Why are senior software developers speaking out against using AI platforms to write code, when some of the largest tech companies are developing the tools to do exactly that?

I find AI code writing to be similar in a lot of ways to self driving cars. Do they drive on the road? Yes. Can you allow them to operate all by themselves? No way, you need an experienced driver to be able to take over things if it fails. Can that experienced driver at least take a nap or watch TV while the car drives itself? No, the whole point is for him to react when the car does something wrong. With AI writing code, it’s kind of the same. Yes, AI can produce code, and yes, a lot of times that code does what it’s supposed to do. But you need an experienced developer not only to fix it when it does not but also to take notice when it only seems to do what it’s supposed, but it actually does something slightly different. Or it does it in a suboptimal way. And the more experienced you are, the more you realise just how much work that AI generated code still needs. That’s why inexperienced developers are more excited about it than experienced ones. And non developers believe that developers will be replaced by AI finally. I should also note that not all software developers are the same. They range from those which would just produce code, to those that would actually design a complex piece of software with a bunch in the middle of the pyramid. The AI so far can only (somewhat) replace the first category, the lower level ones. Which are needed, but not enough in order to actually create deliverable software. When you read an article about someone who used AI only to create something like a small game, it’s usually a software architect who already knew how to build that app and used his knowledge to ask exactly what was needed from the AI and then pieced everything together. We’re still far far away from non developers using AI to create software.