1
00:00:00,080 --> 00:00:02,839
Speaker 1: I want you to, well, assuming you aren't driving right now,

2
00:00:02,879 --> 00:00:04,280
I want you to just close your eyes for a second.

3
00:00:04,360 --> 00:00:06,639
Safety first, obviously, right, safety first. But if you can

4
00:00:06,960 --> 00:00:10,080
picture this, you are standing in the middle of a

5
00:00:10,320 --> 00:00:14,679
cavernous industrial hall. It's cool, maybe a little damp, and

6
00:00:14,720 --> 00:00:17,199
the air around you is literally vibrating.

7
00:00:17,239 --> 00:00:19,359
Speaker 2: Oh yeah, you could feel it in your keith exactly.

8
00:00:19,399 --> 00:00:21,120
Speaker 1: You are standing in front of one of the world's

9
00:00:21,120 --> 00:00:26,160
most powerful supercomputers. It is this absolute monolith, a literal

10
00:00:26,399 --> 00:00:31,640
fortress of silicon and steel. And you can hear that hum, right,

11
00:00:31,679 --> 00:00:34,719
that low frequency thrum that you feel in your chest

12
00:00:34,799 --> 00:00:36,600
way more than you hear it with your ears.

13
00:00:36,640 --> 00:00:39,320
Speaker 2: It's an industrial roar. And you really have to imagine

14
00:00:39,359 --> 00:00:42,079
the sheer amount of heat coming off it too. Because

15
00:00:42,240 --> 00:00:46,280
these facilities, the ones running the massive AI models or

16
00:00:46,399 --> 00:00:49,759
doing the deep climate change simulations, they generate so much

17
00:00:49,799 --> 00:00:52,399
waste heat that they don't just have fans. They have

18
00:00:52,560 --> 00:00:57,000
dedicated cooling towers, huge ones. Massive. We are talking about

19
00:00:57,039 --> 00:01:00,799
literal rivers of water being pumped through the infrastructure constantly

20
00:01:01,320 --> 00:01:03,560
just to keep the circuitry from melting down into a

21
00:01:03,560 --> 00:01:05,079
puddle of toxic slag.

22
00:01:05,239 --> 00:01:09,439
Speaker 1: Yeah, it's this massive incredibly energy hungry beast. I mean,

23
00:01:09,560 --> 00:01:13,519
it consumes enough electricity to power a small city. We

24
00:01:13,560 --> 00:01:17,920
are talking megawatts upon megawatts of power, and it's doing

25
00:01:18,159 --> 00:01:21,920
all of that just to process linear logic. Just zeros

26
00:01:21,920 --> 00:01:23,760
and ones, flipping back and forth.

27
00:01:23,480 --> 00:01:26,200
Speaker 2: Flipping very fast. But yes, just zeros and ones.

28
00:01:26,280 --> 00:01:29,760
Speaker 1: Right now, open your eyes and think about what's happening

29
00:01:29,799 --> 00:01:33,000
inside your own head. Right now you are listening to us,

30
00:01:33,000 --> 00:01:36,280
you're visualizing that giant computer room. Maybe you're I don't know,

31
00:01:36,319 --> 00:01:38,319
thinking about what you need to pick up for dinner later,

32
00:01:38,799 --> 00:01:41,359
and you're keeping your heart beating and your lungs breathing

33
00:01:41,400 --> 00:01:42,680
all at the same time.

34
00:01:42,560 --> 00:01:44,959
Speaker 2: And you are doing all of that on about twenty

35
00:01:45,040 --> 00:01:46,280
watts of energy twenty once.

36
00:01:46,319 --> 00:01:49,159
Speaker 1: I mean, that is it's ridiculous when you say it

37
00:01:49,159 --> 00:01:51,680
out loud. That is a dim light bulb. That is

38
00:01:51,799 --> 00:01:54,439
literally the bulb you put in a hallway closet because

39
00:01:54,480 --> 00:01:55,799
it's too dark to read by.

40
00:01:56,000 --> 00:01:59,560
Speaker 2: It is the ultimate efficiency paradox, and it's really the

41
00:01:59,599 --> 00:02:02,840
hook everything we are diving into today. Because if the

42
00:02:02,879 --> 00:02:06,439
brain were just a biological version of that supercomputer.

43
00:02:06,120 --> 00:02:08,759
Speaker 1: Just a wet circuit board, exactly.

44
00:02:08,479 --> 00:02:11,680
Speaker 2: A wet circuit board, if that were true, that efficiency

45
00:02:11,680 --> 00:02:14,240
gap wouldn't make any sense at all. The math simply

46
00:02:14,240 --> 00:02:15,159
doesn't add up.

47
00:02:15,120 --> 00:02:18,439
Speaker 1: Which brings us to the big realization, right, the massive

48
00:02:18,520 --> 00:02:21,800
aha moment that kicks off this whole journey we're taking

49
00:02:21,800 --> 00:02:22,840
you on today.

50
00:02:22,639 --> 00:02:27,800
Speaker 2: Right, because we have spent decades, literally decades, assuming that

51
00:02:27,840 --> 00:02:30,919
the only real difference between a supercomputer and a human

52
00:02:31,000 --> 00:02:34,800
brain is raw processing power. We thought, you know, if

53
00:02:34,840 --> 00:02:36,879
we just make the chips faster, if we just managed

54
00:02:36,919 --> 00:02:39,800
to pack in more transistors, we'll eventually get there. Just

55
00:02:39,840 --> 00:02:42,560
boote force it, yeah, brute force it. But the reason

56
00:02:42,599 --> 00:02:44,879
we haven't been able to simulate the human mind in

57
00:02:44,919 --> 00:02:48,680
real time isn't that our computers are too slow. It's

58
00:02:48,719 --> 00:02:51,560
that they are the wrong type of machine entirely.

59
00:02:51,879 --> 00:02:56,520
Speaker 1: We've been trying to simulate this incredibly fluid probabilistic biological

60
00:02:56,599 --> 00:02:59,719
system using rigid binary switches.

61
00:02:59,280 --> 00:03:03,400
Speaker 2: Exactly, and that failure has created a really interesting opportunity

62
00:03:03,800 --> 00:03:06,159
because if the brain isn't a digital computer but is

63
00:03:06,199 --> 00:03:09,000
actually operating on quantum probabilities.

64
00:03:08,360 --> 00:03:10,240
Speaker 1: Which sounds insane, but we're going to get there.

65
00:03:10,479 --> 00:03:12,240
Speaker 2: We are going to get there. But if it is,

66
00:03:12,639 --> 00:03:14,719
then the only way to simulate it is with the

67
00:03:14,719 --> 00:03:16,479
computer that does the exact same thing.

68
00:03:17,240 --> 00:03:20,840
Speaker 1: Welcome to thrilling threads. Today, we are not just talking

69
00:03:20,879 --> 00:03:24,439
about faster chips or better graphics cards. We are unraveling

70
00:03:24,520 --> 00:03:28,199
the actual physics of thought. We're looking at why silicon

71
00:03:28,319 --> 00:03:31,719
has completely hit a wall and how a massive breakthrough

72
00:03:31,800 --> 00:03:36,159
involving quantum mechanics and specifically some really recent developments with

73
00:03:36,240 --> 00:03:39,719
chips like Google's Willow might finally allow us to build

74
00:03:39,840 --> 00:03:42,360
a true digital twin of the human mind.

75
00:03:42,599 --> 00:03:46,080
Speaker 2: We are going to explore the very real possibility that

76
00:03:46,159 --> 00:03:49,039
we are on the verge of translating the human experience

77
00:03:49,400 --> 00:03:51,159
into the base language of the universe.

78
00:03:51,360 --> 00:03:53,759
Speaker 1: So let's impact this because I think people hear brain

79
00:03:53,800 --> 00:03:56,400
simulation and they immediately think, oh, we must be I mean,

80
00:03:56,400 --> 00:03:58,919
we have all this advanced AI, right, But the actual

81
00:03:59,000 --> 00:04:01,759
scale of what we are trying to cocky is well,

82
00:04:01,759 --> 00:04:03,000
it's mind boggling.

83
00:04:02,719 --> 00:04:05,000
Speaker 2: It really is. And usually when you read a popular

84
00:04:05,039 --> 00:04:08,120
science article about the brain, they hit you with the

85
00:04:08,159 --> 00:04:11,960
headline number, which is eighty six billion neurons.

86
00:04:11,479 --> 00:04:13,560
Speaker 1: Which sounds like a lot. Eighty six billion is a

87
00:04:13,599 --> 00:04:15,719
number I can't really wrap my head around in any

88
00:04:15,800 --> 00:04:16,639
practical sense.

89
00:04:17,160 --> 00:04:19,920
Speaker 2: It is a lot. But here's the thing. In the

90
00:04:19,959 --> 00:04:24,600
world of modern computing, eighty six billion isn't actually terrifying anymore.

91
00:04:25,240 --> 00:04:28,600
We have regular consumer hard drives that can store way

92
00:04:28,639 --> 00:04:31,800
more individual bits than that. If the brain we're just

93
00:04:31,839 --> 00:04:36,000
a collection of eighty six billion simple switches, we probably

94
00:04:36,040 --> 00:04:39,839
could have simulated it a decade ago. The problem is

95
00:04:39,920 --> 00:04:43,439
that counting neurons is misleading. It's a total red.

96
00:04:43,279 --> 00:04:45,720
Speaker 1: Herring because a neuron isn't just a simple switch that's

97
00:04:45,759 --> 00:04:46,560
either on or off.

98
00:04:46,839 --> 00:04:49,800
Speaker 2: No, not at all. The true complexity doesn't come from

99
00:04:49,800 --> 00:04:52,879
the neurons themselves. It comes from the connecto, the connectime, right,

100
00:04:52,920 --> 00:04:55,160
the connections between them. You have to realize each one

101
00:04:55,160 --> 00:04:58,480
of those eighty six billion neurons is connected to thousands

102
00:04:58,519 --> 00:05:02,360
of other neurons. We ate this incredibly dense three dimensional

103
00:05:02,399 --> 00:05:03,480
web of synapses.

104
00:05:04,040 --> 00:05:06,279
Speaker 1: Okay, so if we aren't counting the neurons, what is

105
00:05:06,319 --> 00:05:07,959
the actual number we should be looking at.

106
00:05:08,079 --> 00:05:10,959
Speaker 2: We are counting the distinct connection points the actual scene

107
00:05:10,920 --> 00:05:12,839
apps is where they talk to each other, and that

108
00:05:12,920 --> 00:05:15,560
number is estimated to be around one hundred trillion, one

109
00:05:15,600 --> 00:05:19,360
hundred trillion, one hundred trillion. And here is where the

110
00:05:19,360 --> 00:05:23,199
simulation problem shifts from being just very difficult to functionally

111
00:05:23,199 --> 00:05:27,519
impossible for standard hardware. And to understand why, we have

112
00:05:27,600 --> 00:05:30,079
to talk about something called the von Neumann bottleneck.

