Codex is horrible and I’m not exaggerating this; it is actually god fucking awful, ultra extremely slow, and makes a ton, A TON!!! of mistakes... Like I need to correct Claude code maybe 2, 3 times and then it gets it right, most of the time it gets it right in one shot. But Codex!!! OMFG. This is like minimum 10 tries! MINIMUM! for the same stuff I’m working on. It's insane that the difference gap is so huge, I would not care if Codex is slower but at least it gets it right! NOOO it’s bad trust me 100% when I say it’s fucking horrible and you will be extremely mad when using it. And this is without any exactifications and 100% honest feedback. It's unbelievably useless; I hate Codex and how they openly lie at OpenAI that it is good. How can Open AI with UNLIMITED money allow themself to be crushed by weaker company like that every single time? HOW is that even possible? Just hire more programmers, there are plenty of really smart people who WANT to work for you! HIRE THEM AND FIX YOUR SHIT, BE BETTER. Open ai is run by clowns🤡. I wish that was not the case because Anthropic at this point can do whatever they want with us I wish we had good competition, but competition is a complete waste of time. Claude code is beyond anything that exists right now unfortunately and that is a FACT. If you read anywhere that someone is saying codex is good they are most likely an AI bot account that is instructed to say that! TRUST ME!!!!!! I know what i'm talking about.