Skip to content
The Chat Archive
gpt-4o-2024-08-06· OpenAI WildChat-4.8M (AI2)

please spellcheck this subtitle and keep the timecodes unchanged: 1 00:00:00,000 -->…

  1. User

    please spellcheck this subtitle and keep the timecodes unchanged:

    1 00:00:00,000 --> 00:00:02,320 What if I told you we could like live through the

    2 00:00:02,321 --> 00:00:05,740 next century of progress in just the next decade?

    3 00:00:06,180 --> 00:00:09,640 That's the idea behind what AI researcher Dario

    4 00:00:09,641 --> 00:00:12,920 Emma Day calls the compressed 21st century.

    5 00:00:13,080 --> 00:00:15,700 And here's why this particular prediction is so intriguing.

    6 00:00:16,300 --> 00:00:16,340 Right.

    7 00:00:16,600 --> 00:00:20,220 Emma Day isn't your typical Silicon Valley tech optimist.

    8 00:00:20,300 --> 00:00:20,520 Right.

    9 00:00:20,740 --> 00:00:26,060 He's both a biologist and an AI expert who runs anthropic.

    10 00:00:26,860 --> 00:00:26,900 Yeah.

    11 00:00:26,901 --> 00:00:30,860 A company known for its focus on AI risks.

    12 00:00:31,380 --> 00:00:31,400 Right.

    13 00:00:31,620 --> 00:00:32,640 Not utopian promises.

    14 00:00:32,900 --> 00:00:33,320 Exactly.

    15 00:00:33,880 --> 00:00:35,500 He's not known for sugar coating things.

    16 00:00:35,720 --> 00:00:35,780 No.

    17 00:00:36,020 --> 00:00:39,400 So when Emma Day says powerful AI could deliver the next 50

    18 00:00:39,401 --> 00:00:42,420 to 100 years of breakthroughs in just 5 to 10,

    19 00:00:43,040 --> 00:00:43,660 it's not just hype.

    20 00:00:44,060 --> 00:00:44,080 Right.

    21 00:00:44,180 --> 00:00:47,200 It's him saying, hold on, things are about to get really wild.

    22 00:00:47,460 --> 00:00:51,820 And powerful AI in Emma Day's world isn't some vague concept.

    23 00:00:52,280 --> 00:00:52,300 Okay.

    24 00:00:52,301 --> 00:00:56,760 He's talking about AI with genius level intellect across multiple fields.

    25 00:00:57,000 --> 00:00:57,300 Oh, wow.

    26 00:00:57,360 --> 00:00:58,360 Not just one party trick.

    27 00:00:59,020 --> 00:01:01,120 Think about a team of the world's brightest minds.

    28 00:01:01,460 --> 00:01:01,640 Okay.

    29 00:01:02,200 --> 00:01:06,820 All working together in a data center, designing experiments, inventing tools,

    30 00:01:07,940 --> 00:01:08,940 essentially operating.

    31 00:01:09,380 --> 00:01:13,800 Like... It's like having a country of geniuses in a data center as he puts it.

    32 00:01:13,980 --> 00:01:14,060 Yeah.

    33 00:01:14,440 --> 00:01:16,080 Which honestly... Good way to put it.

    34 00:01:16,360 --> 00:01:17,920 Doesn't sound that far fetched anymore.

    35 00:01:18,080 --> 00:01:18,960 It really doesn't.

    36 00:01:19,020 --> 00:01:22,080 The big question is, how does this actually play out?

    37 00:01:22,081 --> 00:01:26,140 In the real world, where do we see the most radical upside?

    38 00:01:27,180 --> 00:01:30,740 Well, Emma Day points to biology and health as

    39 00:01:30,741 --> 00:01:33,520 being ripe for this kind of AI acceleration.

    40 00:01:33,960 --> 00:01:34,960 Interesting.

    41 00:01:35,060 --> 00:01:37,320 And this is where his dual background as both a

    42 00:01:37,321 --> 00:01:41,020 biologist and an AI expert is so insightful.

    43 00:01:41,480 --> 00:01:41,580 Yeah.

    44 00:01:41,860 --> 00:01:44,680 He understands the pace of scientific discovery.

    45 00:01:45,120 --> 00:01:45,180 Okay.

    46 00:01:45,480 --> 00:01:49,320 How it often hinges on these rare groundbreaking moments.

    47 00:01:49,321 --> 00:01:49,540 Uh-huh.

    48 00:01:49,640 --> 00:01:50,680 The aha moment.

    49 00:01:50,880 --> 00:01:52,116 Think about CRISPR, for example.

    50 00:01:52,140 --> 00:01:54,620 Right. That completely revolutionized gene editing.

    51 00:01:55,320 --> 00:01:57,680 And it's not hard to imagine how AI could help

    52 00:01:57,681 --> 00:02:00,280 us uncover those kinds of breakthroughs even faster.

    53 00:02:00,640 --> 00:02:01,120 Exactly.

    54 00:02:01,540 --> 00:02:05,720 AI can analyze massive data sets, identify hidden patterns,

    55 00:02:06,160 --> 00:02:08,360 even design and run experiments.

    56 00:02:08,720 --> 00:02:08,860 Wow.

    57 00:02:09,400 --> 00:02:10,720 At a pace we could only dream of.

    58 00:02:10,780 --> 00:02:11,780 Yeah.

    59 00:02:11,860 --> 00:02:14,460 Emma Day's argument is that AI isn't magic.

    60 00:02:15,260 --> 00:02:17,380 But it can drastically compress the timeline.

    61 00:02:17,840 --> 00:02:18,060 Okay.

    62 00:02:18,440 --> 00:02:19,900 From Eureka moment.

    63 00:02:20,220 --> 00:02:20,360 Yeah.

    64 00:02:20,820 --> 00:02:23,140 To real world impact.

    65 00:02:23,780 --> 00:02:27,380 Okay, so it's not like AI is going to magically cure cancer overnight.

    66 00:02:27,880 --> 00:02:29,136 There are still limitations, right?

    67 00:02:29,160 --> 00:02:29,360 Of course.

    68 00:02:29,460 --> 00:02:30,460 Experiments take time.

    69 00:02:30,780 --> 00:02:30,940 Yes.

    70 00:02:31,120 --> 00:02:33,900 And we're still bound by the laws of physics.

    71 00:02:34,260 --> 00:02:36,900 Emma Day isn't saying we can defy the laws of nature.

    72 00:02:37,280 --> 00:02:37,500 Okay.

    73 00:02:37,501 --> 00:02:43,401 But what he's saying is that we can drastically accelerate the pace of discovery.

    74 00:02:43,820 --> 00:02:46,547 Imagine a world where instead of taking 50

    75 00:02:46,548 --> 00:02:49,560 years to achieve a breakthrough takes five.

    76 00:02:49,900 --> 00:02:50,000 Wow.

    77 00:02:50,540 --> 00:02:52,500 That's the kind of compression he's talking about.

    78 00:02:52,540 --> 00:02:55,060 So what does that actually look like for someone like you or me?

    79 00:02:55,500 --> 00:02:55,640 Right.

    80 00:02:55,980 --> 00:02:59,600 What changes in a world where the pace of progress is ten times faster?

    81 00:03:00,280 --> 00:03:01,060 Think about it this way.

    82 00:03:01,240 --> 00:03:01,320 Okay.

    83 00:03:01,500 --> 00:03:04,720 In this compressed 21st century diseases like cancer,

    84 00:03:05,060 --> 00:03:07,500 Alzheimer's, even inherited genetic conditions.

    85 00:03:07,780 --> 00:03:08,320 Oh, wow.

    86 00:03:08,620 --> 00:03:10,300 They could largely be problems of the past.

    87 00:03:10,500 --> 00:03:10,700 Okay.

    88 00:03:10,840 --> 00:03:12,440 Not just within our children's lifetime.

    89 00:03:12,560 --> 00:03:12,680 Yeah.

    90 00:03:12,860 --> 00:03:14,040 But potentially within ours.

    91 00:03:14,260 --> 00:03:14,380 Wow.

    92 00:03:15,100 --> 00:03:16,500 That's not just a medical marvel.

    93 00:03:16,820 --> 00:03:16,900 Right.