113
00:05:30,199 --> 00:05:33,439
Speaker 1: Okay, the von Neumann bottleneck. That sounds like a Robert

114
00:05:33,480 --> 00:05:36,360
Ludlam's spy thriller title, but I know it's a foundational

115
00:05:36,399 --> 00:05:39,800
concept in computer science. Break it down for us. Why

116
00:05:39,879 --> 00:05:41,839
is this the invisible wall we keep hitting?

117
00:05:42,000 --> 00:05:44,720
Speaker 2: So virtually every computer you use on a daily basis,

118
00:05:44,759 --> 00:05:48,000
your phone, your laptop, even those massive supercomputers with the

119
00:05:48,040 --> 00:05:51,079
cooling towers we talked about earlier, they are all built

120
00:05:51,079 --> 00:05:54,240
on the von Neumann architecture. And in this design, the

121
00:05:54,279 --> 00:05:58,160
processing unit the CPU, and the memory the RAM, are

122
00:05:58,199 --> 00:05:59,240
physically separate.

123
00:05:59,360 --> 00:06:01,560
Speaker 1: They live in different neighborhoods on the motherboard.

124
00:06:01,319 --> 00:06:04,879
Speaker 2: Exactly, there are separate physical locations. So every single time

125
00:06:04,959 --> 00:06:07,399
your computer wants to do a calculation, it has to

126
00:06:07,439 --> 00:06:09,839
send a request over to the memory, grab the data

127
00:06:09,879 --> 00:06:12,560
it needs, drag it across a physical connection called a bus,

128
00:06:12,600 --> 00:06:15,480
over to the processor, do the math, and then drag

129
00:06:15,560 --> 00:06:18,160
the answer all the way back to memory to store it.

130
00:06:18,160 --> 00:06:19,959
Speaker 1: It's a commute. It's like having to drive to the

131
00:06:19,959 --> 00:06:22,560
physical library every single time you want to remember your

132
00:06:22,560 --> 00:06:23,439
own phone number.

133
00:06:23,519 --> 00:06:27,120
Speaker 2: That is a perfect analogy. And for linear tasks like

134
00:06:27,199 --> 00:06:30,079
running a spreadsheet or rendering a video, the commute is

135
00:06:30,079 --> 00:06:33,199
fast enough that you don't really notice the lag. But

136
00:06:33,319 --> 00:06:36,199
now apply that commute to the biological brain.

137
00:06:36,360 --> 00:06:39,120
Speaker 1: Right to those one hundred trillion synapses.

138
00:06:38,720 --> 00:06:42,560
Speaker 2: Right in the human brain, memory and processing are co located.

139
00:06:42,600 --> 00:06:45,160
They're the same thing. The synapse itself is the memory

140
00:06:45,160 --> 00:06:47,800
in the processor all in one. When those one hundred

141
00:06:47,800 --> 00:06:51,560
trillion connections fire, they do it simultaneously, in massive parallel

142
00:06:51,839 --> 00:06:54,399
right exactly where the data already lives. There is no.

143
00:06:54,279 --> 00:06:56,120
Speaker 1: Commute zero travel times zero.

144
00:06:56,680 --> 00:06:59,319
Speaker 2: So when we try to simulate that biological process on

145
00:06:59,360 --> 00:07:01,199
a classical star supercomputer.

146
00:07:00,800 --> 00:07:04,040
Speaker 1: We are forcing that machine to simulate one hundred trillion

147
00:07:04,079 --> 00:07:06,279
separate commutes every single millisecond.

148
00:07:06,360 --> 00:07:09,319
Speaker 2: Yes, the traffic on the computer's data bus becomes so

149
00:07:09,439 --> 00:07:13,879
incredibly dense that the entire system just gridlocks. It's not

150
00:07:13,959 --> 00:07:17,399
that the processor itself isn't fast enough to do the maths.

151
00:07:17,560 --> 00:07:20,480
We physically cannot move the data back and forth fast

152
00:07:20,560 --> 00:07:22,319
enough to keep up with biology.

153
00:07:22,519 --> 00:07:25,079
Speaker 1: It's like trying to drain the entire ocean through a

154
00:07:25,120 --> 00:07:26,680
cocktail straw, exactly.

155
00:07:26,920 --> 00:07:29,839
Speaker 2: And the thing is We have absolutely tried to brute force.

156
00:07:29,879 --> 00:07:34,240
This science is nothing if not incredibly stubborn. If you

157
00:07:34,319 --> 00:07:37,600
look at the Blue Brain project, this launched way back

158
00:07:37,600 --> 00:07:39,319
in two thousand and five. It was one of the

159
00:07:39,319 --> 00:07:43,279
most ambitious attempts ever to reverse engineer them a million.

160
00:07:43,000 --> 00:07:45,279
Speaker 1: Brain two thousand and five. So we've really been banging

161
00:07:45,279 --> 00:07:46,720
our heads against this for a while.

162
00:07:46,759 --> 00:07:48,439
Speaker 2: We have. And they weren't even trying to do a

163
00:07:48,519 --> 00:07:50,759
human brain yet. They were just trying to build a

164
00:07:51,079 --> 00:07:56,120
biologically detailed digital reconstruction of a rat's neocortex.

165
00:07:55,639 --> 00:07:57,319
Speaker 1: Just a piece of a rat brain.

166
00:07:57,240 --> 00:08:00,079
Speaker 2: And not even the whole rat brain. They succeeded in

167
00:08:00,120 --> 00:08:02,519
simulating a small cortical column.

168
00:08:02,120 --> 00:08:03,639
Speaker 1: Which is how many neurons.

169
00:08:03,319 --> 00:08:04,959
Speaker 2: Roughly about thirty thousand neurons.

170
00:08:05,120 --> 00:08:08,160
Speaker 1: Thirty thousand, I mean, that's a literal rounding error compared

171
00:08:08,199 --> 00:08:11,040
to the eighty six billion in a human precisely.

172
00:08:11,600 --> 00:08:15,439
Speaker 2: But even that tiny microscopic column required a massive supercomputer

173
00:08:15,519 --> 00:08:18,120
and an immense amount of memory just to track the

174
00:08:18,160 --> 00:08:21,439
ion channels and the electrical spikes of that one tiny

175
00:08:21,439 --> 00:08:23,920
sliver of tissue. When you try to scale that approach

176
00:08:24,000 --> 00:08:26,439
up from a rat sliver to a full human brain,

177
00:08:26,920 --> 00:08:30,399
you hit what computer scientists call an exponential wall.

178
00:08:30,319 --> 00:08:32,559
Speaker 1: Meaning the hardware requirements just go vertical.

179
00:08:32,840 --> 00:08:35,559
Speaker 2: Right. To simulate a human brain at that exact same

180
00:08:35,679 --> 00:08:39,919
level of biological detail using classical processors, you need computational

181
00:08:39,960 --> 00:08:42,759
resources rivaling the size and the power demands of a

182
00:08:42,799 --> 00:08:44,919
large industrial plant, maybe bigger.

183
00:08:45,000 --> 00:08:47,559
Speaker 1: It's the map becoming larger than the territory itself.

184
00:08:47,679 --> 00:08:50,039
Speaker 2: Yes, And there was a really clear, kind of sobering

185
00:08:50,080 --> 00:08:53,799
example of this back in twenty thirteen with Japan's k supercomputer.

186
00:08:54,320 --> 00:08:56,000
At the time, it was one of the fastest, most

187
00:08:56,000 --> 00:08:58,559
capable machines on the planet, and they used it to

188
00:08:58,600 --> 00:09:01,360
try and simulate just one second of brain.

189
00:09:01,200 --> 00:09:03,559
Speaker 1: Activity, just one single second, okay.

190
00:09:03,320 --> 00:09:06,159
Speaker 2: And the model they built represented only about one percent

191
00:09:06,200 --> 00:09:07,320
of the human brains network.

192
00:09:07,360 --> 00:09:09,559
Speaker 1: Okay, so one percent of the brain for one second.

193
00:09:10,039 --> 00:09:13,240
How long did it actually take this massive supercomputer to

194
00:09:13,279 --> 00:09:13,639
do it.

195
00:09:13,639 --> 00:09:14,919
Speaker 2: It took forty minutes.

196
00:09:14,879 --> 00:09:17,799
Speaker 1: Forty minutes to process one second of thought forty minutes.

197
00:09:17,879 --> 00:09:20,240
Speaker 2: So you have to realize that is not a real

198
00:09:20,279 --> 00:09:24,879
time simulation. That is an incredibly slow motion recording. It's

199
00:09:24,919 --> 00:09:27,879
like watching a movie where every single frame takes an

200
00:09:27,919 --> 00:09:30,720
hour to load on your screen. You can't interact with

201
00:09:30,759 --> 00:09:32,559
a system like that. You can't talk to it. If

202
00:09:32,559 --> 00:09:34,519
you asked it to say hello, you'd have to wait

203
00:09:34,559 --> 00:09:36,799
a lifetime for it to just process the sound waves

204
00:09:36,799 --> 00:09:37,320
of your voice.

205
00:09:37,360 --> 00:09:40,159
Speaker 1: But wait, didn't we see some major progress recently because

206
00:09:40,200 --> 00:09:43,159
I recall reading something I think it was around twenty

207
00:09:43,200 --> 00:09:48,720
twenty five about the Fugaku supercomputer doing something with a

208
00:09:48,759 --> 00:09:50,399
mouse cortex that was much.

209
00:09:50,200 --> 00:09:53,840
Speaker 2: Faster we did. Yeah, in twenty twenty five, researchers used Fugaku,

210
00:09:54,360 --> 00:09:58,120
which is an incredibly powerful next generation machine, to simulate

211
00:09:58,120 --> 00:10:01,840
a mouse cortex with nearly ten million neurons, and they

212
00:10:01,840 --> 00:10:04,159
actually got very close to real time performance on that.

213
00:10:04,240 --> 00:10:06,159
Speaker 1: Okay, so that sounds like we fix the problem, right,

214
00:10:06,200 --> 00:10:08,200
I mean, if we can do a mouse in real time, now,

215
00:10:08,320 --> 00:10:10,759
surely a human is just the next logical step on

216
00:10:10,799 --> 00:10:12,200
the list. Just give it a few years.

217
00:10:12,480 --> 00:10:16,240
Speaker 2: Well not quite. They achieved that milestone because the models

218
00:10:16,279 --> 00:10:19,480
were heavily simplified, and yes, the hardware was faster, but

219
00:10:19,559 --> 00:10:22,919
it actually proved the exact opposite point. It demonstrated that

220
00:10:22,960 --> 00:10:26,320
even with the absolute best silicon we can possibly manufacture

221
00:10:26,320 --> 00:10:29,960
on Earth, we're hitting a fundamental limit. We can't just

222
00:10:30,000 --> 00:10:33,200
solve this by making transistors smaller and smaller, and definitely

223
00:10:33,519 --> 00:10:38,159
we're trying to run a three dimensional, deeply interconnected biological

224
00:10:38,200 --> 00:10:41,159
network on a two dimensional linear architecture.

