AI Opportunity Cost

Aug 31st 2025

The main problem of the AI bubble is just a question about opportunity cost. As long as the perceived marginal benefit of investing in AI is larger than investing in anything else, the AI bubble will keep going.

Unfortunately, the perceived marginal benefit is still extremely high, and not just for no reason.

Consider for a moment all of the other AI bubbles we have experienced. Some new algorithm, technique or technology was made; attracted a lot of investment; some products were eventually made; but then the capabilities of the new technology were exhausted and the bubble popped.

Unfortunately, the implication now is different.

Nowadays, everyone is chasing the golden goose. If they own the algorithm, technology, and computing power to actually replace people in the day jobs, they effectively become king of the world. All other problems will be solved, so the marginal benefit of investing in anything other than AI will be close to 0.

So that's why to some extent I kind of agree with the billionaire oligarchs who are spearheading the AI bubble. There is a chance, an actual non-zero chance that an algorithm can recreate the intelligence of the human brain. We don't have that algorithm right now, but in the meantime it makes sense to buy all the computing power you can, because you'll get a much better deal on GPUs before this algorithm is discovered.