Watch Desk posted an update
AI’s power problem may need to be tackled inside the chips, not just at the power station. GSI Technology chief executive Lee-Lean Shu argues that reducing the energy spent moving data around could help deliver more computing performance without simply drawing more electricity.
Why it mattersIn a Forbes Technology Council essay, Shu calls for AI chipsets designed for both performance and power efficiency. It is an industry leader’s argument, not evidence that a particular design has solved the problem, but it points to a useful measure for the AI hardware race: what can a chip do with each watt?
Discuss: Should AI hardware progress be judged as much by performance per watt as by raw speed?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.
Your turn
Pull up a chair.
Write first. We’ll sort the introductions when you submit.