[blazespin]

An AI winter will not happen. We are past the point of no return. The vast majority of research is net positive research.

The only risk is misappropriation of investment, but that has already been smoothed over quickly.

I’d get the complaint if AI was not doing anything useful as it wasn’t in ‘74, but it simply is.

For example, plate recognition, drones and facial recognition. These are brutally efficient technologies. To imagine any government under investing in these is a total misunderstanding of the value that AI brings to military, security, and economic dominance.

It’s a race between nations and the only danger is the computers taking over (not in a general AI way, but in a way that everything is run by algorithms no human can understand)

[not_a_moth]

You sure?

Without a doubt, there's an investment bubble for AI companies in industry: a huge amount of companies have added "AI" to their proposition, where the AI is only marginally useful if not useful at all, for the purpose of marketing and inflating valuation.

On research side, it's now clear the current AI paradigm will not produce the kind of massive, society-shifting promises that were made to investors, which is the focus of the article.

Yes deep learning is here to stay for narrow tasks, but the investment contraction for the rest could certainly feel like a Winter.

[marcosdumay]

> it's now clear the current AI paradigm will not produce the kind of massive, society-shifting promises that were made to investors

It's correct in that it won't produce the society-shifting promises. But there is a huge amount of money on non-society-shifting innovation.

Besides, I really doubt the media predictions are what actually was told to investors.

[mbruce]

This is primarily deep learning which is not general purpose AI

[laichzeit0]

We’re talking about AI here. And deep learning is AI. If you go to university and research or study AI that’s what you’d currently be learning. AGI does not currently exist and is a word used mostly by laymen and the media who don’t understand the current state of AI.