This combined with locally runnable models getting pretty good recently (e.g. Qwen 3.6) tells me that it's time to seriously consider local dev setup again
Besides the cost you get the control, transparency and ability to identify small language models or LoRA you want to serve even more cost effective.
This should become the new Apple's hardware and software play. I am hopeful about the new CEO
Nothing new about that play. They have been heading in this direction for a very long time now.
Perhaps they would have made the basic spell-checker work on MacOS apple silicon then in this long time?