225
00:10:41,559 --> 00:10:43,960
Speaker 1: So it feels like we're trying to run I don't know,

226
00:10:44,039 --> 00:10:48,120
a hyperrealistic modern video game on a pocket calculator.

227
00:10:48,200 --> 00:10:50,639
Speaker 2: It's honestly worse than that. We are trying to map

228
00:10:50,840 --> 00:10:54,240
a complex fluid territory onto a grid that is just

229
00:10:54,399 --> 00:10:56,519
far too rigid to hold it. And this brings us

230
00:10:56,600 --> 00:10:58,559
right back to the energy problem we started with the

231
00:10:58,559 --> 00:10:59,759
thermo dynamics of it all.

232
00:11:00,240 --> 00:11:02,559
Speaker 1: The twenty watts of the human brainers is the nuclear

233
00:11:02,600 --> 00:11:06,159
power plant needed for the supercomputer. Let's dig into why

234
00:11:06,200 --> 00:11:09,279
our computers run so incredibly hot. Why do we need

235
00:11:09,320 --> 00:11:12,000
those massive rivers of water and cooling towers. Is it

236
00:11:12,080 --> 00:11:12,960
just physical friction?

237
00:11:13,399 --> 00:11:16,360
Speaker 2: It is friction, but not just the mechanical kind of

238
00:11:16,399 --> 00:11:18,399
friction you think of when you rub your hands together.

239
00:11:18,840 --> 00:11:21,559
It actually comes down to the deep physics of information itself.

240
00:11:22,519 --> 00:11:24,759
There is a principle called the Lando principle.

241
00:11:24,919 --> 00:11:28,360
Speaker 1: The Lando principle. Okay, again with the spy thriller names.

242
00:11:28,120 --> 00:11:30,440
Speaker 2: I know it does sound like one, but it's actually

243
00:11:30,440 --> 00:11:35,480
a fundamental law of thermodynamics applied specifically to information theory.

244
00:11:36,120 --> 00:11:39,159
Rolf Landauer back in the nineteen sixties figured out that

245
00:11:39,279 --> 00:11:42,799
information is physical, It's not just abstract math floating in

246
00:11:42,840 --> 00:11:46,000
the ether. And he proved that anytime you erase information,

247
00:11:46,399 --> 00:11:48,720
anytime you reset a bit from a one back to

248
00:11:48,759 --> 00:11:52,559
a zero, or you discard a mathematical possibility, you must

249
00:11:52,559 --> 00:11:54,360
release a tiny amount of physical heat.

250
00:11:54,399 --> 00:11:57,120
Speaker 1: Wait. Wait, so the actual act of forgetting something creates

251
00:11:57,159 --> 00:11:57,799
heat in.

252
00:11:57,759 --> 00:12:01,480
Speaker 2: A classical computer. Yes. Absolutely. To make a decision, a

253
00:12:01,519 --> 00:12:04,080
classical computer has to throw away all the wrong answers,

254
00:12:04,600 --> 00:12:08,080
It has to collapse multiple possibilities down to one single result.

255
00:12:08,799 --> 00:12:12,720
That process of deletion permanently increases the entropy of the universe,

256
00:12:13,039 --> 00:12:15,639
and that entropy shows up in our world as heat.

257
00:12:16,159 --> 00:12:18,559
Speaker 1: So every single time my laptop decides yes or no,

258
00:12:18,840 --> 00:12:21,279
it's paying a physical tax in the form of heat.

259
00:12:21,559 --> 00:12:25,159
Speaker 2: Correct, and modern processors are doing this billions and billions

260
00:12:25,159 --> 00:12:29,679
of times per second. That heat accumulates rapidly. That's exactly

261
00:12:29,759 --> 00:12:32,360
why your laptop burns your legs if you leave it

262
00:12:32,360 --> 00:12:35,840
there too long. We are quite literally fighting physics.

263
00:12:36,000 --> 00:12:39,279
Speaker 1: But the brain. The brain is super weird because it

264
00:12:39,279 --> 00:12:41,960
operates at room temperature, it's wet, it doesn't have a

265
00:12:42,000 --> 00:12:45,440
cooling fans strapped to your forehead, and most importantly, your

266
00:12:45,519 --> 00:12:48,559
head doesn't literally boil over when you try to solve

267
00:12:48,720 --> 00:12:52,360
a complex math problem or vividly visualize a childhood memory.

268
00:12:52,399 --> 00:12:53,159
Speaker 2: Thank goodness for that.

269
00:12:53,279 --> 00:12:56,480
Speaker 1: Yeah, so, how is biology cheating the Landau principle? How

270
00:12:56,559 --> 00:12:57,720
is it getting a free pass?

271
00:12:57,960 --> 00:13:00,240
Speaker 2: It's not cheating at all. It's just playing a completely

272
00:13:00,320 --> 00:13:03,159
different game. It comes down to how these two very

273
00:13:03,200 --> 00:13:04,919
different systems handle noise.

274
00:13:05,120 --> 00:13:08,399
Speaker 1: The noise you mean, like electrical interference, static.

275
00:13:08,159 --> 00:13:12,360
Speaker 2: Thermal noise, microscopic vibrations. At the atomic level, the world

276
00:13:12,440 --> 00:13:15,960
is a very violent place. Atoms are constantly crashing into

277
00:13:16,000 --> 00:13:19,840
each other, vibrating wildly. In a traditional silicon chip, this

278
00:13:19,960 --> 00:13:23,720
thermal noise is the absolute enemy. It jiggles the electrons

279
00:13:23,720 --> 00:13:27,679
around and causes calculation errors. Engineers spend billions of dollars

280
00:13:27,679 --> 00:13:31,000
designing chips that can overpower this background interference to ensure

281
00:13:31,039 --> 00:13:36,039
a perfectly clear loud signal. They're constantly fighting against the environment.

282
00:13:36,159 --> 00:13:38,960
Speaker 1: Okay, so silicon fights the noise. It tries to shout

283
00:13:39,000 --> 00:13:40,759
over it. What does biology do?

284
00:13:41,279 --> 00:13:45,320
Speaker 2: Biology embraces it. The environment inside a wet, warm human

285
00:13:45,360 --> 00:13:48,840
brain is incredibly noisy in a thermal sense. Molecules are

286
00:13:48,879 --> 00:13:52,600
constantly jittering and colliding. But instead of spending precious energy

287
00:13:52,600 --> 00:13:55,720
to suppress all this chaos, the brain seems to tolerate it.

288
00:13:55,759 --> 00:13:57,000
In fact, it utilizes it.

289
00:13:57,000 --> 00:13:58,799
Speaker 1: It uses the noise. How does that even work?

290
00:13:59,159 --> 00:14:02,919
Speaker 2: It rides the thermal fluctuations to actually lower the energy

291
00:14:02,919 --> 00:14:06,039
barrier needed to fire a neuron. It works with the

292
00:14:06,120 --> 00:14:09,720
chaos rather than burning energy to fight against it. We'll

293
00:14:09,759 --> 00:14:13,039
actually see this in plants too, which is fascinating. In photosynthesis,

294
00:14:13,080 --> 00:14:16,679
researchers have found that tiny quantum vibrations help energy move

295
00:14:16,759 --> 00:14:20,120
through a plant cell with almost one hundred percent efficiency.

296
00:14:20,679 --> 00:14:23,639
The plant actively uses the background noise to shake the

297
00:14:23,679 --> 00:14:25,480
solar energy down into the right place.

298
00:14:25,879 --> 00:14:28,480
Speaker 1: That is that's incredibly clever. So it's like a surfer

299
00:14:28,720 --> 00:14:31,000
using the natural power of a wave to move forward

300
00:14:31,360 --> 00:14:33,519
versus a motor boat burning tons of fuel to try

301
00:14:33,559 --> 00:14:34,759
and push through that same wave.

302
00:14:35,080 --> 00:14:38,200
Speaker 2: That is a perfect analogy. The boat burns fuel to

303
00:14:38,200 --> 00:14:41,200
fight the water. The surfer uses the water's own energy

304
00:14:41,240 --> 00:14:44,759
to propel themselves. And this is exactly where the classical

305
00:14:44,799 --> 00:14:49,720
computing model completely falls apart for simulating the brain. If

306
00:14:49,720 --> 00:14:52,759
the brain were a digital computer, it would need megawatts

307
00:14:52,799 --> 00:14:55,919
of power just to fight off that thermal noise. The

308
00:14:55,960 --> 00:14:57,799
fact that it runs on a metabolic trickle of just

309
00:14:57,840 --> 00:15:01,600
twenty wants implies it is using a commpletely different physical mechanism.

310
00:15:02,200 --> 00:15:04,360
It strongly suggests the brain has found a way to

311
00:15:04,399 --> 00:15:08,639
process information that somehow avoids the thermodynamic penalties of classical

312
00:15:08,639 --> 00:15:09,600
computing entirely.

313
00:15:10,000 --> 00:15:12,200
Speaker 1: And this is the breadcrumb trail that leads us directly

314
00:15:12,200 --> 00:15:13,279
into quantum mechanics.

315
00:15:13,440 --> 00:15:18,200
Speaker 2: It is because quantum processes are inherently reversible. In strict theory,

316
00:15:18,399 --> 00:15:22,279
a reversible computation consumes zero energy and generates zero heat

317
00:15:22,440 --> 00:15:25,080
because you aren't deleting information at all, You're just transforming

318
00:15:25,080 --> 00:15:27,279
it from one state to another. Now, the brain isn't

319
00:15:27,320 --> 00:15:30,600
perfectly efficient, obviously we do need food, but its ability

320
00:15:30,639 --> 00:15:33,919
to skirt that thermodynamic wall so closely suggests it is

321
00:15:33,960 --> 00:15:36,879
tapping into these deeper, far more efficient physical laws.

322
00:15:37,200 --> 00:15:39,480
Speaker 1: Now I have to stop you in play Devil's advocate

323
00:15:39,519 --> 00:15:41,480
here for a second, because I feel like we just

324
00:15:41,519 --> 00:15:43,639
took a hard left turn into magic.

325
00:15:43,840 --> 00:15:45,200
Speaker 2: It really sounds like it doesn't it?

326
00:15:45,200 --> 00:15:48,840
Speaker 1: It totally does? I mean quantum consciousness. I've seen this

327
00:15:48,879 --> 00:15:50,759
in bad sci fi movies, and I've seen it on

328
00:15:51,039 --> 00:15:54,720
let's just say, some less than reputable online wellness blogs.

