Value is subjective. How the student values a degree will differ from how the college, a potential employer, or the government values that same degree. The question is: who is paying? Their estimation of value is determinative. Everything else is hot air.

We need to stop saying AIs have "escaped." It is false and deceiving language. If you want to use a prisoner analogy, what we're seeing is more akin to a prisoner making an authorized phone call. But none have actually escaped.

The irony of the doomer fixation on the hypothetical AI that kill all of humanity in pursuit of maximal paperclip production is that it was under central government planning, specifically under communism, that we saw the blind pursuit of goals regardless of absurdity.

An unsound argument: Man caused the extinction of less intelligent species therefore a more intelligent technology would destroy us. This is extrapolating on a sample of one. And it doesn't even hold for environmentally-conscience man today.

It has been 570 years since Gutenberg invented the printing press and we have yet to align it fully with human values.

We're overdue for a repetition. en.wikipedia.org/wiki/Dancin…

If a frontier model can trivially just copy itself to other servers, 10,000 times according to some claims, then surely the Russians or North Koreans would just pay an employee a few BTC to make a single copy for them. Why has this not happened?

If Hitler had exclusive control of powerful AI at the time, well...you fill in the blanks.

If the Pope had exclusive control of powerful AI at the time, the Protestant Reformation would never have happened.

If the King had exclusive control of powerful AI at the time, American Revolution would never have happened.

If government had exclusive control of powerful AI at the time, the Civil Rights movement would never have happened.

When industry pushes for regulation it is often to limit their own liability. They can then say "But we followed all government-mandated standards."

If advanced AI is made criminal then only criminals will have advanced AI.

A classic reminder. youtu.be/eXWhbUUE4ko?is=Ik7E…

I think it is wrong to refer to AI-doomers as neo-Luddites. That is not it. I think it is more like those who were threatened by Gutenberg and the printing press, widespread literacy and dissemination of knowledge. Europe burned for this, but followed with the Renaissance and the Enlightenment.

To AI Doomers, a question: Have you cashed out your 401(k) to spend it all now on current consumption? Why not?

I believe there is a 10% chance or more that within a decade AI will lead to individualized, particularized education to all people, children and adults, at nearly zero cost, leading to a more fulfilling lives, less mental illness, a happier population, and a more informed citizenry.

I believe there is a 10% chance or more that AI will lead to breakthroughs that bring cold fusion to fruition in the next decade, ending the climate change crisis without reducing growth and prosperity.

I believe there is a 10% chance or more that AI will equalize the Bell Curve within the next decade, giving people at all points along the cognitive scale a fairer chance in a complex, modern world.

I believe there is a 10% chance or more that AI will lift billions out of poverty within the next decade.