Comment on Wall Street’s AI Bubble Is Worse Than the 1999 Dot-com Bubble, Warns a Top Economist

• Perspectivist@feddit.uk ⁨11⁩ ⁨hours⁩ ago

  Asking investment advice from a system that’s designed to do nothing else but generate natural sounding language based on probabilities is pretty stupid.

  That being said, what’s wrong with this answer? I think it’s more or less a good and balanced take.

  Here’s the first half of it that I left out:

  spoiler

  >“AI” as an investment isn’t one thing—it’s more like a category of bets, ranging from hardware to software to services, each with wildly different risk profiles. So the honest answer is: yes, it can be a good investment—but only if you understand what you’re actually investing in. > >Here’s why that nuance matters: > >Buying Nvidia stock in 2019 was a good AI investment. Buying it now, after a 10x run-up? Much less clear—it’s priced as if they’re the sole arms dealer in a forever war. > >OpenAI, Anthropic, etc. aren’t publicly traded, so retail investors can’t buy them directly. Instead, you get exposure via companies like Microsoft, Amazon, or other backers—meaning you’re not really investing in “AI” directly, but as part of a much broader bundle. > >AI startups and ETFs are all over the place—some are thinly veiled hype vehicles chasing trends, while others are building real infrastructure (like vector databases, chip design tools, or specialized AI services). Picking the wrong one is like investing in Pets.com during the dot-com boom—it sounds techy, but the business might be garbage. > >Thematic ETFs like BOTZ or ROBO give you AI exposure but are diluted by their attempt to hedge across subsectors. They tend to underperform when compared to cherry-picking the winners.

  source