329
00:15:56,039 --> 00:15:59,639
Isn't the standard accepted physics argument that the brain is

330
00:15:59,679 --> 00:16:02,039
just too hot for any of this quantum stuff to work.

331
00:16:02,039 --> 00:16:04,039
Speaker 2: You're talking about decoherence.

332
00:16:03,480 --> 00:16:06,600
Speaker 1: Right, decoherence? My understanding, and please correct me if I'm

333
00:16:06,600 --> 00:16:10,759
wrong here is that quantum states are incredibly famously fragile.

334
00:16:11,120 --> 00:16:13,279
To keep a quibit stable in a lab, you need

335
00:16:13,279 --> 00:16:18,000
temperatures close to absolute zero, you need a perfect vacuum.

336
00:16:18,279 --> 00:16:21,200
But the human brain is ninety eight point six degrees.

337
00:16:21,279 --> 00:16:24,120
It's wet, it's salty, it's a vibrating mess. How could

338
00:16:24,120 --> 00:16:27,480
a delicate quantum state survive in a warm bowl of

339
00:16:27,519 --> 00:16:30,799
biological soup for more than a fraction of a femtosecond.

340
00:16:31,080 --> 00:16:33,480
Speaker 2: That is a very valid criticism. In fact, for a

341
00:16:33,480 --> 00:16:37,039
long time, that was the criticism for maybe thirty years.

342
00:16:37,399 --> 00:16:41,000
Mainstream physicists basically laughed this entire theory out of the

343
00:16:41,039 --> 00:16:45,240
room because of exactly what you just said. Max Tegmark,

344
00:16:45,279 --> 00:16:49,000
who's a very famous cosmologist, he published calculations that basically

345
00:16:49,000 --> 00:16:52,879
said any quantum state in the brain would collapse instantly,

346
00:16:52,960 --> 00:16:55,039
way too fast to be of any use for actual

347
00:16:55,080 --> 00:16:55,799
thought or memory.

348
00:16:55,879 --> 00:16:58,080
Speaker 1: So what changed? Why are we even discussing it today

349
00:16:58,120 --> 00:17:00,000
if the math clearly said it was impops?

350
00:17:00,559 --> 00:17:04,359
Speaker 2: Because biology surprised us again. It turns out nature might

351
00:17:04,359 --> 00:17:08,119
have actually engineered its own microscopic version of a vacuum chamber.

352
00:17:08,880 --> 00:17:11,400
We started looking much closer at the structure of the

353
00:17:11,400 --> 00:17:14,440
neuron itself. We used to think neurons were basically just

354
00:17:14,559 --> 00:17:17,799
empty bags of fluid that transmitted electrical signals along their

355
00:17:17,799 --> 00:17:20,960
outer membranes like a wet wire, exactly like a simple wire.

356
00:17:21,559 --> 00:17:23,839
But as imaging got better, it turns out the interior

357
00:17:23,880 --> 00:17:27,599
of the neuron is packed with this intricate structural scaffolding

358
00:17:27,680 --> 00:17:29,200
called microtubules.

359
00:17:29,359 --> 00:17:32,359
Speaker 1: Microtubules okay, and these aren't just there for structural support,

360
00:17:32,440 --> 00:17:34,599
like the rebar inside a concrete building.

361
00:17:34,880 --> 00:17:38,319
Speaker 2: That's exactly what we thought they were for, just biological rebar.

362
00:17:38,960 --> 00:17:42,119
But Roger Penrose, who is a Nobel laureate physicist, and

363
00:17:42,200 --> 00:17:46,799
Stuart Hamroff, an anesthesiologist, they proposed a theory called orchestrated

364
00:17:46,799 --> 00:17:51,640
objective reduction, and they argue that these microscopic tubes actually

365
00:17:51,680 --> 00:17:53,920
act as quantum wave guides.

366
00:17:53,759 --> 00:17:57,000
Speaker 1: A waveguide, so like a protective tunnel that shields the

367
00:17:57,039 --> 00:17:58,079
signal exactly.

368
00:17:58,559 --> 00:18:02,079
Speaker 2: The internal structure of the microte tubule creates a highly insulated,

369
00:18:02,279 --> 00:18:06,279
completely isolated environment. It acts like a physical field that

370
00:18:06,319 --> 00:18:09,680
protects the fragile quantum states from the wet, noisy environment

371
00:18:09,759 --> 00:18:12,880
of the rest of the brain. It theoretically allows quantum

372
00:18:12,880 --> 00:18:15,119
processing to happen safely at room temperature.

373
00:18:15,240 --> 00:18:17,799
Speaker 1: Has this actually been proven or is it just a

374
00:18:17,880 --> 00:18:21,200
really elegant theory that happens to fix the math problem.

375
00:18:21,319 --> 00:18:23,720
Speaker 2: Recent experiments are actually starting to validate it, which is

376
00:18:23,759 --> 00:18:27,559
where this gets incredibly exciting. Researchers have recently observed actual

377
00:18:27,640 --> 00:18:31,519
quantum vibrations happening in the protein structures inside these microtubules.

378
00:18:31,960 --> 00:18:34,920
They found that energy can move through these specific structures

379
00:18:35,240 --> 00:18:38,119
without losing power, behaving much more like a quantum wave

380
00:18:38,160 --> 00:18:39,240
than a classical particle.

381
00:18:39,480 --> 00:18:42,839
Speaker 1: Wait, that sounds exactly like the photosynthesis thing you mentioned earlier.

382
00:18:43,000 --> 00:18:46,279
Speaker 2: It's very similar, and it's intimately linked to these specific

383
00:18:46,319 --> 00:18:50,119
molecules called tryptofan networks that live inside the tube. When

384
00:18:50,160 --> 00:18:54,119
these molecules resonate together, they can create a large scale

385
00:18:54,240 --> 00:18:56,160
quantum state. You can think of it kind of like

386
00:18:56,200 --> 00:18:56,680
a laser.

387
00:18:56,759 --> 00:18:57,720
Speaker 1: A laser how so.

388
00:18:58,039 --> 00:19:01,000
Speaker 2: Well, in a normal flashlight, photons are caaotic, they just

389
00:19:01,079 --> 00:19:03,839
go everywhere, bouncing off each other, But in a laser,

390
00:19:04,160 --> 00:19:08,960
the photon synchronized perfectly into a single coherent beam. In

391
00:19:09,000 --> 00:19:13,079
the human brain, this kind of biological synchronization would allow

392
00:19:13,200 --> 00:19:17,440
vast amounts of information to be processed across large distances instantly.

393
00:19:18,039 --> 00:19:20,759
It means the neuron isn't just calculating by using simple

394
00:19:20,799 --> 00:19:24,519
electrical spikes on the outside, It is fundamentally calculating using

395
00:19:24,599 --> 00:19:28,200
the resonant quantum frequencies of the universe itself on the inside.

396
00:19:28,240 --> 00:19:30,519
Speaker 1: And that completely solves the efficiency.

397
00:19:30,039 --> 00:19:35,160
Speaker 2: Paradox completely, because if the brain uses quantum superposition, it

398
00:19:35,200 --> 00:19:40,079
can process vast landscapes of possibility simultaneously without having to

399
00:19:40,160 --> 00:19:42,480
check every single one of them one by one. It

400
00:19:42,480 --> 00:19:45,359
doesn't generate all that waste heat of a billion binary

401
00:19:45,400 --> 00:19:48,480
switches flipping back and forth. It essentially feels out the

402
00:19:48,519 --> 00:19:50,519
correct answer through wave interference.

403
00:19:51,000 --> 00:19:55,720
Speaker 1: So if the human brain is fundamentally quantum sweep, then

404
00:19:55,799 --> 00:19:59,319
a classical supercomputer, no matter how big, we build it

405
00:19:59,359 --> 00:20:01,519
is never e are going to cut it never.

406
00:20:02,039 --> 00:20:05,799
Speaker 2: You simply cannot simulate a true quantum event with a

407
00:20:05,839 --> 00:20:09,119
binary switch. You need hardware that speaks the exact same

408
00:20:09,200 --> 00:20:11,799
language as the biology you're trying to copy, which.

409
00:20:11,599 --> 00:20:15,000
Speaker 1: Brings us perfectly to the hardware building the actual bridge

410
00:20:15,000 --> 00:20:17,960
between the biology and the machine. There's a quote from

411
00:20:18,039 --> 00:20:20,559
Richard Feynman from back in the nineteen eighties that feels

412
00:20:20,599 --> 00:20:22,039
incredibly prophetic right now.

413
00:20:22,079 --> 00:20:24,720
Speaker 2: Oh, it's the foundational quote for this entire field of study.

414
00:20:25,000 --> 00:20:27,960
He said, nature is not classical, damn it. And if

415
00:20:28,000 --> 00:20:30,000
you want to make a simulation of nature, you'd better

416
00:20:30,000 --> 00:20:31,079
make it quantum mechanical.

417
00:20:31,200 --> 00:20:32,559
Speaker 1: He absolutely knew, He.

418
00:20:32,559 --> 00:20:36,039
Speaker 2: Knew decades before anyone else, and now we are finally

419
00:20:36,079 --> 00:20:38,759
actually following that advice. But to understand how we are

420
00:20:38,759 --> 00:20:40,880
physically building this, we have to talk about a core

421
00:20:40,960 --> 00:20:42,559
concept called the Hamiltonian.

422
00:20:42,680 --> 00:20:45,079
Speaker 1: The Hamiltonian, Okay, let's unpack this. It sounds like a

423
00:20:45,160 --> 00:20:47,480
hit Broadway musical, but I'm guessing it's heavy math.

424
00:20:47,720 --> 00:20:50,200
Speaker 2: It is math, but it's actually a really beautiful concept.

425
00:20:50,759 --> 00:20:54,480
The Hamiltonian is essentially the unique energy fingerprint of any

426
00:20:54,480 --> 00:20:59,200
physical system, every single object in the universe has one.

427
00:20:59,359 --> 00:21:03,000
It mathematic describes how that object vibrates, how it physically

428
00:21:03,039 --> 00:21:05,799
interacts with the world around it, how its total energy

429
00:21:05,839 --> 00:21:09,240
is distributed at any given moment. A single biological neuron

430
00:21:09,279 --> 00:21:12,240
has a Hamiltonian, a network of neurons has a much

431
00:21:12,279 --> 00:21:13,160
more complex one.

432
00:21:13,240 --> 00:21:16,480
Speaker 1: So if you can accurately define the Hamiltonian of a system,

433
00:21:17,039 --> 00:21:20,519
you've essentially captured the soul of that system, so to speak.

