In hindsight, the answer here was "buy NVIDIA call options"
There was this idea, it may or may not be feasible to predict which companies will lose value due to AI though
Earlier this year, I posted on the rather inactive invite "EA/LW Investing" Discord board about various stocks that I thought would either benefit from or be negatively impacted by AI. I even systematically looked outside the U.S. for ideas. This long/short portfolio has done great this year. Now that it's later in the game and there's been a lot of hype, it might make sense to consider second-order effects, focusing on companies that could benefit indirectly on the long side.
It seems pretty obvious to me, and probably to many other people in the rationalist community, that if AGI goes well, every business that does not control AI or play a role in its production will become virtually worthless. Companies that have no hope of this are obviously overvalued, and those that might are probably undervalued (at least as a group).
And if it goes poorly then it seems unlikely anyone will be able to preserve their financial assets.
The traditional way to express this kind of view would be to buy deep out of the money call options on a basket of stocks of companies that will likely benefit from the AI revolution, i.e. Google, Microsoft, NVIDIA, TSMC and perhaps a few others.
It looks like longest dated call options you can buy are LEAPS which have an expiration data of up to 39 months. I think the odds of AGI happening by that time are low but not zero. I'd prefer something with a longer expiry date.
I've resorted to just buying a bunch of stock in companies like Google, Microsoft, NVIDIA, TSMC and others. But I'm frustrated that I can't better take advantage of everyone else's pigheadedness than by just making the same play as someone who is excited about CUDA.
Is there any way to make a new financial instrument like Michael Burry did in 2006 when he created credit default swaps on CDOs? If anyone knows how to do this for calls on AI stocks, please get in touch.
I think in the long run the only thing that will really matter is control over intelligence and resources, whether that comes through having control over a very powerful AI assistant and a lot of compute available, a lot of money to buy access to these things, or by literally being a superintelligence (likely via AGI-enabled brain uploading).
It's very hard to know what if anything of what we own today can be preserved in a post-singularity world. But it seems plausible that money and resources will give one an advantage in the transition period that could last far beyond the creation of AGI. So it still seems like a sensible play to me to try to maximize investment value.
Does anyone have suggestions as to what financial instruments would help me best express my views on the role of AI in the future? Ideally, it would be something like a bunch of calls with expiration dates in the late 2030s.
The obvious correct move for humans to make right now would be to pause almost all work on making AI better and focus on alignment research. While most of the talented programmers and mathematicians are working on that, others should probably figure out how to ensure an AGI future doesn't immediately collapse into a permanent lightcone-spanning dictatorship enabled by winner-take-all dynamics.
I think there's some possibility we actually do this at a meaningful scale, particularly if there's some kind of pre-AGI disaster that kills a lot of people. So I'd also like to hedge against the possibility of us getting our act together.
I don't have any clear idea of how this should be done in a financial sense. My current career plan is keyed around human genetic augmentation, which will have a very large impact in worlds with long timelines. So in that sense I am betting heavily on AGI either taking much longer than people realize or the world getting its act together and pausing development of this tech.
But financially, I have no better plan than "keep half my savings in boring index funds that will do fine without AGI". But if anyone else has better ideas about how to hedge against the possibility of an AI pause or AGI taking longer than expected I'm all ears.