I was around then. This feels similar from the perspective of stock prices just going up for absolutely absurd reasons—Allbirds would have fit right in—but AI in some ways seems even worse because much of what was being hawked in the dotcom era was at least something that real people would want to buy: usually poorly executed and unsustainably priced but e.g. Kozmo.com wasn't wrong that a ton of people would love to order stuff and have it brought to your home or office. I had some business meetings with pets.com and they were unbelievably bloated—think a room full of people openly admitting that they were just killing time waiting for the IPO so they could cash out their options—but tons of people buy pet food, toys, and medicine online today.

In contrast, a lot of companies are just wedging AI into things where ordinary people don't want or benefit from it anywhere near the level needed to actually pay for it, especially with unreliability being an unsolved problem—the amount you'd pay for something to, say, summarize boring business documents goes down considerably when you have to closely read the original looking for errors and omissions and scrutinize any response to make sure it's not critically flawed.

There's also a weird difference in enthusiasm based on the societal impacts: the web let you do things you couldn't do before, but a lot of AI tools and services are really oriented at your boss’ boss’ boss’ boss’ boss saying they can layoff half of your department and still get more work out of the survivors. There's some cool stuff, yes, but unlike the mood during the dotcom era we now have a prominent tone of dread about deprofessionalization and larger concerns about how businesses will even work if a significant chunk of customers are pushed out of stable employment.