A couple days after the WWDC keynote, starting to see a couple cracks in Apple’s AI strategy. I’ll blog more later. But I think the architecture for small on-device models and larger models in the cloud that are extremely locked down is very smart. This is flexible and can scale, if users accept it.

@manton I'll accept it.

@manton …and a gaping hole in their strategy to reach net zero by 2030. Or is that OpenAI’s problem? 🙄

@paulrobertlloyd How so? They said their own data centers are all 100% renewable energy. I don't count OpenAI, just as I wouldn't count Google's servers just because it's the default search engine. (Although I do think everyone's servers need to be all renewable as soon as possible.)

@numericcitizen Me too. It muddles things so we'll have to see if normal users understand how it all fits together (not that they really should need to).

@manton I Agree. Now, I wonder if this relatively new technology, together with other relatively new technologies will scorch the planet. They are useful, yes, but are we, or rather they, (that live in lands close to the equator), willing to pay the price for this useful, but not necessary technology? I don’t know the whole picture energy-wise, but I know the planet keeps warming at an accelerating rate, and I’m a bit worried and hesitant.

@manton All design decisions have consequences. Pushing requests to ChatGPT means computation on Microsoft’s servers, and famously their climate targets have gone out the window. Then there’s the fact that only newer Apple devices can run the local models, further encouraging hardware upgrades.

@manton I’d encourage you to read this post about the Jevons paradox which explains how energy efficiency gains can often lead to more energy usage rather than less. This seems like a case in point.

@paulrobertlloyd oh nice this is very good

@paulrobertlloyd when I say good I mean bad but good to continue to highlight this very legit concern and it’s not just energy use but also water use