AI Won't Cause Unemployment

This rings so true. It seems socially acceptable to either be a regular social drinker or a renounced drinker.

the idea that this man is a billionaire is mind-blowing. this article is absolutely vapid. picking it apart is more for sport than anything else.

We had two such anti-technology jobs moral panics in the last 20 years — “outsourcing” enabled by the Internet in the 2000’s, and “robots” in the 2010’s. The result was the best national and global economy in human history in pre-COVID 2019, with the most jobs at the highest wages ever.

most people refused to believe that the internet would change anything at all. there was no panic. when bill gates went on letterman and was asked to explain how the internet was different than what already existed, he couldnt. letterman taunted him, saying "ever heard of a tape recorder? ever heard of a radio?" and he was oddly not wrong. thats the extent of the panic in the case of the internet.

AI is certainly different. it is extremely different from all other technologies because it is the only technology that replaces the human mind completely rather than just augment it. the argument of these reverse-luddites is that technology has always been good therefore it always will be good. there is no deeper analysis of whats really going on. there cant be, because it will always lead to the conclusion that our way of life as humans cannot co-exist with AGI.

one must understand that human society is held in place by an economic and technological equilibrium. if AGI were to be created, it cannot be denied that science and technology would almost instantly surge forward by trillions of man-years. we would be transported instantly into an environment where fundamental economic and technological factors are no longer valid. rules would be so different that the very concept of a country would no longer be geopoliticaly tenable. new meta-organisms would be at the center of the new equilibrium and the old human-based meta-organisms that we call countries would cease to exist. in essence, once the singularity happens the most severe and profound power vacuum that has ever existed will be created and the world will transition into a new version of itself where grotesque machine-based meta-organisms find a new equilibrium. this will involve what we call "war" and the idea that the human race could survive this is immediately dismissed. on the other side of the war will be a world where human industry is not just redundant but cosmically irrelevant. if by some miracle we survived the singularity, if our existence posed the slightest inconvenience to the machines, we would all die. this is not true currently because the dominant meta-organism is always human-based. thats the difference.

the absolute best case scenario is that humans somehow manage to parasitically attach themselves to an AGI meta-organism. they would lead lives devoid of agency or purpose except to ogle the machines. except that wouldnt be very interesting because the machines will immediately embark on doing things that a human being could never, ever understand.

people have this delusion that natural selection wont apply to AGI or the singularity. the singularity wont be about you or your fantasy. it wont be about humans or human society.

if anyone else is worried about AGI and the singularity, and wants to help prevent these things, i think we should get in touch. as far as i know, there is no serious political group that is advocating for regulation of AI based on human redundancy. i think its really important to get this ball rolling as soon as possible.

> if AGI were to be created, it cannot be denied that science and technology would almost instantly surge forward by trillions of man years. we would be transported instantly into an environment where fundamental economic and technological factors are no longer valid

I see this claim over and over, and I just don't believe it. Scientific discovery contains three phases: 1. proposing a hypothesis 2. designing an experimental procedure to get confirmation 3. interpreting the results.

Step 1 and 3 might be accelerated by the AI, but it's not clear to me that would give us any benefit because step 2 involves the real world. Say we need to build LHC at Geneva for step 2. How is that instantly going to happen?

Step 2 is the reason not even AGI can iterate faster than the outer loop of experimentation and real world feedback allows for. You can come up with a billion ideas if you can only test one per month, progress won't go fast. Like the COVID mRNA vaccine - it was ready in 3 days and tested in 11 months.

You are not wrong. This is what I would call a severe imagination deficit. Which is of course not a great trait for a venture capitalist. Andreesen is framing things in a very US centric context. Which of course represents the status quo there and the way things are; mostly for historic reasons. That's going to change for two simple reasons. Number 1 would be that it is sub optimal to say the least. It's simply not sustainable. And number 2 is that this is not the way things are elsewhere and the US is in competition with those places. Andreesen is making a closed world assumption and navel gazing. Failing to adapt basically means a slide towards being less relevant, less competitive, etc. Which is of course another way of saying that change is inevitably coming. Whether people like it or not and whether they are ready for it or not. Eventually the economy would collapse and people would adapt. But probably some time before that people will get a clue and start fixing things. Either way, things change. So predicting that things won't, shouldn't or can't change is a foolish prediction to make.

If you want a concrete example of a few places that aren't going to be held back by regulation and bureaucracy: China, India, most of Africa. They have bureaucracies of course but they are also booming economies and very much making the difference via technology. People are going to push boundaries with what is possible technically and some places are going to be more comfortable doing that than others. Ultimately, the willingness to push boundaries is what will lead to breakthrough results. If the US is not willing, too complacent, or too scared, that just means it happens elsewhere.

All this is true regardless of AI. But of course given that we have that, people are going to use that in anger for all sorts of things. Some misguided, some foolish, and some very productive and practical. Andreesen is arguing that flat panels are cheap and education is expensive. Flat panels are cheap because they are not made in the US and the production of those is not bogged down by how inefficient things have become in the US. For the same reason, you are better off getting your education outside of the US. Education is more or less free in other parts of the world. It's only this expensive in the US. Probably silicon valley is the worst place in the world to get one in terms of economics. Which is why a lot of the people working there are coming from abroad. Which is where they get most of their education.

Probably education is at the forefront of what AI can do short term btw. I've played a bit with chat gpt and it's an obviously awesome educational tool. It's like this patient tutor that is willing to explore any subject and answer any question. It might be a bit off with the facts and all the bat shit crazy hallucination. But that sounds like it should be a fixable problem. Once that is fixed, there will be no shortage of teachers any more. Anyone willing to learn will be able to learn and receive an education that is completely tailored to them. Anything, anyone, anywhere. Kids are using chat gpt already to fool their teachers. The next logical step is replacing them entirely. That's not going to take a lot of time. And it won't cost anywhere near a million dollars.