Raw LLM Responses

I think that most of the people that think AGI will happen by 2030 or any reasonable date didn't research topic by any other metric than watching YT videos. First of all, development of AI is EXTREMELY expensive, to the point that only 3 big companies can reasonable do it (google,amazon,ms) and even them have to consider costs and implement some payment plan. The energy is also the concern, such computing power needs very much of it and to make it competitive you would have to make it: -cost low enough( lower than human doing the same job) -perform good enough for people to choose AI product (you can check what is happening under first "only ai" coca cola commercial on yt) -have enough data to train it(which is already the problem because it was okay to use open internet for it but it got flooded with AI content which renders it useless for this now and it is unfeasible to create or find big enough data set to train something that would become AGI) -having sustainable infrastructure to do it, computing power needed for it is extremely HIGH so even if they will make something approaching AGI it would be so cost inefficient to replace human with it is laughable. You could say now "wow retard, but everyone is hyped, they can't be wrong , people invest etc" well yea, they invest and NVidia(which produces most AI chips) and AI companies have overinflated evaluation and HAVE to preach the AI is new big thing otherwise investor will stop pouring money but all this companies operate on loss now(they make less revenue than they have costs but investors money keep them up) but they don't have any reasonable plan how to make money on it so good luck. AI can be used right now to get some people out of work like some people who do some menial task(IE, they choose (on the basis of content(NOT LABEL OR ANY AUTOMATED THING)) where given package go etc. or make less people do same job(accountants with AI tool vs 2 of them) but that as far as it currently goes with current price/efficient ratio