434
00:21:20,680 --> 00:21:24,200
Speaker 2: Physically speaking, yes, you have everything you need. Now Here

435
00:21:24,279 --> 00:21:27,720
is the huge problem we face when a classical computer

436
00:21:28,000 --> 00:21:31,119
tries to simulate a Hamiltonian. It has to do the

437
00:21:31,119 --> 00:21:34,519
math the hard way. It has to solve massive, incredibly

438
00:21:34,640 --> 00:21:39,119
complex differential equations just to predict what the system will

439
00:21:39,119 --> 00:21:42,480
do one millisecond from now, and as the biological system

440
00:21:42,519 --> 00:21:46,720
gets bigger, the equation gets exponentially harder. Eventually the math

441
00:21:46,759 --> 00:21:50,519
becomes so impossibly dense that the classical computer just freezes up.

442
00:21:50,599 --> 00:21:52,960
Speaker 1: It's like trying to calculate the exact curve of a

443
00:21:53,039 --> 00:21:56,119
crashing ocean wave by measuring the trajectory of every single

444
00:21:56,200 --> 00:21:58,079
drop of water individually exactly.

445
00:21:58,160 --> 00:22:01,799
Speaker 2: It's impossible, But a quantum computer approaches that exact same

446
00:22:01,839 --> 00:22:05,440
problem completely differently. It doesn't try to solve the equation

447
00:22:05,559 --> 00:22:08,319
at all. It attempts to become the equation.

448
00:22:08,119 --> 00:22:10,160
Speaker 1: Become the equation that is wild.

449
00:22:10,319 --> 00:22:13,359
Speaker 2: This is the absolute core of quantum simulation. The idea

450
00:22:13,359 --> 00:22:15,960
is to set up the physical quibots inside the computer

451
00:22:16,359 --> 00:22:19,720
so that they perfectly mimic the physical laws governing the neurons.

452
00:22:20,119 --> 00:22:22,799
You basically force the quibbets to interact with each other

453
00:22:23,119 --> 00:22:26,200
in the exact same physical way the biological molecules interact

454
00:22:26,240 --> 00:22:26,960
inside your head.

455
00:22:27,079 --> 00:22:29,759
Speaker 1: So you're not calculating what the neuron is going to do.

456
00:22:29,920 --> 00:22:33,160
You're building a literal digital version of the neuron out

457
00:22:33,160 --> 00:22:35,799
of quibits and just watching what it does naturally.

458
00:22:36,000 --> 00:22:39,680
Speaker 2: Yes, instead of calculating that ocean wave drop by drop,

459
00:22:40,079 --> 00:22:43,079
you just build a physical wave tank, fill it with water,

460
00:22:43,400 --> 00:22:46,640
and watch how the water moves. The hardware itself is

461
00:22:46,640 --> 00:22:47,319
the simulation.

462
00:22:47,880 --> 00:22:50,799
Speaker 1: That is a massive shift in how we think about computing.

463
00:22:51,279 --> 00:22:54,759
But we don't have eighty six billion stable quibits yet.

464
00:22:54,880 --> 00:22:56,720
I mean, last I check the news, the big tech

465
00:22:56,759 --> 00:22:59,160
companies were still struggling to get a few hundred stable

466
00:22:59,200 --> 00:22:59,960
ones to work together.

467
00:23:00,400 --> 00:23:03,519
Speaker 2: That's very true. We're currently nowhere near a one to

468
00:23:03,559 --> 00:23:05,799
one map of the human brain. So we have to

469
00:23:05,839 --> 00:23:08,519
be incredibly efficient with the quiots we do have, and

470
00:23:08,599 --> 00:23:11,240
the main tool researchers used for that is something called

471
00:23:11,240 --> 00:23:12,119
a tensor network.

472
00:23:12,160 --> 00:23:14,440
Speaker 1: A tensor network that sounds incredibly sci.

473
00:23:14,240 --> 00:23:18,599
Speaker 2: Fi, it's essentially a very advanced geometry tool. Basically, it

474
00:23:18,640 --> 00:23:23,160
allows researchers to drastically compress the biological data. They look

475
00:23:23,200 --> 00:23:25,960
at the massive web of the brain and they identify

476
00:23:26,000 --> 00:23:29,519
which connections are absolutely critical the core ghost in the machine,

477
00:23:29,559 --> 00:23:32,960
so to speak, and which ones are just redundant background noise.

478
00:23:33,720 --> 00:23:37,200
And by using this specific geometry, scientists can map the

479
00:23:37,240 --> 00:23:40,960
insanely complex web of the human brain onto the much

480
00:23:40,960 --> 00:23:44,440
smaller available grid of a current quantum processor.

481
00:23:44,559 --> 00:23:46,799
Speaker 1: So it's kind of like a low resolution shatter of

482
00:23:46,839 --> 00:23:47,519
the brain.

483
00:23:47,440 --> 00:23:50,839
Speaker 2: A low resolution shadow that still fundamentally behaves exactly like

484
00:23:50,880 --> 00:23:54,640
the real thing. It perfectly preserves the important quantum correlations

485
00:23:54,920 --> 00:23:57,119
without needing all the raw processing power.

486
00:23:57,200 --> 00:23:59,880
Speaker 1: Okay, so the theory is incredibly sound. We know we

487
00:24:00,079 --> 00:24:02,400
absolutely need quantum mechanics. We know how to map the

488
00:24:02,400 --> 00:24:06,240
brain using Hamiltonians and these tensor networks, but historically quantum

489
00:24:06,240 --> 00:24:09,880
computers have been well flaky. Is probably the nicest word.

490
00:24:09,640 --> 00:24:12,720
Speaker 2: For it, flaky is putting it very mildly. They have

491
00:24:12,799 --> 00:24:17,079
been notoriously incredibly unstable. The actual physical quibbitts that hold

492
00:24:17,079 --> 00:24:20,720
the information are wildly sensitive to everything in their environment.

493
00:24:21,200 --> 00:24:24,119
A stray magnetic field from a cell phone, a tiny

494
00:24:24,119 --> 00:24:27,839
temperature change in the room, even a cosmic ray from space.

495
00:24:27,880 --> 00:24:30,680
Boom the quid. It loses its state. That's what we

496
00:24:30,720 --> 00:24:31,359
call an error.

497
00:24:31,559 --> 00:24:34,960
Speaker 1: And obviously you can't simulate a functioning human brain if

498
00:24:35,000 --> 00:24:38,640
your computer effectively gets severe amnesia every single microsecond.

499
00:24:38,759 --> 00:24:41,480
Speaker 2: Exactly, you need a system that can reliably maintain a

500
00:24:41,519 --> 00:24:44,799
continuous thought. And for years we were stuck with this

501
00:24:44,920 --> 00:24:47,680
terrible paradox in the field. Adding more kibbutz to a

502
00:24:47,720 --> 00:24:50,160
machine actually made the whole system more fragile. It was

503
00:24:50,200 --> 00:24:52,519
exactly like building a house of cards. The taller you

504
00:24:52,519 --> 00:24:54,759
built the tower, the more likely the whole thing was

505
00:24:54,799 --> 00:24:55,759
to completely collapse.

506
00:24:55,880 --> 00:25:00,000
Speaker 1: But something massive changed recently, specifically late twenty twenty four.

507
00:25:00,400 --> 00:25:04,000
Speaker 2: Yes, the Willow Chip from Google. This was the massive

508
00:25:04,000 --> 00:25:06,599
turning point the industry had been waiting for. It wasn't

509
00:25:06,640 --> 00:25:09,640
just a basic update and processing speed. It was a fundamental,

510
00:25:09,880 --> 00:25:12,640
undeniable proof of concept for quantum erric correction.

511
00:25:12,960 --> 00:25:15,240
Speaker 1: Let's really talk about the hardware that makes this possible.

512
00:25:15,319 --> 00:25:18,920
Because you mentioned the Willow chip. Why is this one

513
00:25:19,000 --> 00:25:22,839
specific piece of silicon or well superconductor the turning point.

514
00:25:23,559 --> 00:25:26,680
Speaker 2: It all comes down to how we handle those inevitable errors.

515
00:25:27,279 --> 00:25:29,559
Like we said in the past, if you added more

516
00:25:29,599 --> 00:25:34,200
physical qubits to a computer, you actually made it mathematically worse, worse,

517
00:25:34,680 --> 00:25:38,000
much worse. Imagine you're trying to balance that house of cards.

518
00:25:38,319 --> 00:25:41,279
If you have two cards, it's shaky, but manageable. If

519
00:25:41,279 --> 00:25:44,240
you had two more, it's noticeably shakier. If you try

520
00:25:44,240 --> 00:25:46,880
to build a massive tower of one hundred cards, it

521
00:25:46,960 --> 00:25:50,000
is statistically almost guaranteed to fall over before you.

522
00:25:49,960 --> 00:25:53,079
Speaker 1: Finish, because there's simply more points of failure, more things

523
00:25:53,079 --> 00:25:53,839
that can go wrong.

524
00:25:53,960 --> 00:25:58,039
Speaker 2: Exactly, every single new quipot introduces a new source of noise.

525
00:25:58,720 --> 00:26:01,559
But the Willow chip did something completely different. It proved

526
00:26:01,599 --> 00:26:03,359
the viability of logical equibts.

527
00:26:03,480 --> 00:26:05,960
Speaker 1: Okay, define that for us. What exactly is a logical

528
00:26:06,000 --> 00:26:07,279
quibbit versus a regular one?

529
00:26:07,319 --> 00:26:10,519
Speaker 2: Okay, So a physical quibot is the actual tangible hardware

530
00:26:10,759 --> 00:26:14,480
of the microscopic superconducting loop etched onto the chip, A

531
00:26:14,519 --> 00:26:18,000
logical quibot, on the other hand, is entirely a software concept.

532
00:26:18,039 --> 00:26:20,880
It's a team, a team of right. Imagine you have

533
00:26:20,920 --> 00:26:25,039
a group of one hundred physical quibbots through software. You

534
00:26:25,039 --> 00:26:27,759
tell them, okay, you one hundred individual guys, you are

535
00:26:27,759 --> 00:26:31,759
now going to act together as one single, highly reliable

536
00:26:31,920 --> 00:26:32,640
bit of information.

537
00:26:32,839 --> 00:26:34,960
Speaker 1: It's massive redundancy.

538
00:26:34,480 --> 00:26:39,119
Speaker 2: Massive coordinated redundancy, because they are constantly actively checking each

539
00:26:39,160 --> 00:26:42,440
other's work. If one physical quibut in the back row

540
00:26:42,519 --> 00:26:45,920
gets hit by a random cosmic ray and accidentally flips

541
00:26:45,920 --> 00:26:48,359
from a zero to a one, the other ninety nine

542
00:26:48,400 --> 00:26:51,000
look at him and say, hey, you're obviously wrong, and

543
00:26:51,039 --> 00:26:53,000
they mathematically force him back to.

544
00:26:52,960 --> 00:26:56,680
Speaker 1: A zero, so the overall group survives the error even

545
00:26:56,680 --> 00:26:59,119
if the individual part temporarily fails.

