Comment by thrown-0825
2 days ago
this entire endeavor is a fools errand, and any who has used coding agents for anything more complex than a web tut knows it.
it doesn't matter how much jargon and mathematical notation you layer on top of your black box next token generator, it will still be unreliable and inconsistent because fundamentally the output is an approximation of an answer and has no basis in reality
This is not a limitation you can build around, its a basic limitation of the underlying models.
Bonus points if you are relying on an LLM for orchestration or agentic state, its not going to work, just move on to a problem you can actually solve.
You could more or less use the same reasoning to argue for why humans can't write software.
And you'd be half-right: humans are extremely unreliable, and it takes a lot of safeguards and automated testing and PR reviews etc to get reliable software out of humans.
(Just to be clear, I agree that current models aren't exactly reliable. But I'm fairly sure with enough resources thrown at the problem, we could get reasonably reliable systems out of them.)
There’s a lot of projects with only one and two people behind it and they produce widely used software. Still have to see big tech producing something actually useful with one of those agents they’re touting about.