These “makeshift” structures are housing hardware that costs millions of dollars in total.
“Putting AI servers inside tents, officially called “rapid deployment structures,” is one of the more unique approaches to the AI build-out, Thomas said. They’re certainly not as sturdy as physical buildings made from steel and concrete, with one commenter comparing it to the “classic $10k racing bike with a $9 lock” situation.”
Please stop using AI, please.
If you’re required to use AI for your job, then sabotage the efforts.
I beg of you.
I have been, by focusing on enabling team members to use it exactly like how leadership wants. You want them to use more AI? Okay… they need Agentic harnesses that can do work for them locally. They need MacBook Pros to run the models. They need cloud keys to test different frontier models for different loads. They need governance, observability, repeatability, scheduling, human-in-the-loop…
I’m going to show them that anyone can build a bridge, but only an engineer can build a bridge that barely works. On top of that, I’m going to show them that they’re wrong in believing they want a bridge. All it should take is seeing that they got exactly what they wanted without getting anything that they wanted.
Pure wishful thinking like this is just as useful as thoughts and prayers.
This is a overhyped technology just like the internet was before the dotcom bubble popped.
The best you can hope for is for the AI bubble to burst and then see what AI is like when it normalises.
Because to think this technology will just go away is going to end the same way as for tge people who said the internet is just a temporary fad.
When the dot com bubble bursted, the typical internet speed was 56k modem speed and cellular internet practically didn’t exist. At that point the internet still needed years of exponential technological advancement to allow for stuff like streaming amd mobile services. The difference with the AI bubble is that they try to brute force their way out of infancy by throwing ludicrous amounts of money, energy and other recources at it instead of waiting for the much needed technological progress
wouldn’t it be easier to sabotage tents with data farms in them
I’ve been cruising with the most expensive model at work for a while now. After github’s pricing model change they finally asked us to be more conscious about which model we’re using, but eh, it’s easier to just leave it on Claude 4.8. I figure eventually the costs will catch up with the company.