Even the most sophisticated purpose made medical research tools can't do this in a way that you are imagining. If somehow Apple accomplished mind reading devices in their spare time I'd be so amazed that I'd actually not even care how invasive it is. At least for awhile.
Maybe not today, but with better sensors and neural networks I'd say anything is possible.
That's not what's bothering me most in the long run though. Because there's no way in hell they'll be able to resist the temptation to manipulate our brains directly if they ever develop the capability.
Technology is neither good nor bad, but it enables better and worse with every iteration.
I studied biomedical engineering, it's really difficult to get anything useful out of a human brain non-invasively even if you are Apple. Neural networks are not magic, they can't do much with such low SNR.
LLMs are more an exanoke of punctuated that smooth, even exponential, when looked at vs. 10 years ago, but there’s a difference between no one would expect b/c we don’t know how, and there being a clear articulable reason why not.
patents don't mean something is remotely close to existing. A lot of patents are purely speculative, to either fish for future licensing revenues or to protect the ability to enter a market in the future
