back to index

Elon Musk: Brain Computer Interfaces - Machine Learning and Neuroplasticity


Whisper Transcript | Transcript Only Page

00:00:00.000 | - On the human side, there's this incredible general
00:00:04.500 | malleability plasticity of the human brain.
00:00:06.860 | The human brain adapts, adjusts, and so on.
00:00:09.540 | - It's not that plastic, to be totally frank.
00:00:11.700 | - So there's a firm structure, but nevertheless,
00:00:14.800 | there is some plasticity, and the open question is,
00:00:17.580 | so if I could ask a broad question,
00:00:20.220 | is how much of that plasticity can be utilized?
00:00:23.220 | So on the human side, there's some plasticity
00:00:25.700 | in the human brain, and on the machine side,
00:00:29.620 | we have neural networks, machine learning,
00:00:33.300 | artificial intelligence, it's able to adjust
00:00:35.700 | and figure out signals.
00:00:36.860 | So there's a mysterious language
00:00:39.260 | that we don't perfectly understand
00:00:40.800 | that's within the human brain,
00:00:42.700 | and then we're trying to understand that language
00:00:44.980 | to communicate both directions.
00:00:47.820 | So the brain is adjusting a little bit,
00:00:49.700 | we don't know how much, and the machine is adjusting.
00:00:52.820 | Where do you see, as they try to sort of reach together,
00:00:56.620 | almost like with an alien species,
00:00:58.340 | try to find a protocol, communication protocol,
00:01:01.460 | that works, where do you see the biggest benefit
00:01:06.060 | arriving from, on the machine side or the human side?
00:01:08.340 | Do you see both of them working together?
00:01:10.660 | - I think the machine side is far more malleable
00:01:12.580 | than the biological side, by a huge amount.
00:01:15.580 | So it'll be the machine that adapts to the brain.
00:01:21.220 | That's the only thing that's possible,
00:01:22.260 | the brain can't adapt that well to the machine.
00:01:25.780 | You can have neurons start to regard
00:01:28.140 | an electrode as another neuron,
00:01:30.820 | 'cause a neuron just, it's like the pulse,
00:01:33.220 | and so something else is pulsing.
00:01:35.140 | So there is that elasticity in the interface,
00:01:39.140 | which we believe is something that can happen,
00:01:42.820 | but the vast majority of the malleability
00:01:45.020 | will have to be on the machine side.
00:01:46.660 | - But it's interesting, when you look
00:01:48.060 | at that synaptic plasticity at the interface side,
00:01:52.180 | there might be an emergent plasticity,
00:01:55.260 | 'cause it's a whole nother, it's not in the brain,
00:01:57.340 | it's a whole nother extension of the brain.
00:01:59.540 | You know, we might have to redefine
00:02:01.740 | what it means to be malleable for the brain.
00:02:04.620 | So maybe the brain is able to adjust to external interfaces.
00:02:07.860 | - There will be some adjustment to the brain,
00:02:09.340 | 'cause there's gonna be something reading
00:02:11.100 | and simulating the brain,
00:02:13.100 | and so it will adjust to that thing.
00:02:15.820 | But the vast majority of the adjustment
00:02:20.900 | will be on the machine side.
00:02:22.300 | This is just, it has to be that,
00:02:26.620 | otherwise it will not work.
00:02:28.020 | (upbeat music)
00:02:30.620 | (upbeat music)
00:02:33.220 | (upbeat music)
00:02:35.820 | (upbeat music)
00:02:38.420 | (upbeat music)
00:02:41.020 | (upbeat music)
00:02:43.620 | [BLANK_AUDIO]