What the heck is going on at Apple?

>as critics say Apple, once a tech leader, is behind in the next big wave: artificial intelligence.

Most critics I see deal with the fact that they’re fad chasing and delivering without their flagship polish (for both new products and updates). This narrative is likely to push apple deeper into the well if it becomes the mainstream spin.

As someone who has Apple-everything (just about) and has since 2013 or so, every time I see a headline about Apple being in serious trouble over their “AI failure” I can’t even understand what they’re talking about. Nothing I’ve seen yet is compelling enough I need it in my OS. People can reach Chatgpt on the iPhones just fine. Who cares? Who are these people like “idk might switch to Windows because of Apple’s failed AI strategy” making this an actual problem? I’ve never even understood what their supposed strategy was trying to do (at which it evidently failed? How did it fail? I also don’t follow that.)

Literally every single person in my life is using AI on an almost daily basis. Not just like my co-workers, or some nerdy discord chat, like every single person I regularly interact with from my niece to my mother, my boss to my barista.

People complain about it, it's short falls and idiosyncrasies, but it's only been getting better, both the models and the integration.

There is no future now where LLMs aren't playing a big role. We'll have our CLI luddites who believe computing peaked in 1992 forever, but the rest of society is running full speed towards computers that they can talk to in natural language.

That's why Apple is uneasy. The god-tier technology usability company is on the verge of totally missing out on the greatest revolution in human-computer usability ever. My mother isn't going to want an Apple UI anymore when you just talk to the new computers.

What are they doing with AI besides an improved search if they aren’t developers?

According to OpenAI, ~4% of tokens generated are for software.[1]

LLMs are not a tech tool. SWE is practically a fringe use of LLMs.

See page 16: [1]https://www.nber.org/system/files/working_papers/w34255/w342...