    94 00:03:17,060 --> 00:03:19,200 That's a complete societal game changer.

    95 00:03:19,300 --> 00:03:19,880 It really is.

    96 00:03:20,120 --> 00:03:20,680 And get this.

    97 00:03:20,980 --> 00:03:21,100 Yeah.

    98 00:03:21,520 --> 00:03:25,240 Emma Day even suggests that we could double the human lifespan again.

    99 00:03:25,640 --> 00:03:26,020 It's true.

    100 00:03:26,060 --> 00:03:27,700 Based on the progress we've already made.

    101 00:03:27,860 --> 00:03:27,980 Yeah.

    102 00:03:28,160 --> 00:03:31,440 It's a mind-boggling concept with huge implications.

    103 00:03:31,820 --> 00:03:31,980 Right.

    104 00:03:31,981 --> 00:03:37,080 Imagine the impact on everything from family structures to retirement plans.

    105 00:03:37,081 --> 00:03:37,500 Wow.

    106 00:03:38,100 --> 00:03:41,100 Not to mention the ethical and philosophical questions it raises.

    107 00:03:41,600 --> 00:03:41,780 Right.

    108 00:03:42,040 --> 00:03:44,740 And it's not just about living longer. It's about living better.

    109 00:03:45,140 --> 00:03:45,520 Exactly.

    110 00:03:45,840 --> 00:03:48,220 Emma Day talks about this incredible potential for

    111 00:03:48,221 --> 00:03:54,020 AI to revolutionize not just our physical health, but also our mental well-being.

    112 00:03:54,520 --> 00:03:54,540 Exactly.

    113 00:03:55,180 --> 00:03:59,120 He sees the compressed 21st happening for the mind too.

    114 00:03:59,420 --> 00:03:59,680 Interesting.

    115 00:03:59,940 --> 00:04:00,520 Not just the body.

    116 00:04:00,521 --> 00:04:00,960 Okay.

    117 00:04:01,360 --> 00:04:03,960 And this is where recent advancements in something

    118 00:04:03,961 --> 00:04:06,680 called AI interpretability come into play.

    119 00:04:07,160 --> 00:04:07,400 Okay.

    120 00:04:07,540 --> 00:04:09,460 I have to admit, I'm not familiar with that term.

    121 00:04:09,500 --> 00:04:09,900 Okay, sure.

    122 00:04:10,500 --> 00:04:15,540 What exactly AI interpretability and how does it relate to our minds?

    123 00:04:16,060 --> 00:04:22,320 So, in simple terms, AI interpretability is about understanding how AI thinks.

    124 00:04:22,700 --> 00:04:22,900 Okay.

    125 00:04:23,180 --> 00:04:24,220 Not just what it produces.

    126 00:04:24,680 --> 00:04:24,800 Okay.

    127 00:04:25,100 --> 00:04:25,760 Think of it like this.

    128 00:04:25,761 --> 00:04:26,060 Okay.

    129 00:04:26,280 --> 00:04:33,280 Imagine having a window into the decision -making process of a brilliant scientist.

    130 00:04:33,780 --> 00:04:33,960 Okay.

    131 00:04:34,480 --> 00:04:36,080 Now, apply that to AI.

    132 00:04:36,920 --> 00:04:37,280 Interesting.

    133 00:04:37,740 --> 00:04:41,180 By understanding how AI arrives at its conclusions,

    134 00:04:41,700 --> 00:04:47,620 we can potentially unlock secrets of the human brain far more quickly.

    135 00:04:48,020 --> 00:04:50,720 So, it's almost like AI is holding up a mirror to our own minds.

    136 00:04:51,160 --> 00:04:52,196 That's a great way to put it.

    137 00:04:52,220 --> 00:04:54,340 Helping us understand ourselves better in the process.

    138 00:04:55,120 --> 00:04:55,640 Precisely.

    139 00:04:55,900 --> 00:05:00,280 And this deeper understanding could have profound implications for mental health.

    140 00:05:00,920 --> 00:05:03,620 A mode is very clear that he sees mental well

    141 00:05:03,621 --> 00:05:07,440 -being as being just as crucial as physical health, if not more so.

    142 00:05:07,640 --> 00:05:09,320 I think a lot of people can relate to that.

    143 00:05:09,380 --> 00:05:11,216 We're talking about tackling things like depression,

    144 00:05:11,240 --> 00:05:14,700 anxiety addiction, even personality disorders.

    145 00:05:15,040 --> 00:05:15,100 Yeah.

    146 00:05:15,360 --> 00:05:17,780 With the same urgency and resources that we

    147 00:05:17,781 --> 00:05:20,360 currently direct towards cancer or heart disease.

    148 00:05:20,520 --> 00:05:20,920 Exactly.

    149 00:05:20,921 --> 00:05:21,240 Yeah.

    150 00:05:21,440 --> 00:05:25,640 Amade's vision is to address these issues, not just by treating symptoms.

    151 00:05:26,000 --> 00:05:26,080 Right.

    152 00:05:26,340 --> 00:05:29,900 But by understanding and potentially addressing the root causes.

    153 00:05:30,300 --> 00:05:30,360 Okay.

    154 00:05:30,520 --> 00:05:34,060 And he sees AI as a key driver of that revolution.

    155 00:05:34,560 --> 00:05:37,160 So, how does he see that unfolding?

    156 00:05:37,820 --> 00:05:41,080 How does AI actually help us achieve this mental health revolution?

    157 00:05:41,520 --> 00:05:44,840 Amade lays out like four main avenues.

    158 00:05:45,240 --> 00:05:45,380 Okay.

    159 00:05:45,660 --> 00:05:47,380 Where AI can make a huge difference.

    160 00:05:47,760 --> 00:05:47,900 Okay.

    161 00:05:47,901 --> 00:05:51,600 First, imagine AI accelerating the development of new medications.

    162 00:05:52,280 --> 00:05:53,120 And genetic therapies.

    163 00:05:53,121 --> 00:05:53,940 Mm-hmm.

    164 00:05:53,941 --> 00:05:56,520 Fine-tune to an individual's brain chemistry.

    165 00:05:57,240 --> 00:05:59,260 It's like personalized medicine, but for the mind.

    166 00:05:59,900 --> 00:06:00,020 Exactly.

    167 00:06:00,220 --> 00:06:01,980 No more trial and error with antidepressants.

    168 00:06:02,220 --> 00:06:04,380 It's about getting the right treatment from the get-go.

    169 00:06:04,520 --> 00:06:05,520 Exactly.

    170 00:06:05,600 --> 00:06:10,980 Second, Amade talks about AI enabling incredibly precise interventions in the brain.

    171 00:06:11,160 --> 00:06:11,340 Mm-hmm.

    172 00:06:11,480 --> 00:06:13,340 And we're not talking about crude surgery here.

    173 00:06:13,560 --> 00:06:13,600 Right.

    174 00:06:13,601 --> 00:06:17,460 Think about using AI to target specific neural pathways.

    175 00:06:17,880 --> 00:06:18,000 Okay.

    176 00:06:18,360 --> 00:06:20,680 Potentially addressing conditions like Parkinson's.

    177 00:06:21,120 --> 00:06:21,240 Yeah.

    178 00:06:21,260 --> 00:06:22,800 Or even severe OCD.

    179 00:06:22,960 --> 00:06:23,600 Wow.

    180 00:06:23,980 --> 00:06:24,480 At their root.

    181 00:06:24,780 --> 00:06:27,960 It's like having a microscopic repair crew for the brain.

    182 00:06:28,400 --> 00:06:28,680 Yes.

    183 00:06:28,920 --> 00:06:32,780 Guided by AI to fix things with incredible accuracy.

    184 00:06:33,260 --> 00:06:34,316 That's a great way to put it.

    185 00:06:34,340 --> 00:06:37,020 What's fascinating is how Amade connects this.

    186 00:06:37,300 --> 00:06:37,440 Yeah.

    187 00:06:37,620 --> 00:06:42,340 Not just to fixing what's broken out to potentially enhancing our mental experience.

    188 00:06:42,341 --> 00:06:42,880 That's right.

    189 00:06:43,020 --> 00:06:44,420 He doesn't shy away from that idea.

