Hacker News: MrScruff

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

MrScruff — Sat, 04 Apr 2026 16:38:52 +0000

It's a good question, I've wondered that myself. I haven't used GLM-5 with CC but I've used GLM-4.7 a fair amount, often swapping back and forth with Sonnet/Opus. The difference is fairly obvious - on occasions I've mistakenly left GLM enabled running when I thought I was using Sonnet, and could tell pretty quickly just based on the gap in problem solving ability.

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

MrScruff — Sat, 04 Apr 2026 16:29:40 +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.

Unless I'm misunderstanding what's being described here, running Claude Code with different backend models is pretty common.

https://docs.z.ai/scenario-example/develop-tools/claude

It doesn't perform on par with Anthropic's models in my experience.

New comment by MrScruff in "April 2026 TLDR Setup for Ollama and Gemma 4 26B on a Mac mini"

MrScruff — Fri, 03 Apr 2026 18:59:00 +0000

I've been playing with the open models since the original llama leak. They're getting better over time, are useful for tasks of moderate complexity and it's just cool to have a binary blob of knowledge that you can run locally without an internet connection.

However you should manage your expectations. Whatever the benchmarks say, you'll quickly realise they're not at all competing with Sonnet let alone Opus. Even the largest open weights models aren't really doing that.

New comment by MrScruff in "April 2026 TLDR Setup for Ollama and Gemma 4 26B on a Mac mini"

MrScruff — Fri, 03 Apr 2026 18:43:16 +0000

Haven't really tried GLM5 much but I've used 4.7 quite a bit and it was pretty far from competing with Sonnet at the time, although I saw claims online to the contrary.

New comment by MrScruff in "Say No to Palantir in Europe"

MrScruff — Sun, 29 Mar 2026 16:17:05 +0000

Calling everyone you disagree with a 'bro' doesn't make your point any more convincing.

New comment by MrScruff in "Personal Encyclopedias"

MrScruff — Thu, 26 Mar 2026 08:35:27 +0000

I would say thinking about the indended audience for your creative outlet is a good discipline - even if it's only one person. It often gives the project more of a focus which helps with motivation and makes it more enjoyable.

New comment by MrScruff in "Thoughts on slowing the fuck down"

MrScruff — Wed, 25 Mar 2026 20:20:27 +0000

Honestly a lot of useful software is ‘unimportant’ in the sense that the consequences of introducing a bug or bad code smell aren’t that significant, and can be addressed if needed. It might well be for many projects the time saved not reviewing is worth dealing with bugs that escape testing. Also, it’s entirely possible for software to be both well engineered and useless.

New comment by MrScruff in "Ask HN: AI productivity gains – do you fire devs or build better products?"

MrScruff — Wed, 25 Mar 2026 10:26:37 +0000

For sure, but I haven't written a single piece of software where security would ever be considered a factor. Not all software runs on the web, not all software deals with accounts etc.

New comment by MrScruff in "Goodbye to Sora"

MrScruff — Wed, 25 Mar 2026 08:23:39 +0000

Turns out there are whole categories of software where 'extremely fast and good enough' is what matters, even for skilled software developers.

New comment by MrScruff in "Ask HN: AI productivity gains – do you fire devs or build better products?"

MrScruff — Sun, 22 Mar 2026 19:10:26 +0000

I see a lot of people talk about 'insecure code' and while I don't doubt that's true, there's a lot of software development where security isn't actually a concern because there's no need for the software to be 'secure'. Maintainability is important I'll grant you.

New comment by MrScruff in "Thinking Fast, Slow, and Artificial: How AI Is Reshaping Human Reasoning"

MrScruff — Sun, 22 Mar 2026 14:10:37 +0000

I think this is too broad. If, for example, I get Claude to set up a fine tuning pipeline for rf-detr and it one shots it for me, what have I lost? A learning opportunity to understand the details of how to go about this process, sure. But you could argue the same about relying on PyTorch. Ultimately we all have an overarching goal when engaged in these projects and the learning opportunity might be happening at an entirely different level than worrying about the nuts and bolts of how you build component A of your larger project.

New comment by MrScruff in "AI coding is gambling"

MrScruff — Wed, 18 Mar 2026 19:22:30 +0000

Yeah, I used to enjoy writing code but after a while I realised I actually more enjoy creating tools that I (and other people) liked to use. Now I can do that really quickly even with my very limited free time, at a higher level of abstraction, but it's still me designing the tool.

And despite the amount of people telling me the code is probably awful, the tools work great and I'm happily using them without worrying about the code anymore than I worry about the assembly generated by a compiler.

New comment by MrScruff in "Wired headphone sales are exploding"

MrScruff — Sat, 14 Mar 2026 10:44:01 +0000

I think on the go is the point. I love my AirPod Pros but I wouldn’t listen to them sat at my desk.

New comment by MrScruff in "John Carmack about open source and anti-AI activists"

MrScruff — Fri, 13 Mar 2026 20:36:05 +0000

That would imply that there will never be an adequate open weights coding model. That might be true, but seems unlikely.

New comment by MrScruff in "Ask HN: What Are You Working On? (March 2026)"

MrScruff — Mon, 09 Mar 2026 06:58:57 +0000

I’m building an application for documenting modular patches, mostly for my own use case. It uses ML to recognise the patch points, knobs and toggles from a photo of the front panel. You can then build racks from the scanned modules and then store presets of the knobs and connections which are displayed as simple schematics. Idea is ultimately to have it on an iPad as reference to accompany a live performance. Had some fun fine tuning the cable physics engine.

New comment by MrScruff in "How far back in time can you understand English?"

MrScruff — Wed, 04 Mar 2026 08:21:54 +0000

Is that true though? I don't think the median population member spoke in this way in the 1800s. Obviously in 2026 we have a massive written record of the popular speech patterns of the entire spectrum of society, but that's down to the internet more than anything.

New comment by MrScruff in "My spicy take on vibe coding for PMs"

MrScruff — Wed, 04 Mar 2026 08:12:49 +0000

It should be obvious, particularly from this line:

> It is truly the next thing, and the future, probably happening in the next 2 years, or in 2 years in 2 years.

New comment by MrScruff in "How far back in time can you understand English?"

MrScruff — Sun, 22 Feb 2026 10:23:50 +0000

Seems like some of the initial changes are reflecting more than just the evolving language. He’s comparing someone using informal slang “not gonna lie” against someone writing extremely formally “Hunger, that great leveller, makes philosophers of us all, and renders even the meanest dish agreeable.” which I’m not sure makes sense.

New comment by MrScruff in "The EU moves to kill infinite scrolling"

MrScruff — Sat, 14 Feb 2026 00:26:00 +0000

I suspect there's not a huge amount of overlap between those who would like this banned and those who are targeted by it.

New comment by MrScruff in "The AI Vampire"

MrScruff — Wed, 11 Feb 2026 11:00:33 +0000

There's going to be a long tail of domain-specific tasks that aren't well served by current models for the foreseeable future, but there's also no question the complexity horizon of the SotA models is increasing over time. I've had decent results recently with non-trivial Cuda/MPS code. Is it great code/finely tuned? Probably not but it delivered on the spec and runs fast enough.