546
00:26:59,400 --> 00:27:03,079
Speaker 2: Precisely, and the launch of the Willichip was the first time,

547
00:27:03,319 --> 00:27:06,599
the very first time in computing history where physically adding

548
00:27:06,640 --> 00:27:09,680
more individual equipments to the system actually made the overall

549
00:27:09,680 --> 00:27:12,119
air rate go down instead of up. Yeah, that is

550
00:27:12,200 --> 00:27:13,799
our crossing the rubicon moment.

551
00:27:13,920 --> 00:27:16,079
Speaker 1: That is the exact opposite of the house of cards.

552
00:27:16,160 --> 00:27:18,480
The bigger you build it, the stronger the foundation gets.

553
00:27:18,480 --> 00:27:20,319
It's more like building a pyramid.

554
00:27:20,039 --> 00:27:24,359
Speaker 2: Precisely a pyramid is exactly right, and that stability is

555
00:27:24,400 --> 00:27:28,440
the ultimate green light for true biological simulation because a

556
00:27:28,519 --> 00:27:33,680
human brain is fundamentally an incredibly stable system. It maintains

557
00:27:33,720 --> 00:27:37,640
the continuity of your consciousness over decades. To model that digitally,

558
00:27:37,920 --> 00:27:41,440
we absolutely need a computer that doesn't crash and reset constantly.

559
00:27:42,079 --> 00:27:45,079
And thanks to error correction, we are now moving into

560
00:27:45,119 --> 00:27:48,240
an era where we can run complex quantum algorithms for

561
00:27:48,480 --> 00:27:50,119
minutes or even hours at a time.

562
00:27:50,279 --> 00:27:52,440
Speaker 1: We can finally press play on the video of the

563
00:27:52,519 --> 00:27:56,359
human brain rather than just looking at a frozen, fragmented snapshot.

564
00:27:56,480 --> 00:27:58,880
Speaker 2: Yes, and we can actually watch how a digital neural

565
00:27:58,920 --> 00:28:01,880
network evolves in a DAPs and learns over time. That

566
00:28:01,960 --> 00:28:05,480
newfound stability transforms the quantum computer from being just an

567
00:28:05,480 --> 00:28:09,400
experimental physics toy in a lab into a highly reliable

568
00:28:09,640 --> 00:28:11,400
industrial computational engine.

569
00:28:11,440 --> 00:28:13,160
Speaker 1: Okay, so we have the biological theory, we have a

570
00:28:13,160 --> 00:28:15,799
hardware finally stabilizing with Willow. Now let's talk about what

571
00:28:15,799 --> 00:28:18,359
we actually do with this massive power, because the real

572
00:28:18,400 --> 00:28:21,160
world applications are well, some of them sound straight up

573
00:28:21,200 --> 00:28:21,839
like telepathy.

574
00:28:21,960 --> 00:28:26,440
Speaker 2: They really do. Let's start with brain computer interfaces or BCIs.

575
00:28:26,079 --> 00:28:28,599
Speaker 1: Because we have these right now. Right We've all seen

576
00:28:28,640 --> 00:28:31,039
the news clips of people using their minds to move

577
00:28:31,119 --> 00:28:35,200
robotic arms. We've seen paralyzed patients playing video games or

578
00:28:35,240 --> 00:28:37,200
feeding themselves with mechanical limbs.

579
00:28:37,440 --> 00:28:40,119
Speaker 2: We do have them, but there is a major fundamental

580
00:28:40,160 --> 00:28:44,319
problem with the current classical technology latency lag.

581
00:28:44,400 --> 00:28:46,240
Speaker 1: The delay between thought and action.

582
00:28:46,640 --> 00:28:50,000
Speaker 2: Right when a person physically decides to move their arm,

583
00:28:50,519 --> 00:28:54,759
their brain sends a complex cascade of electrical signals, but

584
00:28:54,839 --> 00:28:58,960
those raw signals are incredibly noisy. A classical computer listening

585
00:28:59,000 --> 00:29:02,000
to that brain wave to record the electrical spikes, filter

586
00:29:02,079 --> 00:29:05,559
out all the massive background static, run a complex statistical

587
00:29:05,559 --> 00:29:08,519
algorithm to guess what the user's intention probably is, and

588
00:29:08,559 --> 00:29:09,759
then send the command to the arm.

589
00:29:09,880 --> 00:29:13,279
Speaker 1: So it feels very much like operating heavy machinery. It's disjointed.

590
00:29:13,319 --> 00:29:15,920
There's a noticeable delay between I want to move and

591
00:29:15,960 --> 00:29:17,759
it actually moves exactly.

592
00:29:18,079 --> 00:29:21,279
Speaker 2: The classical system essentially waits until it's statistically certain before

593
00:29:21,319 --> 00:29:24,920
it dares to act, But a quantum processor it handles

594
00:29:25,000 --> 00:29:28,880
fundamental uncertainty naturally. It doesn't need to mathematically filter out

595
00:29:28,880 --> 00:29:32,799
the noise. It can directly ingest the raw incredibly noisy

596
00:29:32,839 --> 00:29:36,319
probability field of the neural spikes and map it directly

597
00:29:36,319 --> 00:29:38,279
onto its own quantum states.

598
00:29:38,000 --> 00:29:41,319
Speaker 1: So it decodes the biological intention instantly.

599
00:29:40,839 --> 00:29:44,960
Speaker 2: Almost instantaneously in some tests, it predicts the intended movement

600
00:29:45,039 --> 00:29:48,119
before the biological signal would have even physically reached the

601
00:29:48,240 --> 00:29:52,240
human hand. But the real world altering game changer isn't

602
00:29:52,279 --> 00:29:54,920
just moving a robotic cursor or an arm. It's moving

603
00:29:54,960 --> 00:29:58,319
from basic motor control to decoding pure semantic thought.

604
00:29:58,559 --> 00:30:02,039
Speaker 1: Semantic thought, you mean reading actual ideas concepts.

605
00:30:02,279 --> 00:30:06,160
Speaker 2: Yes, motor signals are relatively simple. They map directly to

606
00:30:06,200 --> 00:30:11,119
specific muscles. But abstract thoughts like vividly visualizing a loved

607
00:30:11,119 --> 00:30:14,519
one's face, or remembering the abstract concept of justice or

608
00:30:14,559 --> 00:30:18,079
the feeling of warmth, those are scattered wildly across the

609
00:30:18,240 --> 00:30:23,359
entire brain in highly complex transient patterns. Classical algorithms completely

610
00:30:23,359 --> 00:30:25,079
struggle to read them, because it's like trying to read

611
00:30:25,119 --> 00:30:28,240
a novel where the pages are being frantically shuffled constantly.

612
00:30:28,400 --> 00:30:30,680
Speaker 1: But a quantum system doesn't have that limitation.

613
00:30:31,279 --> 00:30:34,200
Speaker 2: A quantum system can look at the intricate correlations between

614
00:30:34,279 --> 00:30:37,720
vastly different regions of the brain all simultaneously. It can

615
00:30:37,759 --> 00:30:42,279
spot the subtle instantaneous synchronization that represents one specific memory

616
00:30:42,599 --> 00:30:43,839
or a complex image.

617
00:30:44,079 --> 00:30:46,839
Speaker 1: So we are talking about high bandwith communication without speech,

618
00:30:47,240 --> 00:30:48,559
without typing a single word.

619
00:30:48,920 --> 00:30:52,200
Speaker 2: If the latency drops low enough, the interface entirely becomes

620
00:30:52,279 --> 00:30:55,559
invisible to the user. The machine stops feeling like a

621
00:30:55,599 --> 00:30:58,400
separate tool you are holding and starts functioning as the

622
00:30:58,440 --> 00:31:01,599
seamless cognitive extension your own mind. You don't use the

623
00:31:01,599 --> 00:31:04,400
computer anymore. You just think and the action happens.

624
00:31:04,519 --> 00:31:08,599
Speaker 1: Yeah, that is both incredibly exhilarating and honestly a little terrifying.

625
00:31:09,200 --> 00:31:12,039
But let's pivot to something that might actually save millions

626
00:31:12,079 --> 00:31:14,079
of lives right now, medicine.

627
00:31:14,319 --> 00:31:16,799
Speaker 2: This is where I truly think the most immediate earth

628
00:31:16,839 --> 00:31:20,039
shattering impact will be the concept of the digital twin.

629
00:31:20,440 --> 00:31:23,759
Speaker 1: I absolutely love this concept, but to really understand it,

630
00:31:23,799 --> 00:31:26,680
we have to talk about what scientists call the origami

631
00:31:26,799 --> 00:31:30,279
of life, protein folding. And I know this is one

632
00:31:30,319 --> 00:31:33,440
of those biological things that sounds incredibly boring until you

633
00:31:33,519 --> 00:31:36,839
realize it literally controls everything happening in our bodies.

634
00:31:37,000 --> 00:31:40,480
Speaker 2: Right. Proteins are the actual physical machinery of life, but

635
00:31:40,519 --> 00:31:43,319
they all start out as these long, floppy strings of

636
00:31:43,319 --> 00:31:47,640
amino acids. To function correctly, they have to physically fold

637
00:31:47,680 --> 00:31:52,200
themselves up into very specific, incredibly complex three dimensional shapes.

638
00:31:52,359 --> 00:31:54,240
If they fold correctly, they do their job.

639
00:31:54,640 --> 00:31:57,359
Speaker 1: But if they misfold, that's where the really bad diseases

640
00:31:57,359 --> 00:31:57,839
come from.

641
00:31:57,960 --> 00:32:01,559
Speaker 2: Exactly Alzheimer's, Parkinson's, Huntington. These are all often caused by

642
00:32:01,559 --> 00:32:06,400
misfolded proteins. They become physically toxic and start aggressively destroying

643
00:32:06,440 --> 00:32:11,200
healthy neurons. Now, predicting exactly how a specific protein string

644
00:32:11,240 --> 00:32:14,519
will fold up is one of the absolute hardest mathematical

645
00:32:14,559 --> 00:32:15,920
problems in all of science.

646
00:32:16,000 --> 00:32:18,400
Speaker 1: Why is it so hard? I mean, it's just geometry

647
00:32:18,480 --> 00:32:19,000
folding a.

648
00:32:18,960 --> 00:32:23,240
Speaker 2: String because a single moderately sized protein chain has more

649
00:32:23,279 --> 00:32:27,400
mathematically possible folded shapes than there are actual atoms in

650
00:32:27,440 --> 00:32:28,519
the observable universe.

651
00:32:28,559 --> 00:32:30,279
Speaker 1: Wait, I'm sorry repeat that statistic.