    190 00:06:44,540 --> 00:06:44,640 Okay.

    191 00:06:44,980 --> 00:06:50,040 His third avenue is all about AI helping us build incredibly accurate

    192 00:06:50,041 --> 00:06:53,480 computational models of how our minds work.

    193 00:06:53,920 --> 00:06:58,340 Imagine using these models to not only understand conditions like autism.

    194 00:06:58,820 --> 00:06:59,040 Okay.

    195 00:06:59,120 --> 00:07:00,120 Or schizophrenia better.

    196 00:07:00,600 --> 00:07:00,680 Right.

    197 00:07:00,960 --> 00:07:06,040 But to develop personalized learning plans that optimize how we learn and create.

    198 00:07:06,620 --> 00:07:08,100 So instead of just treating deficits.

    199 00:07:08,680 --> 00:07:08,820 Exactly.

    200 00:07:08,821 --> 00:07:13,040 We're talking about using AI to unlock our full cognitive potential.

    201 00:07:13,360 --> 00:07:14,416 I like that way I put it again.

    202 00:07:14,440 --> 00:07:15,600 That's incredibly exciting.

    203 00:07:15,840 --> 00:07:16,920 But also a little daunting.

    204 00:07:17,300 --> 00:07:17,360 Yeah.

    205 00:07:17,600 --> 00:07:19,480 What about the ethical implications of all this?

    206 00:07:19,620 --> 00:07:21,800 So that's something Amadei addresses head on.

    207 00:07:22,040 --> 00:07:22,240 Okay.

    208 00:07:22,540 --> 00:07:25,381 He doesn't pretend these advancements won't come with tough questions.

    209 00:07:25,580 --> 00:07:25,720 Right.

    210 00:07:26,160 --> 00:07:28,992 But he argues that the potential benefits, particularly

    211 00:07:28,993 --> 00:07:33,600 in alleviating suffering, are too great to ignore.

    212 00:07:33,960 --> 00:07:34,040 Right.

    213 00:07:34,041 --> 00:07:36,840 And his fourth avenue is perhaps the most accessible.

    214 00:07:37,300 --> 00:07:37,480 Okay.

    215 00:07:37,980 --> 00:07:40,270 AI powered therapy available to everyone.

    216 00:07:40,920 --> 00:07:41,140 Okay.

    217 00:07:41,360 --> 00:07:43,320 Not just those who can afford traditional therapy.

    218 00:07:43,480 --> 00:07:44,480 That's huge.

    219 00:07:45,540 --> 00:07:48,533 Imagine a world where everyone has access

    220 00:07:48,534 --> 00:07:52,080 to affordable, effective mental health care.

    221 00:07:52,380 --> 00:07:53,440 It's a powerful idea.

    222 00:07:53,780 --> 00:07:55,060 Tailored to their unique needs.

    223 00:07:55,520 --> 00:07:55,700 Mm-hmm.

    224 00:07:55,920 --> 00:07:59,440 That could be a true game changer for global well-being.

    225 00:08:00,040 --> 00:08:03,200 And that brings us to another crucial aspect of Amadei's vision.

    226 00:08:03,201 --> 00:08:03,700 Okay.

    227 00:08:04,060 --> 00:08:08,760 The impact of this compressed 21st century on global inequality.

    228 00:08:09,380 --> 00:08:09,580 Okay.

    229 00:08:09,800 --> 00:08:14,420 He's very clear that while AI has the potential to lift billions out of poverty.

    230 00:08:14,840 --> 00:08:14,920 Right.

    231 00:08:15,200 --> 00:08:16,700 It requires intentional effort.

    232 00:08:17,140 --> 00:08:17,340 Right.

    233 00:08:17,540 --> 00:08:19,020 Because technology itself is neutral.

    234 00:08:19,200 --> 00:08:19,360 Exactly.

    235 00:08:19,700 --> 00:08:21,920 It's how we choose to develop it and deploy it.

    236 00:08:22,040 --> 00:08:22,280 That's right.

    237 00:08:22,520 --> 00:08:25,360 That will determine whether it benefits everyone or just widens

    238 00:08:25,361 --> 00:08:27,260 the gap between the haves and the have-nots.

    239 00:08:27,340 --> 00:08:27,860 Precisely.

    240 00:08:27,960 --> 00:08:28,100 Okay.

    241 00:08:28,580 --> 00:08:30,600 And Amadei sees this playing out on several levels.

    242 00:08:30,601 --> 00:08:30,860 Yep.

    243 00:08:31,140 --> 00:08:33,340 First, there's the issue of distributing those

    244 00:08:33,341 --> 00:08:35,200 incredible health advances we talked about.

    245 00:08:35,240 --> 00:08:35,260 Okay.

    246 00:08:35,720 --> 00:08:37,360 Think about the eradication of smallpox.

    247 00:08:37,640 --> 00:08:38,300 Mm-hmm.

    248 00:08:38,301 --> 00:08:40,740 A disease that once ravaged entire populations.

    249 00:08:41,240 --> 00:08:41,440 Yeah.

    250 00:08:42,020 --> 00:08:44,240 With AI-powered strategies and logistics.

    251 00:08:44,740 --> 00:08:44,820 Right.

    252 00:08:45,220 --> 00:08:48,920 We could do the same for diseases like malaria, tuberculosis.

    253 00:08:49,420 --> 00:08:51,781 And even common parasitic infections that

    254 00:08:51,782 --> 00:08:54,040 disproportionately affect developing countries.

    255 00:08:54,300 --> 00:08:57,660 It's like a global health initiative, but supercharged by AI.

    256 00:08:58,040 --> 00:08:58,260 Exactly.

    257 00:08:58,261 --> 00:09:01,400 And with AI-optimizing distribution networks.

    258 00:09:02,120 --> 00:09:05,480 Even the most remote communities could benefit from these breakthroughs.

    259 00:09:05,600 --> 00:09:06,600 Absolutely.

    260 00:09:06,920 --> 00:09:11,400 Amadei also talks about AI's potential to ignite a second green revolution.

    261 00:09:11,820 --> 00:09:12,820 Okay.

    262 00:09:12,960 --> 00:09:15,000 Imagine AI-optimizing crop yields.

    263 00:09:15,400 --> 00:09:15,780 Mm-hmm.

    264 00:09:15,920 --> 00:09:17,560 Developing drought-resistant seeds.

    265 00:09:17,920 --> 00:09:18,080 Right.

    266 00:09:18,420 --> 00:09:20,740 And revolutionizing food production and distribution.

    267 00:09:21,240 --> 00:09:21,320 Wow.

    268 00:09:21,740 --> 00:09:24,580 We're talking about tackling hunger on a global scale.

    269 00:09:25,120 --> 00:09:25,200 Sure.

    270 00:09:25,480 --> 00:09:27,500 Not just increasing profits for a select few.

    271 00:09:27,501 --> 00:09:29,160 That's a powerful vision.

    272 00:09:30,580 --> 00:09:32,360 But let's be realistic for a moment.

    273 00:09:32,540 --> 00:09:32,560 Okay.

    274 00:09:32,740 --> 00:09:36,261 We've seen how technology can sometimes exacerbate existing inequalities.

    275 00:09:36,340 --> 00:09:36,440 Right.

    276 00:09:36,800 --> 00:09:37,760 Even with the best of intentions.

    277 00:09:37,800 --> 00:09:38,800 Of course.

    278 00:09:39,220 --> 00:09:43,080 What about access to these advancements within developed countries?

    279 00:09:43,640 --> 00:09:43,680 Right.

    280 00:09:43,880 --> 00:09:48,420 How do we ensure that everyone benefits not just a wealthy elite?

    281 00:09:48,920 --> 00:09:51,480 Amadei acknowledges that this is a valid concern.

    282 00:09:51,820 --> 00:09:52,020 Okay.

    283 00:09:52,021 --> 00:09:57,620 He's not naive about the potential for AI to be used for profit over people.

    284 00:09:58,300 --> 00:10:01,280 But he argues that at least in democratic societies.

    285 00:10:02,040 --> 00:10:05,014 Market forces combined with citizen demand will

    286 00:10:05,015 --> 00:10:08,220 drive down costs and increase access over time.

