Ep. 147 Energy and AI 能源与人工智能（英文播客）

Today, we're excited to present a special cross-over episode with the Podcast "Preparing for AI" hosted by Matt Cartwright BEM 柯明龙 and Jimmy Rhodes.
The guest for this episode is Anders Hove, long-time host of the Environment China podcast, and Beijing-based Matt Cartwright interviews him about the impact of AI on the clean energy transition. Anders ‘references several points related to China, such as efforts to locate data centers in colder regions, and the results of experiments seeking to make data centers respond to the output of renewables. We covered quite a bit of ground and hope you enjoy it!
- How much energy AI is using and how much it will use - and how that compares with renewable energy production now and in the future.
- What technologies AI could help develop or optimize, and their importance (or lack thereof) compared to basic policy incentives for scaling up clean energy that already exists.
- How different AI functions (text, image, video) have dramatically different implications for energy use, so it's hard to predict how much energy it will use since AI isn't one thing.
- How poor disclosure from data centers and AI companies hurt sustainability. Ironically, while AI boosters claim AI will help optimize renewable energy and weather forecasts, often relying on public data, AI could be much more helpful by just predicting its OWN energy use and how to optimize that -- but companies treat this as a secret, probably because they don't know themselves.