[Yizahi]

Let's imagine for a second that this a few billion dollars per year to Google is correct. Why do you assume that it covers everything to be done by Google itself - from hosting to running actual servers? Apple may very well pay Google a licensing fee, take a trained LLM and run inference themselves locally or even at a yet another 3rd party for example a datacenter corporation or any mix of these. And then a true real cost of running just the inference on every Apple device would be separated into a completely different org payment flows, very obscured and higher than just a license fee.

I'm not saying that this is what really happens. I'm saying that believing a CEO is as foolish and as grounded in reality as believing Ed Zitron.

[dofm]

> Why do you assume that it covers everything to be done by Google itself - from hosting to running actual servers?

I don't, and that's the point, isn't it?

It's the keys to a substantial chunk of the kingdom for $1B a year. Literally they are getting, for a very small price, the right to distill their own models from Gemini.

Is there money in this for someone with a data centre? Possibly. Is there money in it for NVIDIA? Possibly.

But either way, that's not OpenAI or Anthropic, is it?

[aix1]

> It's the keys to a substantial chunk of the kingdom for $1B a year. Literally they are getting, for a very small price, the right to distill their own models from Gemini.

Here is a different interpretation: Apple bought the rights to distill and use a smaller version of one unspecified model in the Gemini family (there are many such models).

The distillation will be carried out at Google's data centres so that the original weights never leave Google premises.

For this to be keys to be kingdom it would need to cover all current and future models and would need to be very permissive with regards to distillation parameters and allowed uses of the distilled model.

I expect the reality to be somewhere between these two extremes.