@robinsonmeyer@heatmap_news Do you think AI is overdiscussed as part of the problem? I see way more coverage of it in mainstream media relative to the other issues, often citing to utility systems that are outliers in terms of data center concentration.
This really isn’t that useful, though? Textiles is a huge industry that demands systemic, global regulation. Nutrition labels exist because people have different nutritional needs, but people don’t have different textile sustainability needs.
Energy will soon become the biggest bottleneck in AI.
Gigawatt datacenters are already being built; larger ones are on the way; countries will struggle to produce enough power to satisfy demand.
Distributed training will become increasingly important, helping balance consumption across global supply.
I don't wanna rag too hard on the episode, there's plenty of good stuff in there, it's just not that likely that AI will really expand energy demand that much.
Rare L from @tracyalloway and @TheStalwart on today's Odd Lots. Framing the data center energy issue around AI misses what's really going on. A quick thread:
It’s time for our first #EnergyBar of the year! Join us on Feb. 1 at Greentown Boston for our signature networking event feat. startups and the climate innovation ecosystem: https://t.co/7Msq91P4OL
The first person who can figure out how to build a
1) daycare
2) remote co-working space
3) coffee shop
into a single building with a subscription... will become a billionaire in 2024.