While I wholeheartedly agree with this, the game plan will be to use automation to lower prices on good and services while keeping wages low, if the people are hand to mouth they won't be able to make those choices. They won't have enough discretionary income to have those choices. Walmart brings in cheap products to a town and puts the ma and pa shops out of business, then turns around and pays employees less than the ma and pa shop would -- so everybody in the town has no alternative but to shop/work at Walmart. All the productivity and its associated wealth flows up to the 1%. This has been happening since at least the 1980s. AI is just another link in that chain.

What is the long game strategy here for Google? Glut the web with art, content and programming that is all derivative, and them make those items copyrighted? How will this be good for creativity and innovation? How will this be good for humanity? How about an AI that runs a company -- ethically -- for basically nothing? Let the working class own the company and have it run by AI and eliminate the greed and psychopathy at the top.

Certainly nothing in the posted video wouldn't have been caught by a good IDE during the coding process. I still haven't seen chat GPT do anything that looks like actual cognition, no high level thinking. Its not at a point where it can fix real problems yet.

I keep trying to make ChatGPT a useful tool. I have used it to generate some legal language for contracts, and while it creates a useful starting point, the wording and details need to be heavily reviewed and modified before they can be used. As far as coding, for Rust, its answers are useless for all but the most trivial examples. I find it faster 9 times out of 10 to just look for the answer I need rather than trying to get the shit-code from GPT to work. Here is a good example: I wrote a little challenge for my adult offspring as part of an Easter treasure hunt. The challenge was to get a small Rust program of a dozen lines or so to compile (the clue was in an array of binary values). The code spit out a little clue. My son tried and tried to get Chat GPT to interpret the answer for him. It gave all sorts of unhelpful answers as though they were going to help. When they pasted the code directly into GPT it confidently gave them a clue that sounded reasonable. It was actually quite comical, because they thought they had beat the challenge, until I told them the very plausible sounding clue was 100% wrong! I finally steered them to an online rust playground and they got the correct answer in seconds. Chat GPT was worse than useless, it was a convincing distraction. I have yet to have someone show me it can be trusted, or relied upon for anything more than an amusing distraction. If you are using this in your professional life, I would take caution and review everything it does.