    287 00:10:08,380 --> 00:10:08,600 Okay.

    288 00:10:09,140 --> 00:10:10,840 Think about how quickly cell phones went from

    289 00:10:10,841 --> 00:10:12,920 luxury items to then practically ubiquitous.

    290 00:10:13,420 --> 00:10:15,680 That's a good point. It seemed like almost overnight

    291 00:10:15,681 --> 00:10:17,880 everyone in their grandmother had a smartphone.

    292 00:10:18,560 --> 00:10:18,740 Yeah.

    293 00:10:18,880 --> 00:10:21,100 And it wasn't just about the technology becoming cheaper.

    294 00:10:21,101 --> 00:10:21,600 Right.

    295 00:10:21,760 --> 00:10:24,360 It was also about people recognizing the value.

    296 00:10:24,640 --> 00:10:25,000 Exactly.

    297 00:10:25,220 --> 00:10:26,400 Demanding to be connected.

    298 00:10:26,840 --> 00:10:26,980 Yes.

    299 00:10:27,160 --> 00:10:29,300 So you're saying a similar dynamic could play out?

    300 00:10:29,600 --> 00:10:29,920 Yes, right.

    301 00:10:30,100 --> 00:10:32,180 With these AI-powered advancements.

    302 00:10:32,760 --> 00:10:33,760 Exactly.

    303 00:10:33,800 --> 00:10:36,860 As people experience the transformative power of these technologies.

    304 00:10:37,240 --> 00:10:37,340 Right.

    305 00:10:37,780 --> 00:10:39,580 Whether it's a life -saving medical treatment.

    306 00:10:39,620 --> 00:10:39,800 Okay.

    307 00:10:40,320 --> 00:10:44,420 Or personalized education that unlocks a child's full potential.

    308 00:10:44,960 --> 00:10:45,280 Mm-hmm.

    309 00:10:45,540 --> 00:10:48,680 They'll demand access for themselves and their families.

    310 00:10:48,681 --> 00:10:51,080 So we're not just talking about making AI affordable.

    311 00:10:51,400 --> 00:10:54,000 We're talking about like, I don't know, a fundamental shift.

    312 00:10:54,440 --> 00:10:54,500 Right.

    313 00:10:54,920 --> 00:10:56,960 In how we value these advancements.

    314 00:10:57,500 --> 00:11:00,040 How we ensure that they're seen as like a public good.

    315 00:11:00,260 --> 00:11:00,480 Mm-hmm.

    316 00:11:00,760 --> 00:11:03,300 Not just another luxury for the privileged few.

    317 00:11:03,660 --> 00:11:03,780 Right.

    318 00:11:04,020 --> 00:11:05,280 And that connects to another concern.

    319 00:11:05,281 --> 00:11:08,440 Amade raises what he calls the opt-out problem.

    320 00:11:08,680 --> 00:11:08,780 Okay.

    321 00:11:09,120 --> 00:11:10,200 It's not just about access.

    322 00:11:10,480 --> 00:11:10,580 Right.

    323 00:11:10,860 --> 00:11:15,500 But about those who might choose to forgo the benefits of AI altogether.

    324 00:11:16,020 --> 00:11:16,120 Okay.

    325 00:11:16,240 --> 00:11:16,660 This is interesting.

    326 00:11:16,661 --> 00:11:19,180 We haven't really dug into this side of things yet.

    327 00:11:19,560 --> 00:11:21,979 Who are these people and why would they opt

    328 00:11:21,980 --> 00:11:24,120 out of something potentially so beneficial?

    329 00:11:24,180 --> 00:11:27,060 It's a diverse group with complex reasons.

    330 00:11:27,340 --> 00:11:27,480 Okay.

    331 00:11:27,860 --> 00:11:29,660 Some might distrust the technology.

    332 00:11:29,661 --> 00:11:30,080 Yeah.

    333 00:11:30,240 --> 00:11:32,720 Due to misinformation or religious beliefs.

    334 00:11:32,721 --> 00:11:33,300 Mm-hmm.

    335 00:11:33,301 --> 00:11:34,760 Others might simply fear change.

    336 00:11:35,240 --> 00:11:35,520 Okay.

    337 00:11:35,640 --> 00:11:36,940 Especially such rapid change.

    338 00:11:37,340 --> 00:11:37,460 Yeah.

    339 00:11:37,680 --> 00:11:38,680 And prefer the familiar.

    340 00:11:38,880 --> 00:11:38,980 Right.

    341 00:11:39,440 --> 00:11:41,940 Even if it means missing out on life-changing advancements.

    342 00:11:42,340 --> 00:11:44,120 It's almost like a self-imposed disadvantage.

    343 00:11:44,180 --> 00:11:44,740 Yeah.

    344 00:11:44,900 --> 00:11:45,400 You could say that.

    345 00:11:45,401 --> 00:11:47,840 And as AI becomes more integrated into society.

    346 00:11:48,320 --> 00:11:48,360 Right.

    347 00:11:48,640 --> 00:11:51,800 Those who opt out could find themselves increasingly left behind, right?

    348 00:11:51,820 --> 00:11:52,220 Exactly.

    349 00:11:52,740 --> 00:11:56,680 It's not hard to imagine a scenario where access to education.

    350 00:11:57,060 --> 00:11:57,160 Right.

    351 00:11:57,640 --> 00:11:57,800 Healthcare.

    352 00:11:58,520 --> 00:11:58,580 Yeah.

    353 00:11:59,160 --> 00:12:00,160 Even job opportunities.

    354 00:12:01,260 --> 00:12:04,100 Hinges on engagement with AI in some way.

    355 00:12:04,220 --> 00:12:05,980 Which creates a whole new set of challenges.

    356 00:12:06,320 --> 00:12:06,560 It does.

    357 00:12:06,840 --> 00:12:09,480 It's not enough to just develop these incredible advancements.

    358 00:12:09,760 --> 00:12:10,060 Yeah, right.

    359 00:12:10,440 --> 00:12:12,620 We also need to address the underlying reasons.

    360 00:12:13,300 --> 00:12:13,420 Right.

    361 00:12:13,421 --> 00:12:15,480 Why some people might reject them.

    362 00:12:15,680 --> 00:12:18,320 And Amadeus sees a critical role for education here.

    363 00:12:18,500 --> 00:12:19,500 Okay.

    364 00:12:19,640 --> 00:12:22,300 Clear, honest communication about AI.

    365 00:12:22,740 --> 00:12:22,860 Right.

    366 00:12:23,140 --> 00:12:25,040 Both its risks and its potential.

    367 00:12:25,260 --> 00:12:25,580 Mmm.

    368 00:12:25,980 --> 00:12:27,220 It's about building trust.

    369 00:12:27,600 --> 00:12:27,720 Yeah.

    370 00:12:28,060 --> 00:12:31,640 And ensuring that everyone feels included in this technological revolution.

    371 00:12:31,980 --> 00:12:32,060 Nope.

    372 00:12:32,100 --> 00:12:34,200 Not just the early adopters or tech elite.

    373 00:12:34,460 --> 00:12:34,820 Of course.

    374 00:12:35,200 --> 00:12:37,960 But even if we address these issues of access and adoption,

    375 00:12:38,200 --> 00:12:41,660 there's another layer of concern that Amade doesn't shy away

    376 00:12:41,661 --> 00:12:45,900 from the impact of this compressed 21st century.

    377 00:12:46,200 --> 00:12:46,240 Yeah.

    378 00:12:46,580 --> 00:12:47,760 On the very meaning of work.

    379 00:12:48,120 --> 00:12:48,740 It's a big one.

    380 00:12:48,900 --> 00:12:49,060 Right.

    381 00:12:49,340 --> 00:12:51,700 Because if AI is automating so many jobs.

    382 00:12:52,060 --> 00:12:52,220 Yeah.

    383 00:12:52,480 --> 00:12:54,120 What happens to our sense of purpose.

    384 00:12:54,200 --> 00:12:54,760 Right.

    385 00:12:54,860 --> 00:12:55,820 Our place in the world.

    386 00:12:55,821 --> 00:12:56,380 Mmm-hmm.

    387 00:12:56,460 --> 00:12:59,560 Are we all destined to become like those pampered pets.

