LLMs like what is being overhyped cannot take over the world.

They are not constant-thought processors. They are the most overbuilt autocomplete of all time.

The “take over the world” problem is the capitalists who see dollar signs in automating away humans.

It's OK, these stages are not supposed to be sequential. They can go through them in any order, and even cyclically return to other stages. Even full acceptance can be relapsed from time to time.

If ai allows the improvements hoped for, it's likely to increase productivity for humanity just like all other major advances did.

It's politics and laws that decide who benefits from it. Under our current systems, wealthy people and corporations benefit, but likey we'll see increasing taxes on ai output and increases in ubi type schemes for people.

AI is a bubble currently but it is an advancement in tech that provides benefit. Just nowhere near the benefit to be actual intelligence.

Human jobs get eliminated by new tech all the time. This just has and will be rapid, which leads to upheaval. Unfortunately, it's precisely at a time when the world is already moving towards war and authoritarianism, so it's particularly bad timing.