Raw LLM Responses

But ultimately this question is a bit flawed, getting rid of software engineers is about closing the loop of automation. The same problem exists in manufacturing industry, and repair for that matter. We today have the technology to automate every single restaurant in the world, we could do it completely, we could infact also fully automate farming, and manufacturing, but the cost to do it all at once is impractical, and for most tasks it is just simply cheaper and less risky to still have humans in the loop. At some point that wont be true anymore, and it will probably rapidly be cheaper to have fully automated factories that are just managed by humans. But that takes way more than being able to do each little task and fix every little error in an automated way. You need to know every single problem that might arise can be solved automatically, or you will still need humans to serve as problem solving redundancy. It is all the small problems that fucks it up, we can't garuantee that the production line wont come to a halt and not be repaired by the repair bots just yet, and some simple tasks are just very fiddly and bespoke, and are hard to automate. Its just the shear amount of gaming out of small things in a bespoke way, that stands in the way of automating pretty much every single manual job humans do today from car factories to diners, and closing that loop depends heavily on building the resources to do all that bespoke engineering work, it is mostly not even worth it beyond a certain point, even warehouses are easier to run with humans in the loop as of right now. With software engineering, the problem is identical except it is also harder, because it includes everything to be solved above. Its the small details and problem solving redundancy that is the real killer, not standard applications, more so bespoke implementation.