    388 00:12:59,980 --> 00:13:00,100 Right.

    389 00:13:00,600 --> 00:13:02,960 Completely reliant on our AI overlords.

    390 00:13:03,440 --> 00:13:06,680 Well, Amadeus actually pushes back against this idea.

    391 00:13:06,940 --> 00:13:07,100 Okay.

    392 00:13:07,460 --> 00:13:10,840 That our worth is solely defined by our economic output.

    393 00:13:10,841 --> 00:13:11,180 Right.

    394 00:13:11,580 --> 00:13:15,120 By how much we can produce or achieve in a traditional sense.

    395 00:13:15,740 --> 00:13:17,600 He argues that meaning comes from many sources.

    396 00:13:18,080 --> 00:13:18,280 Okay.

    397 00:13:18,580 --> 00:13:19,580 Relationships.

    398 00:13:19,700 --> 00:13:20,700 Creativity.

    399 00:13:20,900 --> 00:13:21,340 Personal passions.

    400 00:13:21,780 --> 00:13:22,020 Mmm-hmm.

    401 00:13:22,040 --> 00:13:25,440 Even just the simple joy of experiencing life itself.

    402 00:13:25,840 --> 00:13:26,020 Yeah.

    403 00:13:26,360 --> 00:13:29,360 And in a world where AI takes care of many of the necessities.

    404 00:13:29,920 --> 00:13:29,960 Right.

    405 00:13:30,200 --> 00:13:31,840 We might have even more time and freedom.

    406 00:13:32,220 --> 00:13:32,360 Yeah.

    407 00:13:32,780 --> 00:13:34,580 To explore those other avenues of fulfillment.

    408 00:13:34,980 --> 00:13:37,200 It's like shifting from a culture of productivity.

    409 00:13:37,660 --> 00:13:37,720 Yes.

    410 00:13:38,060 --> 00:13:39,320 To a culture of possibility.

    411 00:13:39,740 --> 00:13:40,100 I like that.

    412 00:13:40,101 --> 00:13:43,500 Where we're free to pursue what truly interests us.

    413 00:13:43,700 --> 00:13:43,760 Right.

    414 00:13:43,920 --> 00:13:44,920 What brings us joy.

    415 00:13:45,100 --> 00:13:45,280 Exactly.

    416 00:13:45,580 --> 00:13:47,260 Without the pressure of economic necessity.

    417 00:13:47,700 --> 00:13:48,020 Mmm-hmm.

    418 00:13:48,240 --> 00:13:49,560 But that raises another question.

    419 00:13:50,980 --> 00:13:51,120 Okay.

    420 00:13:51,700 --> 00:13:53,380 If AI is doing most of the work.

    421 00:13:53,780 --> 00:13:53,880 Right.

    422 00:13:54,120 --> 00:13:57,020 How do we ensure everyone benefits from this abundance.

    423 00:13:57,560 --> 00:13:57,680 Right.

    424 00:13:58,140 --> 00:14:00,520 Rather than just to select few who own the AI.

    425 00:14:01,000 --> 00:14:04,160 So Amadei acknowledges that the traditional economic models.

    426 00:14:04,480 --> 00:14:04,640 Okay.

    427 00:14:05,000 --> 00:14:07,160 Based on scarcity and competition.

    428 00:14:07,700 --> 00:14:07,780 Right.

    429 00:14:07,781 --> 00:14:12,800 Might need a serious rethink in this compressed 21st century.

    430 00:14:12,980 --> 00:14:13,240 Okay.

    431 00:14:13,640 --> 00:14:16,700 He talks about the need for a more equitable distribution of resources.

    432 00:14:17,040 --> 00:14:17,180 Okay.

    433 00:14:17,420 --> 00:14:19,580 Perhaps through something like a global wealth tax.

    434 00:14:19,920 --> 00:14:20,260 Mmm-hmm.

    435 00:14:20,261 --> 00:14:25,781 Or a system where AI itself helps to manage and distribute the abundance it creates.

    436 00:14:26,200 --> 00:14:28,540 So it's not just about humans redistributing wealth.

    437 00:14:29,100 --> 00:14:29,160 Right.

    438 00:14:29,260 --> 00:14:31,900 It's about designing AI systems that are inherently more equitable.

    439 00:14:32,640 --> 00:14:34,980 That prioritize the well-being of all people.

    440 00:14:35,260 --> 00:14:36,300 Not just profit margins.

    441 00:14:36,301 --> 00:14:36,840 Right.

    442 00:14:37,160 --> 00:14:40,360 It's a huge challenge but also an incredible opportunity.

    443 00:14:40,860 --> 00:14:41,460 It really is.

    444 00:14:41,520 --> 00:14:43,400 To rethink our relationship with work.

    445 00:14:43,780 --> 00:14:43,980 Yeah.

    446 00:14:44,100 --> 00:14:44,400 With wealth.

    447 00:14:45,080 --> 00:14:45,540 With each other.

    448 00:14:45,720 --> 00:14:46,100 Exactly.

    449 00:14:46,420 --> 00:14:52,500 And that's what makes Amadei's vision of a compressed 21st century so compelling.

    450 00:14:53,080 --> 00:14:53,180 Yeah.

    451 00:14:53,420 --> 00:14:54,800 It's not some utopian fantasy.

    452 00:14:55,360 --> 00:14:55,540 Right.

    453 00:14:55,720 --> 00:14:58,340 Where technology magically solves all our problems.

    454 00:14:58,520 --> 00:14:58,740 Of course.

    455 00:14:59,080 --> 00:15:00,080 It's a call to action.

    456 00:15:00,280 --> 00:15:00,440 Okay.

    457 00:15:00,680 --> 00:15:04,640 A challenge to step up and actively shape this future.

    458 00:15:04,920 --> 00:15:05,160 Right.

    459 00:15:05,161 --> 00:15:10,420 To ensure that it reflects our values, our aspirations for a better world.

    460 00:15:10,640 --> 00:15:12,921 It's a future that's both exhilarating and a bit daunting.

    461 00:15:13,380 --> 00:15:13,420 Yeah.

    462 00:15:13,520 --> 00:15:15,600 Full of possibilities and potential pitfalls.

    463 00:15:16,220 --> 00:15:16,440 For sure.

    464 00:15:16,640 --> 00:15:19,520 But as Amadei reminds us, it's not predetermined.

    465 00:15:20,320 --> 00:15:20,540 It's not.

    466 00:15:20,700 --> 00:15:23,154 It's a future that we all of us have a hand

    467 00:15:23,155 --> 00:15:25,500 in creating through the choices we make today.

    468 00:15:25,880 --> 00:15:26,200 It's true.

    469 00:15:26,600 --> 00:15:27,820 So what role will you play?

    470 00:15:28,320 --> 00:15:28,800 Yeah, that's the question.

    471 00:15:29,280 --> 00:15:31,880 That's something to ponder as we wrap up this deep dive.

    472 00:15:32,240 --> 00:15:32,440 Yes.

    473 00:15:32,441 --> 00:15:35,080 Into Dario Amadei's bold vision of the future.

    474 00:15:35,440 --> 00:15:36,020 It's a good good one.

    475 00:15:36,260 --> 00:15:37,960 We'll leave you with that thought and a gentle

    476 00:15:37,961 --> 00:15:40,500 reminder to subscribe for more mind-expanding

    477 00:15:40,501 --> 00:15:43,860 explorations at the cutting edge of AI and beyond.

    478 00:15:44,560 --> 00:15:45,560 Until next time.

    479 00:15:45,620 --> 00:15:48,020 Until next time, keep asking those big questions.

    480 00:15:48,520 --> 00:15:48,820 Keep learning.

    481 00:15:49,120 --> 00:15:53,121 Keep pushing the boundaries of what's possible and most importantly, keep learning.

  2. Assistant

    Here's a spellchecked version of the subtitle with timecodes unchanged:

    1
    00:00:00,000 --> 00:00:02,320
    What if I told you we
    could live through the

    2
    00:00:02,321 --> 00:00:05,740
    next century of progress in
    just the next decade?

    3
    00:00:06,180 --> 00:00:09,640
    That's the idea behind what
    AI researcher Dario

    4
    00:00:09,641 --> 00:00:12,920
    Amodei calls the
    compressed 21st century.

