single maintainer and they seem to be backlogged or working on other stuff. If there was an aider fork that ran forward with capabilities I'd happily switch.

That said, I haven't tried Claude Code firsthand, only saw friends using it. I'm not comfortable letting agents loose on my production codebase.

I’m glad that Go went the other way around: compilation speed over optimization.

For the kind of work I do — writing servers, networking, and glue code — fast compilation is absolutely paramount. At the same time, I want some type safety, but not the overly obnoxious kind that won’t let me sloppily prototype. Also, the GC helps. So I’ll gladly pay the price. Not having to deal with sigil soup is another plus point.

I guess Google’s years of experience led to the conclusion that, for software development to scale, a simple type system, GC, and wicked fast compilation speed are more important than raw runtime throughput and semantic correctness. Given the amount of networking and large - scale infrastructure software written in Go, I think they absolutely nailed it.

But of course there are places where GC can’t be tolerated or correctness matters more than development speed. But I don’t work in that arena and am quite happy with the tradeoffs that Go made.