2 ms·I agree. I don't find gpt-3.5 worth using for real work as there are too many failures.by ParetoOptimal 4y agoI agree. I don't find gpt-3.5 worth using for real work as there are too many failures.