[ben_w]

> The mind-reading stuff is still a fantasy and will be so also in the future.

Fantasy today, sure — judging by advertising categories I was getting placed in at least a few years back, the AI don't generally have the ability to infer much from revealed preferences, despite the odd headline every so often.

But in the future?

Even ignoring the possibility that the aforementioned advertising-AI get better, there's Neuralink and whatever competition it ends up getting.

[jruohonen]

> Even ignoring the possibility that the aforementioned advertising-AI get better, there's Neuralink and whatever competition it ends up getting.

I am skeptical. AI, Neuralink, or whatever might infer better on many things, sure, but these certainly will not be able to infer what non-dogmatic people think about, say, philosophy or politics already because humans have always a capability to change their minds, while AI is terribly bad at reorienting itself.

[ben_w]

Wires on synapses, however, are very good at sensing exactly what's up.

They're too expensive right now, and I really hope Neuralink turns out to be as much of a business disaster as buying Twitter, because if it's as much of a world-changer as Tesla or SpaceX…

…well, if cheap-and-good BCI comes, there's too many dystopian ways for it to be abused for comfort.

[jruohonen]

> ... well, if cheap-and-good BCI comes, there's too many dystopian ways for it to be abused for comfort.

I am all for dystopia; already with VR/XR/etc. you can probably infer a lot, especially once they start (or have done already?) scanning retina movements and such. But still: stimuli is a one thing and mind-reading (thought control?) is another.

[ben_w]

BCI is literal mind-reading: when the thought goes ping, it shows up on the wire.

The current limitation is that the number of sense-wires we can put on a BCI implant is a factor of ~2^30 smaller than the number of synapses in a human brain.