Hacker News Reader

Comment by Kuyawa | Hacker News Reader

* rolls eyes *

Code is solved, no discussion about that. Whatever piece of code you think you can do better than AI, you my friend, are wrong. There will be people that will take months, perhaps years to understand that but make no mistake, code is solved and apps will cost nothing to build, nothing to copy and nothing to implement or maintain

And the good thing? it will get better, faster and cheaper than free. You can't see it? Not my problem, all yours, get your harness du jour and start learning, well no, you will be laid off no matter what you do, so, I have no advice to give

View on HN

> code is solved and apps will cost nothing to build,

it probably depends on apps. For some generic boilerplate fitness tracker maybe.

But I write some complicated code daily using all Frontier models (I switch between them), and in all cases it takes many back and force iterations to get to the code quality (clean design, no obvious bugs) up to my standards.

There is absolutely not a single line of code human created, no matter how abstract or esotheric, that an AI model can not code faster, cheaper and better. None, not one, in the whole history of mankind and computing

Now, code generated by AI, of course it takes time to generate by our current models, trillions of parameters analyzed to produce an output takes time and resources. You don't take credit for that, I stopped doing that. It is the model that deserves the credit. We are not writing complicated code, we are delegating that to AI models and they make take time to resolve the problems they may face, but they still will do it a million times faster than the best of us

Coding remains solved

looks like we are in dissareement

You aren't coding though. The LLM is coding and you are specifying.

It is just a function. A next token prediction function. Garbage-in/Garbage-out is as true as ever.

Coding is solved. Specification and translating requirements into specifications is an unbounded domain and unsolvable by definition.

What LLMs change is that now you develop the specification through iterative implementation and testing.

You start with a weak specification which produces slop, then you progressively build the specification with different approaches to automated testing. The tests are the specification. Ultimately, what you ship is the tests. They are the only thing that proves the functionality of the program. An LLM can produce code that passes any test suite you give it. If the result is bad, the problem is not the code, the problem is the specification.

> What LLMs change is that now you develop the specification through iterative implementation and testing.

historically coding involved both logic design and actual writing code. When people say coder they always meant both.

So, you are trying to win argument by inventing new meaning of term.

Reply on news.ycombinator.com