    5
    00:00:13,080 --> 00:00:15,700
    And here's why this particular
    prediction is so intriguing.

    6
    00:00:16,300 --> 00:00:16,340
    Right.

    7
    00:00:16,600 --> 00:00:20,220
    Amodei isn't your typical
    Silicon Valley tech optimist.

    8
    00:00:20,300 --> 00:00:20,520
    Right.

    9
    00:00:20,740 --> 00:00:26,060
    He's both a biologist and an
    AI expert who runs Anthropic.

    10
    00:00:26,860 --> 00:00:26,900
    Yeah.

    11
    00:00:26,901 --> 00:00:30,860
    A company known for its focus on AI risks.

    12
    00:00:31,380 --> 00:00:31,400
    Right.

    13
    00:00:31,620 --> 00:00:32,640
    Not utopian promises.

    14
    00:00:32,900 --> 00:00:33,320
    Exactly.

    15
    00:00:33,880 --> 00:00:35,500
    He's not known for sugar coating things.

    16
    00:00:35,720 --> 00:00:35,780
    No.

    17
    00:00:36,020 --> 00:00:39,400
    So when Amodei says powerful AI
    could deliver the next 50

    18
    00:00:39,401 --> 00:00:42,420
    to 100 years of breakthroughs in
    just 5 to 10,

    19
    00:00:43,040 --> 00:00:43,660
    it's not just hype.

    20
    00:00:44,060 --> 00:00:44,080
    Right.

    21
    00:00:44,180 --> 00:00:47,200
    It's him saying, hold on,
    things are about to get really wild.

    22
    00:00:47,460 --> 00:00:51,820
    And powerful AI in Amodei's
    world isn't some vague concept.

    23
    00:00:52,280 --> 00:00:52,300
    Okay.

    24
    00:00:52,301 --> 00:00:56,760
    He's talking about AI with genius
    level intellect across multiple fields.

    25
    00:00:57,000 --> 00:00:57,300
    Oh, wow.

    26
    00:00:57,360 --> 00:00:58,360
    Not just one party trick.

    27
    00:00:59,020 --> 00:01:01,120
    Think about a team of
    the world's brightest minds.

    28
    00:01:01,460 --> 00:01:01,640
    Okay.

    29
    00:01:02,200 --> 00:01:06,820
    All working together in a data center,
    designing experiments, inventing tools,

    30
    00:01:07,940 --> 00:01:08,940
    essentially operating.

    31
    00:01:09,380 --> 00:01:13,800
    Like... It's like having a country of
    geniuses in a data center as he puts it.

    32
    00:01:13,980 --> 00:01:14,060
    Yeah.

    33
    00:01:14,440 --> 00:01:16,080
    Which honestly...
    Good way to put it.

    34
    00:01:16,360 --> 00:01:17,920
    Doesn't sound that far-fetched anymore.

    35
    00:01:18,080 --> 00:01:18,960
    It really doesn't.

    36
    00:01:19,020 --> 00:01:22,080
    The big question is,
    how does this actually play out?

    37
    00:01:22,081 --> 00:01:26,140
    In the real world,
    where do we see the most radical upside?

    38
    00:01:27,180 --> 00:01:30,740
    Well, Amodei points to
    biology and health as

    39
    00:01:30,741 --> 00:01:33,520
    being ripe for this kind
    of AI acceleration.

    40
    00:01:33,960 --> 00:01:34,960
    Interesting.

    41
    00:01:35,060 --> 00:01:37,320
    And this is where his dual
    background as both a

    42
    00:01:37,321 --> 00:01:41,020
    biologist and an AI expert
    is so insightful.

    43
    00:01:41,480 --> 00:01:41,580
    Yeah.

    44
    00:01:41,860 --> 00:01:44,680
    He understands the pace
    of scientific discovery.

    45
    00:01:45,120 --> 00:01:45,180
    Okay.

    46
    00:01:45,480 --> 00:01:49,320
    How it often hinges on
    these rare groundbreaking moments.

    47
    00:01:49,321 --> 00:01:49,540
    Uh-huh.

    48
    00:01:49,640 --> 00:01:50,680
    The aha moment.

    49
    00:01:50,880 --> 00:01:52,116
    Think about CRISPR, for example.

    50
    00:01:52,140 --> 00:01:54,620
    Right. That completely revolutionized
    gene editing.

    51
    00:01:55,320 --> 00:01:57,680
    And it's not hard to imagine
    how AI could help

    52
    00:01:57,681 --> 00:02:00,280
    us uncover those kinds of
    breakthroughs even faster.

    53
    00:02:00,640 --> 00:02:01,120
    Exactly.

    54
    00:02:01,540 --> 00:02:05,720
    AI can analyze massive data sets,
    identify hidden patterns,

    55
    00:02:06,160 --> 00:02:08,360
    even design and run experiments.

    56
    00:02:08,720 --> 00:02:08,860
    Wow.

    57
    00:02:09,400 --> 00:02:10,720
    At a pace we could only dream of.

    58
    00:02:10,780 --> 00:02:11,780
    Yeah.

    59
    00:02:11,860 --> 00:02:14,460
    Amodei's argument is that
    AI isn't magic.

    60
    00:02:15,260 --> 00:02:17,380
    But it can drastically
    compress the timeline.

    61
    00:02:17,840 --> 00:02:18,060
    Okay.

    62
    00:02:18,440 --> 00:02:19,900
    From Eureka moment.

    63
    00:02:20,220 --> 00:02:20,360
    Yeah.

    64
    00:02:20,820 --> 00:02:23,140
    To real-world impact.

    65
    00:02:23,780 --> 00:02:27,380
    Okay, so it's not like AI is
    going to magically cure cancer overnight.

    66
    00:02:27,880 --> 00:02:29,136
    There are still limitations, right?

    67
    00:02:29,160 --> 00:02:29,360
    Of course.

    68
    00:02:29,460 --> 00:02:30,460
    Experiments take time.

    69
    00:02:30,780 --> 00:02:30,940
    Yes.

    70
    00:02:31,120 --> 00:02:33,900
    And we're still bound by
    the laws of physics.

    71
    00:02:34,260 --> 00:02:36,900
    Amodei isn't saying we can
    defy the laws of nature.

    72
    00:02:37,280 --> 00:02:37,500
    Okay.

    73
    00:02:37,501 --> 00:02:43,401
    But what he's saying is that we can
    drastically accelerate the pace of discovery.

    74
    00:02:43,820 --> 00:02:46,547
    Imagine a world where
    instead of taking 50

    75
    00:02:46,548 --> 00:02:49,560
    years to achieve a
    breakthrough takes five.

    76
    00:02:49,900 --> 00:02:50,000
    Wow.

    77
    00:02:50,540 --> 00:02:52,500
    That's the kind of compression
    he's talking about.

    78
    00:02:52,540 --> 00:02:55,060
    So what does that actually look like
    for someone like you or me?

    79
    00:02:55,500 --> 00:02:55,640
    Right.

    80
    00:02:55,980 --> 00:02:59,600
    What changes in a world where the pace
    of progress is ten times faster?

    81
    00:03:00,280 --> 00:03:01,060
    Think about it this way.

    82
    00:03:01,240 --> 00:03:01,320
    Okay.

    83
    00:03:01,500 --> 00:03:04,720
    In this compressed 21st century
    diseases like cancer,

    84
    00:03:05,060 --> 00:03:07,500
    Alzheimer's, even
    inherited genetic conditions.

    85
    00:03:07,780 --> 00:03:08,320
    Oh, wow.

    86
    00:03:08,620 --> 00:03:10,300
    They could largely be problems
    of the past.

    87
    00:03:10,500 --> 00:03:10,700
    Okay.

    88
    00:03:10,840 --> 00:03:12,440
    Not just within our children's lifetime.

    89
    00:03:12,560 --> 00:03:12,680
    Yeah.

    90
    00:03:12,860 --> 00:03:14,040
    But potentially within ours.

    91
    00:03:14,260 --> 00:03:14,380
    Wow.

    92
    00:03:15,100 --> 00:03:16,500
    That's not just a medical marvel.

    93
    00:03:16,820 --> 00:03:16,900
    Right.

