Comment by vatsachak
6 hours ago
These things still can't solve problems in the right way. The benchmarks prove that they can solve problems. But in practice they will make your codebase look like the output of some compilation process
6 hours ago
These things still can't solve problems in the right way. The benchmarks prove that they can solve problems. But in practice they will make your codebase look like the output of some compilation process
I'm afraid this is "cope".
There's hardly any work you can think of which can't be done faster / beter with ai assistance, when your role is of reviewing and directing. If you have an anti-example, would like to hear.