The AI coding agent that disappoints

I am using Claude Code myself and there is people talking about that AI Coding agents just don't live up to standards, that it misses stuff and does not follow instructions, making it daunting to work with it. You can see it here AI Coding Sucks on youtube.

I can relate, using Claude Code now for 5 months, it does not sees to amaze how often it falls through and does simple things wrong.

Still helps

But I am using coding agents still because it helps me type less still, in programming when running commands from the commandline even 1 character wrong makes a mistake. This mistake the coding agent never makes, it does not mistype by one character.

It is able to work independently, it can fix small stuff that you would never have bother fixing, but can verify in 10 seconds.

It can migrate code and fix bugs if it has tests that are failing. That is a huge help.

Optimistic future

AI agents are more like a tractor, it is much more efficient at turning over the ground than doing it manually, but it is still just a tool.

Andrej Karpathy talks about talking information out of the models and keeping the working/thinking and I think that is the right method.

He also speaks about entropy. An idea could be to have an AI look at the first AI's work to catch the silly mistakes, but first, that is really slow if you want that today, second is that Andrej talks about entropy collapse, that we are not seeing AI act as humans since they are too focused.

So I believe it is the future, but that we have not seen the final form of the AI coding agent.