Like the turn of the millennium, I think we’re in a bubble. Money is pouring in, chasing AGI. The bots are impressive, but the ROI does not meet the hype. I think we’re going to have an AI bust that will rival the Dot-com bust.

2001 came, and companies like Pets.com, GeoCities, and eToys went bankrupt. But the internet didn’t become irrelevant, right? Of course not. Trillions have been made since then on the internet. The same will happen with AI. The long game will play out big. In 10-20 years, some of the MAG 7 will be irrelevant. But there will also be a few monstrous companies that are just a couple of guys today. Just as we had a few big winners in internet search and social media, there will be a handful of big winners in the AGI space.

Put yourself in 2001 as an internet company and ask, “What would you have done?” You could have become the expert on SEO on the few platforms that survived. Or, you could have found a niche product within e-commerce and focused on it with old-fashioned business principles on profitability.

What does that look like now if we do, in fact, have a bubble that’s about to burst? Small, focused, trained ML models and narrowly integrated AI. Instead of trying to build a robot butler that washes the dishes, folds the laundry, and waters the plants, we need a pool pump that can take in sensor data and weather forecasts and automatically adjust the chemical balance preemptively to prevent algae bloom. Instead of self-driving cars, we need boats that can predict the next engine fault before you get to the lake so the dealer is ready with the part before Memorial Day weekend. Excavators that can optimize their operation to maintain total uptime.

  • We’re in a bubble like the Dot-com era.
  • Just as the Dot-com bust did not kill the internet, this bust will not kill AI.
  • The big promises are overhyped.
  • We little guys can win big by chasing the small, incremental promise