It’s a real concern that the best Apple can do is use OpenAI instead of its own technology.
This strategy makes a lot of sense if you believe that progress will plateau at the current level with small incremental improvements in reliability and UX. Apple really only pays a price if failing to develop their own foundation model results in their phones not selling, which doesn't seem to be the case, regardless of how much of an embarrassment Siri is. Right now, Microsoft, OpenAI, Meta, Google, and Grok are all bludgeoning each other with trillions of dollars and creating a perfect situation for Apple to swoop in in a couple years when reliability is sorted out and GPU compute has crashed to rock-bottom prices.
"Providing tokens" is the bottom layer of the value chain. If you can't build a moat around it, it's a race to the bottom because you can only compete on price.
Apple doesn't compete on the bottom of the value chain.
Google must keep up with others because it is ad company. And oh boy the potential of influence with AI models.
Neither of these seem like a shame so much as leveraging SOTA.
That doesn’t mean that Apple shouldn’t be following the space closely but I think it’s a mistake to think that normal people are changing their buying decisions based on any of this yet. One trade off here is that Google is spending a ton of money but as of yet is massively in the red on that investment, which means they’ll stick with their own LLM while Apple is free to switch to whoever the market leader is without a large sunk cost factoring into the decision.
I am very very far from a google fanboi.
My perception is that a huge percentage of the mass market just like OpenAI because they were the first to market and still have the most name recognition. Even my coworker who works in DevOps says “Gemini sucks, Claude sucks” even though he has never once tried either of them and has never looked at a single benchmark comparison.