[xg15]

> The causes of AI are plainly the curiosity of researchers and the greed of corporations who hope to make money with it.

This doesn't completely add up though. The current flavor of AI took 10-15 years of massive research and capital investment to be developed: Think of the effort of scraping most of the web for training data, then running hundreds of the most powerful GPUs available for a year for the pretraining, then paying thousands of workers to label the data for RLHF. There has to have been a conscious management decision to pursue that specific direction of research, no matter how many resources it would require - and mere scientific curiosity doesn't strike me as sufficient of a reason to explain the expenses.

Expectations of profit would, but if that was the driving factor, then it's weird how haphazardly the whole sector goes about monetization strategies. ChatGPT was presented as a sort of open-ended tech demo with not even any specific purpose. Right now, tech companies are almost desperately shoving AI into about any existing product they can think of, usually for free and often even against the preferences of their users.

This doesn't look like a successful monetization strategy for me - if anything, AI looks like the world's most elaborate case of investor storytelling.

I don't want to rule out that they'll eventually find a business model for AI, but it seems weird to commit to a technology which requires this kind of extreme resource investment to be useful without having any idea what you actually want to do with it, once you have it.

[airstrike]

> > The causes of AI are plainly the curiosity of researchers and the greed of corporations who hope to make money with it. This is exceedingly evident. They shout if from the rooftops.

> There has to have been a conscious management decision to pursue that specific direction of research, no matter how many resources it would require - and mere scientific curiosity doesn't strike me as sufficient of a reason to explain the expenses.

Sure, but that's why the GP said "curiosity of researchers" and "the greed of corporations". They didn't claim it was mere scientific curiosity so this criticism of their argument does not hold.

> Expectations of profit would, but if that was the driving factor, then it's weird how haphazardly the whole sector goes about monetization strategies.

The fact that people are bad at monetizing it doesn't negate the fact that expectations of future profit were a driving factor.

[tim333]

If you listen to most of them the use case is AGI. Once you get there the AGI robots can develop better robots exponentially and so produce an almost infinite amount of stuff. Subject to resources but still a lot.

[6510]

If you have 10m you buy a boat, 100m an airplane, with friends you spend 10m on a party which is nothing? What fun products are there if the budget is many billions? The ones in endless pursuit of simply more money are mentally ill which is hard to ignore for the rest of the club. They try some immortality but then you have to stop living the good life. May try some mars colonization but it turns out the space bicycle isn't quite ready.

They do have to do something or else the reality catches up with them. The reality that you don't need that much wealth and that it might have been better if you took less from others. This post purchase rationalization also requires training your psychopathy (you cant let peoples suffering get to you or you have to pay to help them. You need to learn to ignore it.) Witch fits perfectly with the project described in the article.