In the dotcom era, I had friends working at e-commerce startups selling items you could easily find at a store. They even had to buy from the same wholesale suppliers, and try to undersell retail, even though they had additional shipping cost (offset a little by not having to pay local sales tax). So they ate the losses because VCs told them they had to show the only metric was positive customer growth (not profit). All business ideas were “add e-commerce to X.”
In the 2008 crash, even though it was triggered by real-estate debt, a lot of the same tech dynamics were at play, except “add mobile to X.”
A lot of present day AI companies are following the same path. “Add AI to X.”
What’s different this time is that there’s a lot more hardware involved, in the form of GPU and data center expansion. After dotcom, we were left over with a lot of fiber, telco, and home internet expansion which was still usable. This time, it’s not clear what the data centers will be good for if AI crashes out. Maybe crypto-mining.
it’s not clear what the data centers will be good for if AI crashes out. Maybe crypto-mining.
not very useful for society neither
utility companies are also taking a risk, many people isn’t aware how expensive is to build the infrastructure to supply electric power to the new data centers
The unbridled enthusiasm is the same.
In the dotcom era, I had friends working at e-commerce startups selling items you could easily find at a store. They even had to buy from the same wholesale suppliers, and try to undersell retail, even though they had additional shipping cost (offset a little by not having to pay local sales tax). So they ate the losses because VCs told them they had to show the only metric was positive customer growth (not profit). All business ideas were “add e-commerce to X.”
In the 2008 crash, even though it was triggered by real-estate debt, a lot of the same tech dynamics were at play, except “add mobile to X.”
A lot of present day AI companies are following the same path. “Add AI to X.”
What’s different this time is that there’s a lot more hardware involved, in the form of GPU and data center expansion. After dotcom, we were left over with a lot of fiber, telco, and home internet expansion which was still usable. This time, it’s not clear what the data centers will be good for if AI crashes out. Maybe crypto-mining.
not very useful for society neither
utility companies are also taking a risk, many people isn’t aware how expensive is to build the infrastructure to supply electric power to the new data centers