Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: State of the Art of Coding Models, According to Hacker News Commenters

hnup.date

Post reply on HN
Print view View original

71–80 of 96 posts

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#71

Post by 2ndorderthought »

miyoji wrote: post #69
2ndorderthought wrote: post #30

Earlier quoted context omitted.

They are all stealing from each other just like how they all stole from us. Grok supposedly admitted to distilling from open ai for instance. I think it's also unfair to say their success is solely due to stealing data. They are contributing a lot of advances to the literature about what they are doing. The proof is in the results we have 27b models you can vibe code with. Not 1t+ It's murky sure. But there are smear…

> I think it's also unfair to say their success is solely due to stealing data. Their models would be completely useless if they didn't train on stolen data, so no, it's not unfair at all.

Name one company selling AI models who didn't steal their data or aquire it through dubious or unethical means?

If anything it's a Robinhood story.

Regardless I doubt they would be useless at all. Alibaba has access to tons of data and they make qwen. Qwen models are insane for their size.

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#72

Post by miyoji »

2ndorderthought wrote: post #71
miyoji wrote: post #69

Earlier quoted context omitted.

> I think it's also unfair to say their success is solely due to stealing data. Their models would be completely useless if they didn't train on stolen data, so no, it's not unfair at all.

Name one company selling AI models who didn't steal their data or aquire it through dubious or unethical means? If anything it's a Robinhood story. Regardless I doubt they would be useless at all. Alibaba has access to tons of data and they make qwen. Qwen models are insane for their size.

> Name one company selling AI models who didn't steal their data or aquire it through dubious or unethical means?

That's what I said. I was absolutely referring to OpenAI and Anthropic as well as the Chinese models.

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#73

Post by julianlam »

asnelt wrote: post #65
julianlam wrote: post #38

Interesting that Gemma 4 didn't crack the top 10. I've been experimenting with the 26B-A4B model with some surprisingly good results (both in inference speed and code quality — 15 tok/s, flying along!), vs my last few experiments with Devstral 24B. Not sure whether I can fit that 35B Qwen model everybody's so keen on, on my 32GB unified RAM. However I think I may be in the minority of HN commenters exploring models f…

Can you elaborate on your setup? What harness are you using with Gemma 4 on your 32GB machine?

[dead]

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#74

Post by 2ndorderthought »

julianlam wrote: post #40
esperent wrote: post #36

Earlier quoted context omitted.

It's well priced but does that have much relevance for "state of the art coding models", specifically? I wouldn't use Gemini 3 Flash or GPT 5.4 mini for anything except the most trivial work, although both are useful for basic exploratory work. So I'm using a heavy model for the bulk of the work and the cost of that so far outweighs the light model that the light model cost is effectively irrelevant.

It's so interesting to see the wild pendulum swings of LLM sentiment here. If one likes a model then it's capable of one-shotting entire apps. Otherwise it's "only suitable for the most trivial tasks". Never in between.

It's so true. I bet 80% of questions normal people even ask chatgpt/copilot could be answered with an 8b model trained on recent data.

I don't think people realize how small the gap between free to cheap models have from frontier models. It's going to be commoditized a lot faster than marketing will catch up. Once cash gets tight or prices rise it's more or less done for.

Especially considering some of the small free/cheap models can one shot code now.

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#75

Post by 2ndorderthought »

tokkkie wrote: post #39

more users = more complaints. negativity just means popularity. kimi...?

That's not really true. It also indicates how widespread the issues are. The graph lets you normalize by user reports by the way.

Kimi specifically kimi2.6 from moonshotai is a new open weight cheap model that has the performance of frontier models. The price point is it's main draw

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#76

Post by julianlam »

fredcallagan wrote: post #70

Judging by how things are moving ( pricing models, limits, harness patchy updates ), it feels like the real salvation will be a combination of more mature OS models and some open source harness setup like OpenCode or similar. I'm feeling like OS models are nearly there, and with the proper setup and harness might already be there. What are the general thoughts on this ?

I only started playing around with local inference a couple weeks ago. Prior to that I was just using Gemini via web since it came with my Workspace subscription, but I did not want to be reliant on the cloud.

Others will have a better idea since they've been messing around with local inference longer than I, but I am quite impressed with the models I have been loading on my laptop with only iGPU. As of this week I no longer feel like I am playing second fiddle with slow inference and small models. Gemma 4 (and maybe Qwen3.5, haven't tried it yet) seem to have changed the game this month!

Even with trying some absolutely shiiiiite models (I only had 16GB unified RAM at the start), I was suitably impressed that I splashed the $300 to double my RAM. I am happy that this one time cost was enough to break through to smarter models and faster inference. No ongoing cloud costs!

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#77

Post by 2ndorderthought »

julianlam wrote: post #76
fredcallagan wrote: post #70

Judging by how things are moving ( pricing models, limits, harness patchy updates ), it feels like the real salvation will be a combination of more mature OS models and some open source harness setup like OpenCode or similar. I'm feeling like OS models are nearly there, and with the proper setup and harness might already be there. What are the general thoughts on this ?

I only started playing around with local inference a couple weeks ago. Prior to that I was just using Gemini via web since it came with my Workspace subscription, but I did not want to be reliant on the cloud. Others will have a better idea since they've been messing around with local inference longer than I, but I am quite impressed with the models I have been loading on my laptop with only iGPU. As of this week I n…

It's awesome. Even on a trash computer you can run a small model that works just about as good as anything else for basic questions for free and no privacy issues. It's gotta be the future.

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#78

Post by 2ndorderthought »

miyoji wrote: post #72
2ndorderthought wrote: post #71

Earlier quoted context omitted.

Name one company selling AI models who didn't steal their data or aquire it through dubious or unethical means? If anything it's a Robinhood story. Regardless I doubt they would be useless at all. Alibaba has access to tons of data and they make qwen. Qwen models are insane for their size.

> Name one company selling AI models who didn't steal their data or aquire it through dubious or unethical means? That's what I said. I was absolutely referring to OpenAI and Anthropic as well as the Chinese models.

I misunderstood you. I am so used to people just slamming local models over data issues and I knee jerk replied to you. Sorry.

Re: Show HN: State of the Art of Coding Models, According to Hacker News Commenters

#79

Post by dzink »

My push for trying local was the wildly unpredictable but systematic performance of large models like Opus and ChatGPT. It feels like at different times of day or week they are getting degraded beyond belief. I don’t know if it is deliberate, a function of demand, or a function of the models themselves. We are all learning the shape of this space by trying. I need to be able to rely on consistent performance - and maybe that means putting some harness of benchmarks between models and maybe it means between different inference providers, and maybe local.
Post reply on HN