    94
    00:03:17,060 --> 00:03:19,200
    That's a complete societal game changer.

    95
    00:03:19,300 --> 00:03:19,880
    It really is.

    96
    00:03:20,120 --> 00:03:20,680
    And get this.

    97
    00:03:20,980 --> 00:03:21,100
    Yeah.

    98
    00:03:21,520 --> 00:03:25,240
    Amodei even suggests that we could
    double the human lifespan again.

    99
    00:03:25,640 --> 00:03:26,020
    It's true.

    100
    00:03:26,060 --> 00:03:27,700
    Based on the progress we've already made.

    101
    00:03:27,860 --> 00:03:27,980
    Yeah.

    102
    00:03:28,160 --> 00:03:31,440
    It's a mind-boggling concept
    with huge implications.

    103
    00:03:31,820 --> 00:03:31,980
    Right.

    104
    00:03:31,981 --> 00:03:37,080
    Imagine the impact on everything from
    family structures to retirement plans.

    105
    00:03:37,081 --> 00:03:37,500
    Wow.

    106
    00:03:38,100 --> 00:03:41,100
    Not to mention the ethical and
    philosophical questions it raises.

    107
    00:03:41,600 --> 00:03:41,780
    Right.

    108
    00:03:42,040 --> 00:03:44,740
    And it's not just about living
    longer. It's about living better.

    109
    00:03:45,140 --> 00:03:45,520
    Exactly.

    110
    00:03:45,840 --> 00:03:48,220
    Amodei talks about this
    incredible potential for

    111
    00:03:48,221 --> 00:03:54,020
    AI to revolutionize not just our physical
    health, but also our mental well-being.

    112
    00:03:54,520 --> 00:03:54,540
    Exactly.

    113
    00:03:55,180 --> 00:03:59,120
    He sees the compressed 21st happening
    for the mind too.

    114
    00:03:59,420 --> 00:03:59,680
    Interesting.

    115
    00:03:59,940 --> 00:04:00,520
    Not just the body.

    116
    00:04:00,521 --> 00:04:00,960
    Okay.

    117
    00:04:01,360 --> 00:04:03,960
    And this is where recent
    advancements in something

    118
    00:04:03,961 --> 00:04:06,680
    called AI interpretability
    come into play.

    119
    00:04:07,160 --> 00:04:07,400
    Okay.

    120
    00:04:07,540 --> 00:04:09,460
    I have to admit,
    I'm not familiar with that term.

    121
    00:04:09,500 --> 00:04:09,900
    Okay, sure.

    122
    00:04:10,500 --> 00:04:15,540
    What exactly is AI interpretability,
    and how does it relate to our minds?

    123
    00:04:16,060 --> 00:04:22,320
    So, in simple terms, AI interpretability
    is about understanding how AI thinks.

    124
    00:04:22,700 --> 00:04:22,900
    Okay.

    125
    00:04:23,180 --> 00:04:24,220
    Not just what it produces.

    126
    00:04:24,680 --> 00:04:24,800
    Okay.

    127
    00:04:25,100 --> 00:04:25,760
    Think of it like this.

    128
    00:04:25,761 --> 00:04:26,060
    Okay.

    129
    00:04:26,280 --> 00:04:33,280
    Imagine having a window into the decision-making process of a brilliant scientist.

    130
    00:04:33,780 --> 00:04:33,960
    Okay.

    131
    00:04:34,480 --> 00:04:36,080
    Now, apply that to AI.

    132
    00:04:36,920 --> 00:04:37,280
    Interesting.

    133
    00:04:37,740 --> 00:04:41,180
    By understanding how AI arrives
    at its conclusions,

    134
    00:04:41,700 --> 00:04:47,620
    we can potentially unlock secrets
    of the human brain far more quickly.

    135
    00:04:48,020 --> 00:04:50,720
    So, it's almost like AI is holding up
    a mirror to our own minds.

    136
    00:04:51,160 --> 00:04:52,196
    That's a great way to put it.

    137
    00:04:52,220 --> 00:04:54,340
    Helping us understand ourselves better
    in the process.

    138
    00:04:55,120 --> 00:04:55,640
    Precisely.

    139
    00:04:55,900 --> 00:05:00,280
    And this deeper understanding could have
    profound implications for mental health.

    140
    00:05:00,920 --> 00:05:03,620
    Amodei is very clear that
    he sees mental well

    141
    00:05:03,621 --> 00:05:07,440
    -being as being just as crucial
    as physical health, if not more so.

    142
    00:05:07,640 --> 00:05:09,320
    I think a lot of people
    can relate to that.

    143
    00:05:09,380 --> 00:05:11,216
    We're talking about tackling
    things like depression,

    144
    00:05:11,240 --> 00:05:14,700
    anxiety, addiction,
    even personality disorders.

    145
    00:05:15,040 --> 00:05:15,100
    Yeah.

    146
    00:05:15,360 --> 00:05:17,780
    With the same urgency and
    resources that we

    147
    00:05:17,781 --> 00:05:20,360
    currently direct towards cancer
    or heart disease.

    148
    00:05:20,520 --> 00:05:20,920
    Exactly.

    149
    00:05:20,921 --> 00:05:21,240
    Yeah.

    150
    00:05:21,440 --> 00:05:25,640
    Amodei's vision is to address these issues,
    not just by treating symptoms.

    151
    00:05:26,000 --> 00:05:26,080
    Right.

    152
    00:05:26,340 --> 00:05:29,900
    But by understanding and potentially
    addressing the root causes.

    153
    00:05:30,300 --> 00:05:30,360
    Okay.

    154
    00:05:30,520 --> 00:05:34,060
    And he sees AI as a
    key driver of that revolution.

    155
    00:05:34,560 --> 00:05:37,160
    So, how does he see that unfolding?

    156
    00:05:37,820 --> 00:05:41,080
    How does AI actually help us
    achieve this mental health revolution?

    157
    00:05:41,520 --> 00:05:44,840
    Amodei lays out like four main avenues.

    158
    00:05:45,240 --> 00:05:45,380
    Okay.

    159
    00:05:45,660 --> 00:05:47,380
    Where AI can make a huge difference.

    160
    00:05:47,760 --> 00:05:47,900
    Okay.

    161
    00:05:47,901 --> 00:05:51,600
    First, imagine AI accelerating the
    development of new medications.

    162
    00:05:52,280 --> 00:05:53,120
    And genetic therapies.

    163
    00:05:53,121 --> 00:05:53,940
    Mm-hmm.

    164
    00:05:53,941 --> 00:05:56,520
    Fine-tuned to an
    individual's brain chemistry.

    165
    00:05:57,240 --> 00:05:59,260
    It's like personalized medicine,
    but for the mind.

    166
    00:05:59,900 --> 00:06:00,020
    Exactly.

    167
    00:06:00,220 --> 00:06:01,980
    No more trial and
    error with antidepressants.

    168
    00:06:02,220 --> 00:06:04,380
    It's about getting the right treatment
    from the get-go.

    169
    00:06:04,520 --> 00:06:05,520
    Exactly.

    170
    00:06:05,600 --> 00:06:10,980
    Second, Amodei talks about AI enabling
    incredibly precise interventions in the brain.

    171
    00:06:11,160 --> 00:06:11,340
    Mm-hmm.

    172
    00:06:11,480 --> 00:06:13,340
    And we're not talking about
    crude surgery here.

    173
    00:06:13,560 --> 00:06:13,600
    Right.

    174
    00:06:13,601 --> 00:06:17,460
    Think about using AI to
    target specific neural pathways.

    175
    00:06:17,880 --> 00:06:18,000
    Okay.

    176
    00:06:18,360 --> 00:06:20,680
    Potentially addressing conditions
    like Parkinson's.

    177
    00:06:21,120 --> 00:06:21,240
    Yeah.

    178
    00:06:21,260 --> 00:06:22,800
    Or even severe OCD.

    179
    00:06:22,960 --> 00:06:23,600
    Wow.

    180
    00:06:23,980 --> 00:06:24,480
    At their root.

    181
    00:06:24,780 --> 00:06:27,960
    It's like having a microscopic repair
    crew for the brain.

    182
    00:06:28,400 --> 00:06:28,680
    Yes.

