Yeah that’s it. If it ran locally it wouldn’t regularly show ‘rate limit exceeded’ messages. Its happening because it’s running server side in meta land.
One thing it sure as hell doesn’t encrypt is the links in your messages. They’re clearly sent to Meta to be ‘unfurled’. You can tell because sometimes a zoom like will unfurl to ‘too many requests try later’ so it’s obviously being done by some massive bot.
Nvidia seem to be trying to rebrand them as ‘AI factories’, because factories means jobs. Only once the data center is built it employs maybe a dozen full time staff. There just isn’t any work to do, especially in a dc that is just for one massive client with racks and racks if identical equipment.
It’s not that they don’t technically work. It’s just they’re no longer efficient compared to newer versions that can do more with less power. So to remain competitive you need to upgrade otherwise your cost to execute a model is too high.
Hyperscalers used to write GPU’s down to zero value after three years, over the last couple of years they’ve all increased this to six.
The trouble with the railways comparison is that after investing tons of cash the railways were built. With AI the GPUs have no value after 6 years (if that). So the investment must continue forever. It’s madness.
International business class still comes with metal silverware and proper plates today. First class might have a mother of pearl spoon to go with your caviar, if it’s fancy enough.
No one is rolling around a ham to carve next to your seat but international first class can still be plenty fancy on the few airlines that still do it.
Same way they treat everything else, it’s the only way they know how