Hacker News Reader

Comment by redox99 | Hacker News Reader

Interesting. Since GPT4 came out I strongly believed LLMs would replace human programming. By November 2025 when opus 4.5 came out I thought it was pretty established already and most people agreed with that premise.

By human programming I mean humans typing code.

View on HN

I think the experience of the gpt4/gemini2/claude2-3 era of models was widely variable depending on the use case and how much reference material there was on the internet. I'm told it was very good at producing react code for example, but my experience was it often failed to handle Rust or Kotlin type checking and failed pretty often (in less immediately detectable ways) on PHP. So looking at the trajectory from the initial copilot (could auto compete fast inverse square root and method level problems that you could also just google), to GPT4 (which from my experience, still couldn't code) and hearing that the training data by that point was "most of the internet" it wasn't that clear to me that it would reach the point it has today.

Obviously the models themselves have trended bigger since which has helped and also just the tooling and harnesses around it and the models being trained for that use case has achieved a lot since that q4 2025 window which is about the first time it became functional for me.

I think most people failed at extrapolating, seeing the trend and imagining what is possible. Maybe they didn’t want to. This is still the issue today. Some developers were overconfident about their understanding of how AI works, without really understanding deep learning at all. They assumed the limitations and shortcomings at the time were definitive. Tooling and harnesses were not a big invention at all. They were obvious from day one but took time to build and turn them into somewhat mature products we have today.

The training data back then was sooooo bad. You'd open something at random in Common Crawl and it was truly unbelievable that models (like llama) were able to output coherent stuff. The architecture was truly miraculous if it could train well with terabytes of garbage.

There was so much low hanging fruit back then in terms of training.

Can you point out an example where someone made that mistake?

I think humans typing code is basically dead. But I also think typing code was always a minority of the work. System design, requirements, the big picture, that’s still engineer led.

LLMs can write code. What they can’t do is build systems based on vague desires from business people. Because nobody can, you need to distill and refine nebulous ideas into technical requirements and design. And that’s what engineers do.

I also think this is inherently a human endeavor. It’s about relationships and understanding. LLMs can’t accomplish this because they don’t have the political power or the ability to push back. And we probably don’t want them to have that power.

Reply on news.ycombinator.com