Note: This isn’t the place to debate/convince others of your position on AI winter. There are plenty of threads on Lesswrong for that already.
There are already a few posts on how to profit from rapid AI progress in the near future. I figured it would be worthwhile to also discuss the opposite scenario—a third AI winter, if nothing else to hedge against it.
I don’t expect many people to agree with my pessimistic view on future AI progress, especially considering the recent hype around GPT-3 and AlphaFold. But if you’re predicting another AI winter, it’s precisely the best time to take a bet, namely when expectations are high.
However, I find it quite difficult to operationalize such a position. Disentangling deep learning from the rest of tech is hard. It doesn’t make sense to short Nvidia stock since their main business is still gaming hardware. Tesla may well succeed in the electric cars/solar energy industries even without full self-driving. Alphabet, Microsoft and Facebook all could go on just fine if deep learning turns out to be a mirage.
That’s why I’ve come here for advice. I have intentionally left out any precise definition of ‘AI winter’ because I don’t want to rule out any potential ideas that only fit some scenarios. Any input would be greatly appreciated.