Unless we muster the political will to stop AI development, internationally, until we can be certain of our ability to durably imbue it with the intrinsic desire to keep humans around, doing human things.
That's been tried several times now and has a tendency to end very badly for capital. You'd think folks with even a grade school level of historical literacy would know better than to stick a fork in that outlet.
I haven't seen evidence of that. I see evidence of rapid advances in task length, general capabilities, and research and development capabilities in AI, and generality, price, and autonomy in robotics.
How much headroom in these capabilities do you believe we have, before a data center can protect and maintain itself and an on-site power plant? Before robots can run a robot factory?
If you showed a conversation between Terry Tao or Steve Yegge and their AI collaborators to someone from 2021, they would consider it beyond obvious that it's AGI. Today, we know they still have some shortcomings; but in another 5 years, what looks to us today like it's beyond obviously ASI may well be enough for catastrophic, irrecoverable outcomes.
Good luck eating that.
In practice this currently means voting for political options who can correctly identify concentration of power as the root cause of most of our current and future problems, and who pledge to actually do something about it.
We witnessed the same thing with looms and other automation in the Industrial Revolution. Capital that helps you produce more. But owners faced with increased competition under commoditized production see their profit margins fall. Thus they will turn to squeezing workers - the source of value - for profit in the newly commoditized landscape - exactly what happened during the Industrial Revolution. It was only when workers got their act together and organized that this decline was stopped and reversed.
Of course the structure exists because we allow it, that's the easy part. Hard part is - why do we allow it?
However, that's not useful in predicting what capital owners will do, because they follow their local incentives. "If everyone keeps doing X, we will all be worse off" does not help unless you can create local incentives that point toward an equilibrium where everyone stops doing X.
In this case, no capital owner is individually better off by unilaterally refusing to chase more efficient returns on their capital. We would need an international agreement, with enforcement mechanisms, like I mentioned above.
It's not the AI you have to convince, it's your government and the people running tech companies. Dario Amodei was cheering for AI to take all programming jobs (along with the others). If that happened, it would be an unmitigated disaster for millions of people. Imagine a student who comes out of a CS major with tons of student debt. How much sympathy does Dario feel for this person? Getting him to STFU would be a good first step.
> the political will to stop AI development
The reason that's not likely is that it's an arms race. You stop AI research here, but how can you trust that China and Russia are doing the same? Unlike nuclear bombs, the potential harms are less tangible.
I don't need to imagine this student, I'm friends with some who are going through this right now. They graduated almost a year ago and haven't found work yet. One of them jokes about suicide often and I don't know how to help him
The social contract between labour and capital has been frayed for a long time, but it is near breaking now. It's going to get worse, maybe a lot worse, before it gets better. If it ever does get better
Perhaps one of the secondary effects of AI replacing developers will be mobilising a group of smart, motivated people to the left
(It's always interesting to think of the secondary effects which kick in past a certain point of growth. High-multiple stock valuations often fail to take these into account. For the East India Company, for example — your company can keep growing until it's the size of a country. But suddenly other countries treat you as a foreign power rather than a pet.)
I am on this site because it is one of the less shitty places on the Internet (in terms of usability, privacy etc.) to have some form of discussion, but I never identified as a "hacker", "techie", "entrepreneur" or "temporarily embarrassed billionaire". AI didn't change my view on anything, except it has shown me how blind and naive people can be.
Of course I tend to focus on aspects that are being discussed here (context of software engineering).
My personal bet would be that in the medium term, there will be a reversal of the idiotic belief that you can immediately just lay off developers because of LLMs. If your developers are more productive because of LLMs, you still have an advantage by having more developers than the competition. There's also a lot of institutional knowledge that's just not documented. You fire key people, you can cripple your organization.
In the longer term, I think AI will eventually take jobs, and unfortunately, it will have major negative societal impact. I doubt that our governments will be proactive in trying to anticipate this. They will just play damage control. There's probably going to be an anti-AI social movement. You'll have the confluence of more and more disinformation and AI slop online along with more and more job loss. There are probably going to be riots. Some people think UBI is inevitable. I think the problem is that if the government puts UBI in place, they will only give you the minimum necessary so that you don't starve. Just enough to afford to rent a bedroom, eat processed food and stay online all day.
The real split are between those that support science and technology, and those that hate science and technology and want to see more children die.