Why is Apple struggling so hard with LLMs? I get that they're a hardware company, but surely that can't be the full story. They must be suffering from some case of extreme perfectionism or something like that. I would assume it'd be very difficult to get an LLM to perform consistently and reliably without making it useless.
They refuse to pony up for competent talent in ML. Meta, OpenAI, and others out-compensate Apple for ML talent by a factor of 2-4x.
Especially after China has entered the game and DeepSeek disintegrating most of the moat that existed around those companies.
And Apple can simply use what others created instead of burning billions into R&D which might not come up with anything noteworthy.