These are AI’s that need to be Online and their effectiveness comes from huge computing power on datacenters. Apple at this point was going for an on device mobile hardware AI.
According to the latest reports, Apple was also considering cloud AI services. Siri today uses Apple’s servers for plenty of things, so this is clearly not a hard requirement for them.
Regardless, Apple needs to offer competitive services, no matter how they chose to implement them. This is too big of an inflection point for them to sit by twiddling their thumbs for another few years until these models can run on-device.
Apple has invested as much money as any other on the field
That does not seem to be the case. Microsoft, Meta, and Google are clearly ahead of Apple in AI research. Or if Apple has spent equivalent money, it’s clearly being used very inefficiently.
If the only people who can dictate these rules are the companies that actively creating and have the majority of the market then this would be only OpenAI. But a lot of companies invested money on OpenAI, so should not they be able to pitch in?
I suppose the more salient point is that Cook has no leverage. Apple’s big, but they don’t have a meaningful presence in generative AI, so they can’t lead by example. Or in other words, they’d be setting rules to restrict others, not really themselves. And without their own competitive offerings, they have a perverse incentive to artificially restrict AI development to diminish the competitions’ advantage. To this day, that’s an active strategy they employ for e.g. web apps, so it’s difficult to believe anything they propose today is in good faith.
According to the latest reports, Apple was also considering cloud AI services. Siri today uses Apple’s servers for plenty of things, so this is clearly not a hard requirement for them.
Regardless, Apple needs to offer competitive services, no matter how they chose to implement them. This is too big of an inflection point for them to sit by twiddling their thumbs for another few years until these models can run on-device.
That does not seem to be the case. Microsoft, Meta, and Google are clearly ahead of Apple in AI research. Or if Apple has spent equivalent money, it’s clearly being used very inefficiently.
I suppose the more salient point is that Cook has no leverage. Apple’s big, but they don’t have a meaningful presence in generative AI, so they can’t lead by example. Or in other words, they’d be setting rules to restrict others, not really themselves. And without their own competitive offerings, they have a perverse incentive to artificially restrict AI development to diminish the competitions’ advantage. To this day, that’s an active strategy they employ for e.g. web apps, so it’s difficult to believe anything they propose today is in good faith.