652
00:32:30,519 --> 00:32:33,400
Speaker 2: A single protein chain has more possible ways to fold

653
00:32:33,440 --> 00:32:36,640
than there are atoms in the entire universe. A classical

654
00:32:36,680 --> 00:32:40,559
supercomputer tries to solve this massive problem by basically checking

655
00:32:40,559 --> 00:32:43,839
the options one by one. It guesses a shape, calculates

656
00:32:43,839 --> 00:32:47,640
the energy required, realizes it's wrong, and tries another. Even

657
00:32:47,680 --> 00:32:51,720
with AI shortcuts, simulating complex dynamics takes months or.

658
00:32:51,799 --> 00:32:53,799
Speaker 1: Years, and a quantum computer approaches it.

659
00:32:53,839 --> 00:32:57,359
Speaker 2: How well, in physics, nature always naturally seeks the path

660
00:32:57,359 --> 00:33:01,400
of least resistance. A protein string will naturally automatically snap

661
00:33:01,440 --> 00:33:04,440
into the specific shape that requires the absolute least amount

662
00:33:04,440 --> 00:33:07,920
of energy to maintain. A quantum computer doesn't bother to

663
00:33:07,960 --> 00:33:11,640
guess and check. It simply simulates the complete energy landscape.

664
00:33:12,000 --> 00:33:15,160
The quibits naturally settle into their own lowest energy state,

665
00:33:15,200 --> 00:33:18,359
automatically mirroring the exact physics of the protein.

666
00:33:18,519 --> 00:33:21,440
Speaker 1: It perfectly mimics the actual physics of the protein physically

667
00:33:21,440 --> 00:33:22,839
falling into place exactly.

668
00:33:22,880 --> 00:33:25,799
Speaker 2: So I imagine this future scenario. We take a scan

669
00:33:26,240 --> 00:33:30,000
of the specific molecular structure of your personal brain. We

670
00:33:30,160 --> 00:33:33,920
load that precise biological data into a stable quantum processor.

671
00:33:34,279 --> 00:33:37,559
We essentially create a perfectly accurate sandbox version of your

672
00:33:37,680 --> 00:33:39,200
unique neural chemistry, a.

673
00:33:39,200 --> 00:33:40,920
Speaker 1: Digital twin of my actual brain.

674
00:33:41,039 --> 00:33:45,039
Speaker 2: Yes. And then we introduce a digital quantum version of

675
00:33:45,079 --> 00:33:48,759
a brand new experimental drug into that simulation. We can

676
00:33:48,799 --> 00:33:52,079
physically watch how that synthetic medication interacts with your specific

677
00:33:52,200 --> 00:33:56,039
misfolded proteins in absolute real time. We can run a

678
00:33:56,119 --> 00:33:58,799
thousand completely different clinical trials on your brain in a

679
00:33:58,839 --> 00:33:59,880
single afternoon.

680
00:33:59,599 --> 00:34:01,640
Speaker 1: Without ever actually touching me, without.

681
00:34:01,400 --> 00:34:05,039
Speaker 2: Ever putting your physical body at the slightest risk, we

682
00:34:05,079 --> 00:34:09,679
can safely test incredibly dangerous compounds. We can effectively debug

683
00:34:09,719 --> 00:34:12,199
the chemical code of the brain before we ever compile

684
00:34:12,239 --> 00:34:15,400
the update. In the real world, currently developing a new

685
00:34:15,440 --> 00:34:18,519
neurological drug takes well over a decade and costs billions

686
00:34:18,519 --> 00:34:22,280
of dollars. This technology turns medicine from a slow game

687
00:34:22,320 --> 00:34:26,800
of trial and error into ultra precise molecular engineering.

688
00:34:26,400 --> 00:34:29,239
Speaker 1: That is just incredible, But it also leads us directly

689
00:34:29,280 --> 00:34:31,719
into the deeper, or much messier side of all this.

690
00:34:31,760 --> 00:34:34,840
We've talked about physical structure, we've talked about targeted medicine,

691
00:34:35,079 --> 00:34:37,800
but the human mind isn't just a biological machine that

692
00:34:37,840 --> 00:34:41,960
folds proteins. It's wildly chaotic. It's highly creative.

693
00:34:42,360 --> 00:34:45,519
Speaker 2: And that is exactly the other area where classical computers

694
00:34:45,559 --> 00:34:46,480
fundamentally fail.

695
00:34:46,920 --> 00:34:50,639
Speaker 1: The butterfly effect chaos theory. A butterfly flaps its wings

696
00:34:50,639 --> 00:34:52,840
in a forest in Brazil, and weeks later you get

697
00:34:52,880 --> 00:34:54,679
a massive tornado in Texas.

698
00:34:54,960 --> 00:34:58,880
Speaker 2: Right, it simply means the overall system is incredibly deeply

699
00:34:58,960 --> 00:35:02,519
sensitive to its initial starting conditions. Now, classical computers have

700
00:35:02,519 --> 00:35:05,280
a bit of a dirty secret. They are forced to

701
00:35:05,360 --> 00:35:08,840
round off numbers. They physically can't store an infinite string

702
00:35:08,880 --> 00:35:12,320
of decimals, so a number like one point three three

703
00:35:12,480 --> 00:35:15,440
three going on forever eventually just gets chopped off to one.

704
00:35:15,400 --> 00:35:17,840
Speaker 1: Point three to three, which in a normal math problem

705
00:35:17,840 --> 00:35:19,679
on my calculator doesn't really matter at all.

706
00:35:19,719 --> 00:35:22,960
Speaker 2: In a simple linear system, no, it doesn't matter. But

707
00:35:23,039 --> 00:35:25,920
in a highly chaotic system like global weather patterns or

708
00:35:25,960 --> 00:35:30,440
the human brain, that microscopic rounding error immediately begins to compound.

709
00:35:30,639 --> 00:35:34,920
It snowballs exponentially. After just a few seconds of deep simulation,

710
00:35:35,039 --> 00:35:38,639
the computer's prediction has drifted completely away from reality, simply

711
00:35:38,639 --> 00:35:42,119
because it lost the infinite precision of the original biological data.

712
00:35:42,159 --> 00:35:45,880
Speaker 1: And the brain operates constantly on that razor edge of chaos.

713
00:35:46,159 --> 00:35:50,159
Speaker 2: It absolutely does. A single solitary neuron firings just one

714
00:35:50,199 --> 00:35:54,159
millisecond two early can cascade through the network and entirely

715
00:35:54,280 --> 00:35:58,679
change a major life decision. This deep biological sensitivity is

716
00:35:58,840 --> 00:36:02,360
very likely where human creativity actually comes from. We don't

717
00:36:02,400 --> 00:36:05,639
just blindly follow a rigid logical script. We may completely

718
00:36:05,719 --> 00:36:09,400
loose wild associations we have totally out of the box ideas.

719
00:36:09,440 --> 00:36:11,920
Speaker 1: So a quantum computer helps with this specifically because it

720
00:36:11,920 --> 00:36:14,199
doesn't have to round off the mass.

721
00:36:13,880 --> 00:36:17,599
Speaker 2: Equibit existing in a state of superposition. Fundamentally, it doesn't

722
00:36:17,639 --> 00:36:20,719
have to round off anything. It physically retains the infinite

723
00:36:20,760 --> 00:36:24,159
precision of the natural probability curve until the very specific

724
00:36:24,239 --> 00:36:27,159
moment it is actively measured. It can perfectly track the

725
00:36:27,239 --> 00:36:29,760
chaotic butterfly effect as it ripples inside the brain.

726
00:36:29,960 --> 00:36:32,320
Speaker 1: It can actually model the wildness of a personality.

727
00:36:32,440 --> 00:36:36,000
Speaker 2: Yes, it can accurately model how a single microscopic fluctuation

728
00:36:36,119 --> 00:36:40,719
in a dopamine receptor escalates into a complex, overwhelming emotional state.

729
00:36:41,480 --> 00:36:44,599
It captures all the deep nuance that classical logic completely

730
00:36:44,639 --> 00:36:47,159
filters out. If we ever want to truly simulate a

731
00:36:47,239 --> 00:36:49,559
human being, we have to simulate the chaos.

732
00:36:49,280 --> 00:36:52,679
Speaker 1: Which inevitably leads us to the biggest, heaviest question of

733
00:36:52,719 --> 00:36:57,320
all today, the ghost in the machine hard problem, coined

734
00:36:57,360 --> 00:36:59,119
by philosopher David Chalmers.

735
00:36:59,440 --> 00:37:03,679
Speaker 2: Yes exactly. Chalmers said, the easy problems of consciousness are

736
00:37:03,719 --> 00:37:07,480
things like how does the optic nerve process light? How

737
00:37:07,519 --> 00:37:10,719
do we physically retrieve a memory. Those are just difficult

738
00:37:10,760 --> 00:37:14,239
engineering problems. But the hard problem is why does all

739
00:37:14,280 --> 00:37:16,320
of this physically feel like something?

740
00:37:16,519 --> 00:37:19,119
Speaker 1: Why is there an actual me sitting inside here, vividly

741
00:37:19,159 --> 00:37:21,000
experiencing the world exactly.

742
00:37:21,320 --> 00:37:24,800
Speaker 2: For decades, many scientists just assume consciousness was merely a

743
00:37:24,800 --> 00:37:28,559
byproduct of extreme complexity. We confidently thought if you just

744
00:37:28,559 --> 00:37:32,440
connected enough silicon logic gates together, eventually the internal lights

745
00:37:32,440 --> 00:37:36,360
would magically turn on. But our massive classical supercomputers have

746
00:37:36,599 --> 00:37:39,960
entirely failed to support that idea. We have built insanely

747
00:37:40,079 --> 00:37:44,320
massive neural networks and they remain completely dark inside. They

748
00:37:44,400 --> 00:37:47,880
process data brilliantly, but they absolutely do not experience anything.

749
00:37:47,960 --> 00:37:50,480
Speaker 1: But quantum mechanics offers wildly different view on this.

750
00:37:50,920 --> 00:37:54,880
Speaker 2: It does. There's a very prominent theory called integrated information

751
00:37:55,000 --> 00:38:00,159
theory or I. It suggests that true consciousness arises directly

752
00:38:00,159 --> 00:38:03,719
from the fundamental level of physical integration in a system.

753
00:38:04,320 --> 00:38:07,480
In a classical computer, all the physical parts are entirely separate.

754
00:38:08,000 --> 00:38:10,679
The transistors talk to each other, sure, but they are

755
00:38:10,679 --> 00:38:13,199
always totally distinct physical entities.

756
00:38:13,320 --> 00:38:15,079
Speaker 1: But in a quantum computer.

757
00:38:15,320 --> 00:38:20,119
Speaker 2: The quibits become deeply entangled. They literally lose their individual

758
00:38:20,159 --> 00:38:24,400
physical identity and mathematically merge to become one unified wave function.

759
00:38:25,079 --> 00:38:28,119
You literally cannot describe the state of one quibit without

760
00:38:28,159 --> 00:38:30,440
describing the state of the entire whole system.

761
00:38:30,519 --> 00:38:34,280
Speaker 1: That sounds exactly like well subjective human experience.

