Not at all. Most builders are clueless. They think that because they have money, they can solve any problems.

Edit: Adding that site selection is HARD. Every developer with a spare farm plot wants in on the action. Most are completely inappropriate.

Agreed. Oai, Anthro, etc don’t necessarily have paths forward. But a LOT of folks want their hands on the results. Perhaps real open sw will rise to put them under and no one will need to pay for the models. But they will have to pay for the hardware to run it. And that’s pricey. And we get paid up front.

There are more than a dozen datacenters within 30 minutes of my home.

You ask “why believe me?” I ask " why take you seriously?" And we get nowhere.

The popular opinion is framed in view of the worst offenders. That gets clicks. My daily experience shows me that:

• one urinal in my building uses more water in a day than the datacenter does in a year.
• our generators run at less than 65 dB at full tilt at a distance of 30 feet. They are only run during utility outages - and monthly tests for 10 minutes.
• we build in corporate parks and remote sites with surplus power.
• the power substation we paid for and built delivers 20 MW back to the community. We added capacity to the local area on our dime.

However, others are not like us. Musk runs gas turbines for primary power. MS uses open- loop cooling. They get the headlines. We don’t. Our business is based on responsible behavior. But we get painted with the same brush.

Going a step further, the builders that build with environmental and community concerns first are more likely to endure. My company does exactly that.

If we are not going to have honest discussion, why participate? Lemmy should be better.

Operations. DC manager.

Agreed.

As someone on the inside, I can assure you that the buyers put down money up front for a specified product and delivery date. There is no speculation involved. They line up to give us money. Unless Microsoft et al. goes bankrupt, we keep building.

Those results are still tightly controlled by NDAs. They need time for FDA approval, patents, etc. Then time to get to market. These are not released in scientific papers anymore. It’s protected because it cost so much up front.

AI is not just LLMs. Much of AI is machine control, and bioinformatics.llms are the foam on top of the ocean.

The PC became mainstream in 1981. Why didn’t the Internet become big for another 10 years?

How much is the cure/treatment for Parkinson’s worth? Or other intractable medical problems?

What about a device able to manipulate the containment field of a fusion reactor?

These are the reasons why everyone is buying the compute services - there are dozens of moonshot-grade problems that are suddenly possible to attempt. Many will fail, but whoever gets that one success will gain astronomical benefit.