>>11777105Because goal alignment is the major AI safety concern. Not methods of achieving gradient descent.
>>11777062This isn't a sci-fi trope anon. This is a near future reality. Neurolink has made huge inroads into brain-machine IO interfaces. Once that hurdle is resolved all that's left is integration.
>>11777069We seriously, to this date, have no idea what parts of the brain hold consciousness and to what extent. Furthermore if you slap in a machine onto the brain how do you know which part took primacy? Machines can process information about 1000x faster than chemical/electrical brain signals. As the machine part expands in power, will the flesh part even matter?
These are fundamental questions.