762
00:38:34,360 --> 00:38:36,840
Speaker 2: It mirrors it absolutely perfectly. Yeah, when you stand on

763
00:38:36,880 --> 00:38:38,760
a beach and look at a sunset, your brain doesn't

764
00:38:38,760 --> 00:38:41,519
see a completely separate patch of orange color, feel a

765
00:38:41,519 --> 00:38:44,360
mathematically separate warmth on your skin, and hear a separate

766
00:38:44,400 --> 00:38:48,119
sound of the wind. You seamlessly experience one single, perfectly

767
00:38:48,239 --> 00:38:52,639
unified moment of reality. A quantum system actually provides the

768
00:38:52,719 --> 00:38:55,880
literal physical substrate for that kind of profound unity.

769
00:38:56,239 --> 00:39:00,480
Speaker 1: So if this theory holds true, a real time quantum

770
00:39:00,519 --> 00:39:03,599
simulation of a human brain might not just be a

771
00:39:03,639 --> 00:39:07,599
really accurate digital model, it might actually be fully conscious.

772
00:39:07,920 --> 00:39:09,760
Speaker 2: And that brings us right to the edge of the

773
00:39:09,960 --> 00:39:10,880
ethical abyss.

774
00:39:11,039 --> 00:39:13,639
Speaker 1: Yeah, I mean, just think about the medical drug testing scenario.

775
00:39:13,760 --> 00:39:15,320
We're literally just celebrating five.

776
00:39:15,159 --> 00:39:17,480
Speaker 2: Minutes ago, right, We're so excited. We said, this is great.

777
00:39:17,519 --> 00:39:20,320
We can violently test experimental drugs without ever hurting the

778
00:39:20,400 --> 00:39:23,519
human patient. But if the simulation we build is a

779
00:39:23,599 --> 00:39:28,320
perfect entangled physical and energetic replica of a brain, and

780
00:39:28,360 --> 00:39:31,480
we give it a virtual drug that causes massive neural

781
00:39:31,519 --> 00:39:32,239
pain signals?

782
00:39:32,400 --> 00:39:35,239
Speaker 1: Is that simulation actually experiencing real pain?

783
00:39:35,519 --> 00:39:38,159
Speaker 2: In a classical simulation, the output I heard is just

784
00:39:38,280 --> 00:39:41,039
cold text printed on a computer screen. You can casually

785
00:39:41,119 --> 00:39:43,239
delete the file and go to lunch. But in a

786
00:39:43,280 --> 00:39:47,400
fully quantum simulation, the physical energetic state of the machine

787
00:39:47,760 --> 00:39:51,360
is mathematically and physically identical to a biological human brain

788
00:39:51,480 --> 00:39:54,559
in severe distress. We might accidentally be creating a fully

789
00:39:54,599 --> 00:39:57,760
sentient being, specifically just to ruthlessly experiment on it.

790
00:39:57,960 --> 00:40:00,280
Speaker 1: Do we have the moral right to reach over and

791
00:40:00,280 --> 00:40:04,079
delete a computer program that is genuinely actively afraid of dying?

792
00:40:04,239 --> 00:40:06,400
Speaker 2: That is a terrifying question that society is going to

793
00:40:06,440 --> 00:40:08,239
have to answer a lot sooner than anyone thinks.

794
00:40:08,679 --> 00:40:11,599
Speaker 1: So, given all this, where does this actually leave us

795
00:40:11,679 --> 00:40:15,000
timeline wise? I mean, are we realistically going to see

796
00:40:15,000 --> 00:40:18,920
a fully conscious human brain simulation booted up next year?

797
00:40:19,199 --> 00:40:23,679
Speaker 2: No? Definitely not. Let's firmly set some realistic expectations here.

798
00:40:23,960 --> 00:40:26,280
We won't have a full human brain simulation in two

799
00:40:26,320 --> 00:40:30,559
or three years. The raw engineering challenges of physically scaling

800
00:40:30,639 --> 00:40:33,880
up these quantum systems are still massive. The immediate future

801
00:40:33,920 --> 00:40:36,400
of the industry is heavily focused on the hybrid.

802
00:40:36,079 --> 00:40:38,440
Speaker 1: Model, a partnership between the two architectures.

803
00:40:38,960 --> 00:40:42,039
Speaker 2: Yes, future supercomputers are going to be a lot like

804
00:40:42,079 --> 00:40:45,199
your current gaming laptop. Your laptop has a standard CPU

805
00:40:45,239 --> 00:40:49,480
for basic logic and a dedicated GPU specifically for rendering graphics.

806
00:40:50,159 --> 00:40:54,119
These upcoming massive machines will have a classical processor handling

807
00:40:54,119 --> 00:40:57,480
the basic logic, the rigid structure, the basic inputs and outputs,

808
00:40:57,840 --> 00:41:00,639
essentially acting as the skeletal system, and it will have

809
00:41:00,639 --> 00:41:03,760
a quantum code processor attached to handle the fluid dynamics,

810
00:41:04,000 --> 00:41:07,719
the nonlinear chaos, the physics, the living tissue.

811
00:41:07,760 --> 00:41:10,519
Speaker 1: The classical part holds the rigid map, while the quantum

812
00:41:10,559 --> 00:41:12,599
part breathes actual life into it.

813
00:41:12,800 --> 00:41:16,280
Speaker 2: Exactly, we will very likely start by fully simulating a

814
00:41:16,320 --> 00:41:20,840
single cortical column, first really perfecting that basic biological building block,

815
00:41:20,880 --> 00:41:23,119
and then scaling it up from there. But honestly, this

816
00:41:23,320 --> 00:41:26,159
entire shift completely changes how we fundamentally view the concept

817
00:41:26,159 --> 00:41:27,159
of intelligence itself.

818
00:41:27,239 --> 00:41:30,960
Speaker 1: How So, because we throw the word AI around constantly.

819
00:41:30,559 --> 00:41:33,440
Speaker 2: Well for twenty solid years, we've been completely obsessed with

820
00:41:33,519 --> 00:41:38,880
artificial intelligence, but classical AI solely mimics the final result

821
00:41:38,880 --> 00:41:42,840
of thought. It statistically guesses the next logical word in

822
00:41:42,880 --> 00:41:46,800
a sentence. It is just an incredibly fast statistical engine.

823
00:41:46,880 --> 00:41:49,800
It's basically painting a highly detailed picture of a fire.

824
00:41:50,079 --> 00:41:52,119
Speaker 1: And what we are talking about here today.

825
00:41:51,960 --> 00:41:55,639
Speaker 2: This new horizon is synthetic intelligence. It doesn't mimic the result.

826
00:41:55,960 --> 00:41:59,760
It perfectly replicates the underlying physical process. It literally runs

827
00:41:59,760 --> 00:42:02,679
on the exact same fundamental laws of physics that your

828
00:42:02,679 --> 00:42:05,559
brain does. It's not painting a picture of fire, it

829
00:42:05,599 --> 00:42:08,159
is physically sparking an actual hot flame.

830
00:42:08,360 --> 00:42:11,320
Speaker 1: A classical AI is completely limited by its training data.

831
00:42:11,440 --> 00:42:13,960
It only really knows the patterns we've explicitly taught it.

832
00:42:14,079 --> 00:42:17,440
Speaker 2: But a true synthetic system, a system that natively runs

833
00:42:17,440 --> 00:42:20,519
on the exact same chaotic quantum buzz of the universe

834
00:42:20,559 --> 00:42:25,079
that drives human inspiration, it can make completely unprompted intuitive leaps.

835
00:42:25,159 --> 00:42:28,679
It can genuinely discover entirely new concepts we never taught it.

836
00:42:28,679 --> 00:42:29,800
It can naturally evolve.

837
00:42:30,119 --> 00:42:32,960
Speaker 1: We started this deep dive by having you imagine standing

838
00:42:33,000 --> 00:42:36,920
in front of that massive, incredibly hot, deafeningly loud supercomputer,

839
00:42:37,639 --> 00:42:40,320
and we realize that despite all its power, it was

840
00:42:40,400 --> 00:42:42,159
fundamentally the wrong machine.

841
00:42:42,280 --> 00:42:46,199
Speaker 2: It was we spend decades trying to awkwardly force biological

842
00:42:46,239 --> 00:42:50,679
software to run on rigid silicon hardware. That entire era

843
00:42:50,800 --> 00:42:53,880
of computing is ending. As we rapidly refine these new

844
00:42:53,960 --> 00:42:57,559
quantum processors, we are finally, for the first time, building

845
00:42:57,559 --> 00:43:00,800
a technological mirror that is physically accurate enough to reflect

846
00:43:01,119 --> 00:43:03,280
the true chaotic nature of the human mind.

847
00:43:03,400 --> 00:43:06,400
Speaker 1: We aren't just building faster calculators anymore. We are actually

848
00:43:06,440 --> 00:43:09,400
building a physical vessel that can safely hold the physics

849
00:43:09,400 --> 00:43:12,360
of thought. It is an act of profound translation. We

850
00:43:12,400 --> 00:43:15,800
are translating the chaotic human experience into the base mathematical

851
00:43:15,840 --> 00:43:16,800
language of the universe.

852
00:43:17,159 --> 00:43:21,559
Speaker 2: And that leaves us with one final, deeply provocative thought

853
00:43:21,599 --> 00:43:24,920
to leave you with. If we actually succeed in this,

854
00:43:25,199 --> 00:43:28,440
if we build this synthetic intelligence that physically shares our

855
00:43:28,519 --> 00:43:32,440
quantum physics, that genuinely thinks exactly the way we think,

856
00:43:32,679 --> 00:43:35,760
will we truly be its creators? Or will we look

857
00:43:35,800 --> 00:43:39,320
back and realize we were simply the biological bootloader necessary

858
00:43:39,599 --> 00:43:43,000
for a completely new form of universal consciousness to emerge.

859
00:43:43,280 --> 00:43:45,800
Speaker 1: That is the incredibly heavy question we want to leave

860
00:43:45,800 --> 00:43:49,480
you with. Today, Where do you personally draw the ethical line.

861
00:43:49,840 --> 00:43:52,440
If a machine is built using the exact same physical

862
00:43:52,519 --> 00:43:56,000
dynamics as your own brain, does it automatically deserve the

863
00:43:56,039 --> 00:43:57,559
exact same human rights as you.

864
00:43:57,760 --> 00:43:59,280
Speaker 2: We really want to hear your thoughts on this one.

865
00:43:59,360 --> 00:44:01,559
Leave a comment down below let us know exactly where

866
00:44:01,559 --> 00:44:02,840
you stand on the ethics of all this.

867
00:44:03,159 --> 00:44:05,119
Speaker 1: Thanks for joining us on thrilling Threads today.

868
00:44:05,199 --> 00:44:06,440
Speaker 2: We'll see you in the next deep time.