AI will reduce the demand for "knowledge" and cultural workers (writing, music, art, programming, data analysts, etc, basically everyone who doesn't do manual labour) by 90-99% with the remaining workers being AI helpers/"prompt-engineers". Once this transition picks up steam, it will happen so quickly that the rust-belt-ification of the midwest will look measured and quant in comparison, and we saw how government didn't act to help people that got dislocated through absolutely 0 fault of their own. And that's not even getting into how AI/ML solutions often only hit the 80 mark in the 80/20 rule and cause bizarre issues for people that run afoul of the system (see: Google's ML driven account takedown processes with youtube/play store), while winning purely on their low marginal cost.

If AI was paired with an explicit promise of a generous UBI then I would merely personally choose to not consume AI content and not be against it in general. But that isn't how the market economy works unfortunately.

That seems less "anti-AI" and more anti our current economic systems, the inequality they produce, and their increasing incompatibility with the present and future. The "incompatibility" being between our values and the resulting outcomes. Our values tell us that AI and automation freeing up immense amounts of human labour/effort so that it can be directed to new (creative) ends should be wonderful and celebrated, but the reality is that it's bringing about a frightening future where an increasingly small few control the vast majority of the power in the world, and direct it towards "profitable" ends which are largely divorced from human values.

Yep essentially that is my thinking! AI/ML just hyper-charges the ability of capital to produce with an ever shrinking amount of required labor. And in our free-market with no guard rails every uptick in unemployment leads to more misery.

Yup. I'll add that it's not just unemployment that hurts us (under our current economic systems), but also underemployment, as we see with the "gig economy". Some people will overly focus on the unemployment stats, and naively conclude that all's well despite many people not living up to their potential, and barely making ends meet.

I've heard this argument over and over. I used to be an avid believer in this myself.

However, this has been the fear of people for years, with every new tech and every time, it seems like some new thing replaced the old, or even added to it. The electric guitar, the synth, DAWs, they never replaced musicians, they just gave them more tools, yet there were always nay sayers.

I see this happening in business too. People haven't been fired en masse due to tech and we're currently having massive shortages in the work force. Even with McDonald's moving to self-service screens. I'm getting less and less scared of tech taking over and removing people from everything.

It's easy to discount a disrupting technology when you see yourself as far removed from it.

- Autonomous McDonalds doesn't bother me because I don't work at McDonalds.

- Stable Diffusion doesn't bother me because I'm not an artist.

- Copilot doesn't bother me because my code requires creativity and critical thinking.

But then again, Copilot doesn't need to be "good" before it starts to take away programmer jobs. It just needs to be better than the worst programmer on your team.

And yet, McDonald's and many restaurants in general have been almost permanently short staffed since Covid. In spite of attempts at automation.

I have yet to see anything impressive from Copilot. You still can not describe a set of business requirements to it and have it translate that into a full stack application that factors for performance, security, GDPR, usability, etc. It's at best a form of auto complete.

And yet they have personnel shortages, which was my point. Where one job disappears, another opens up.

> [new technology] will reduce the demand for [some class of] workers

Is the exact argument used against literally every new technology.

The difference is the speed and generality of the change brought on by this specific tech. That is why AI/ML is so heavily capitalized right now, people with capital know the power of this. Previous changes were painful (automation of auto manufacturing + offshoring of auto manufacturing led directly to political backlash for decades) but they were scoped to individual industries at a time and with enough time for society to adjust. I think AI/ML is a step change in the speed of this and we won't be able to just muddle through it like we did with manufacturing automation.

At least that is my thought on how this is different, I am aware that this has been said for basically every previous technological advancement.

This is the exact same argument people have made about every new piece of technology since the beginning of written history.