I don’t agree with the assumption that any of the LLMs are super intelligent. They aren’t even close to actual human intelligence, and most real-world experience has shown that they are best used as a tool, like a graphing calculator.
We are also a long way off from a fully automated industry. The energy required for a fully staffed AI workforce is staggering, the current usage of AI isn’t even sustainable, let alone powering millions of physical bots plus the AI needed for piloting them.
LLMs have zero intelligence. They can’t think, at all. Zero capacity for thought. Its autofill. It takes the input and uses it to output the most likely response. That’s it, full stop. Unfortunately the assumption that its intelligent is making people even dumber.