    183
    00:06:28,920 --> 00:06:32,780
    Guided by AI to fix
    things with incredible accuracy.

    184
    00:06:33,260 --> 00:06:34,316
    That's a great way to put it.

    185
    00:06:34,340 --> 00:06:37,020
    What's fascinating is how
    Amodei connects this.

    186
    00:06:37,300 --> 00:06:37,440
    Yeah.

    187
    00:06:37,620 --> 00:06:42,340
    Not just to fixing what's broken out to
    potentially enhancing our mental experience.

    188
    00:06:42,341 --> 00:06:42,880
    That's right.

    189
    00:06:43,020 --> 00:06:44,420
    He doesn't shy away from that idea.

    190
    00:06:44,540 --> 00:06:44,640
    Okay.

    191
    00:06:44,980 --> 00:06:50,040
    His third avenue is all about AI helping
    us build incredibly accurate

    192
    00:06:50,041 --> 00:06:53,480
    computational models
    of how our minds work.

    193
    00:06:53,920 --> 00:06:58,340
    Imagine using these models to not
    only understand conditions like autism.

    194
    00:06:58,820 --> 00:06:59,040
    Okay.

    195
    00:06:59,120 --> 00:07:00,120
    Or schizophrenia better.

    196
    00:07:00,600 --> 00:07:00,680
    Right.

    197
    00:07:00,960 --> 00:07:06,040
    But to develop personalized learning plans
    that optimize how we learn and create.

    198
    00:07:06,620 --> 00:07:08,100
    So instead of just treating deficits.

    199
    00:07:08,680 --> 00:07:08,820
    Exactly.

    200
    00:07:08,821 --> 00:07:13,040
    We're talking about using AI to
    unlock our full cognitive potential.

    201
    00:07:13,360 --> 00:07:14,416
    I like that way I put it again.

    202
    00:07:14,440 --> 00:07:15,600
    That's incredibly exciting.

    203
    00:07:15,840 --> 00:07:16,920
    But also a little daunting.

    204
    00:07:17,300 --> 00:07:17,360
    Yeah.

    205
    00:07:17,600 --> 00:07:19,480
    What about the ethical implications
    of all this?

    206
    00:07:19,620 --> 00:07:21,800
    So that's something Amodei
    addresses head-on.

    207
    00:07:22,040 --> 00:07:22,240
    Okay.

    208
    00:07:22,540 --> 00:07:25,381
    He doesn't pretend these advancements
    won't come with tough questions.

    209
    00:07:25,580 --> 00:07:25,720
    Right.

    210
    00:07:26,160 --> 00:07:28,992
    But he argues that the
    potential benefits, particularly

    211
    00:07:28,993 --> 00:07:33,600
    in alleviating suffering,
    are too great to ignore.

    212
    00:07:33,960 --> 00:07:34,040
    Right.

    213
    00:07:34,041 --> 00:07:36,840
    And his fourth avenue is
    perhaps the most accessible.

    214
    00:07:37,300 --> 00:07:37,480
    Okay.

    215
    00:07:37,980 --> 00:07:40,270
    AI-powered therapy available to everyone.

    216
    00:07:40,920 --> 00:07:41,140
    Okay.

    217
    00:07:41,360 --> 00:07:43,320
    Not just those who can
    afford traditional therapy.

    218
    00:07:43,480 --> 00:07:44,480
    That's huge.

    219
    00:07:45,540 --> 00:07:48,533
    Imagine a world where
    everyone has access

    220
    00:07:48,534 --> 00:07:52,080
    to affordable, effective
    mental health care.

    221
    00:07:52,380 --> 00:07:53,440
    It's a powerful idea.

    222
    00:07:53,780 --> 00:07:55,060
    Tailored to their unique needs.

    223
    00:07:55,520 --> 00:07:55,700
    Mm-hmm.

    224
    00:07:55,920 --> 00:07:59,440
    That could be a true game
    changer for global well-being.

    225
    00:08:00,040 --> 00:08:03,200
    And that brings us to another
    crucial aspect of Amodei's vision.

    226
    00:08:03,201 --> 00:08:03,700
    Okay.

    227
    00:08:04,060 --> 00:08:08,760
    The impact of this compressed 21st
    century on global inequality.

    228
    00:08:09,380 --> 00:08:09,580
    Okay.

    229
    00:08:09,800 --> 00:08:14,420
    He's very clear that while AI has the
    potential to lift billions out of poverty.

    230
    00:08:14,840 --> 00:08:14,920
    Right.

    231
    00:08:15,200 --> 00:08:16,700
    It requires intentional effort.

    232
    00:08:17,140 --> 00:08:17,340
    Right.

    233
    00:08:17,540 --> 00:08:19,020
    Because technology itself is neutral.

    234
    00:08:19,200 --> 00:08:19,360
    Exactly.

    235
    00:08:19,700 --> 00:08:21,920
    It's how we choose to develop
    it and deploy it.

    236
    00:08:22,040 --> 00:08:22,280
    That's right.

    237
    00:08:22,520 --> 00:08:25,360
    That will determine whether it benefits
    everyone or just widens

    238
    00:08:25,361 --> 00:08:27,260
    the gap between the haves
    and the have-nots.

    239
    00:08:27,340 --> 00:08:27,860
    Precisely.

    240
    00:08:27,960 --> 00:08:28,100
    Okay.

    241
    00:08:28,580 --> 00:08:30,600
    And Amodei sees this playing
    out on several levels.

    242
    00:08:30,601 --> 00:08:30,860
    Yep.

    243
    00:08:31,140 --> 00:08:33,340
    First, there's the issue
    of distributing those

    244
    00:08:33,341 --> 00:08:35,200
    incredible health advances we
    talked about.

    245
    00:08:35,240 --> 00:08:35,260
    Okay.

    246
    00:08:35,720 --> 00:08:37,360
    Think about the eradication of smallpox.

    247
    00:08:37,640 --> 00:08:38,300
    Mm-hmm.

    248
    00:08:38,301 --> 00:08:40,740
    A disease that once
    ravaged entire populations.

    249
    00:08:41,240 --> 00:08:41,440
    Yeah.

    250
    00:08:42,020 --> 00:08:44,240
    With AI-powered strategies and logistics.

    251
    00:08:44,740 --> 00:08:44,820
    Right.

    252
    00:08:45,220 --> 00:08:48,920
    We could do the same for
    diseases like malaria, tuberculosis.

    253
    00:08:49,420 --> 00:08:51,781
    And even common
    parasitic infections that

    254
    00:08:51,782 --> 00:08:54,040
    disproportionately affect
    developing countries.

    255
    00:08:54,300 --> 00:08:57,660
    It's like a global health initiative,
    but supercharged by AI.

    256
    00:08:58,040 --> 00:08:58,260
    Exactly.

    257
    00:08:58,261 --> 00:09:01,400
    And with AI-optimizing
    distribution networks.

    258
    00:09:02,120 --> 00:09:05,480
    Even the most remote communities could
    benefit from these breakthroughs.

    259
    00:09:05,600 --> 00:09:06,600
    Absolutely.

    260
    00:09:06,920 --> 00:09:11,400
    Amodei also talks about AI's potential to
    ignite a second green revolution.

    261
    00:09:11,820 --> 00:09:12,820
    Okay.

    262
    00:09:12,960 --> 00:09:15,000
    Imagine AI-optimizing crop yields.

    263
    00:09:15,400 --> 00:09:15,780
    Mm-hmm.

    264
    00:09:15,920 --> 00:09:17,560
    Developing drought-resistant seeds.

    265
    00:09:17,920 --> 00:09:18,080
    Right.

    266
    00:09:18,420 --> 00:09:20,740
    And revolutionizing food production
    and distribution.

    267
    00:09:21,240 --> 00:09:21,320
    Wow.

    268
    00:09:21,740 --> 00:09:24,580
    We're talking about tackling hunger
    on a global scale.

    269
    00:09:25,120 --> 00:09:25,200
    Sure.

    270
    00:09:25,480 --> 00:09:27,500
    Not just increasing profits for
    a select few.

    271
    00:09:27,501 --> 00:09:29,160
    That's a powerful vision.

    272
    00:09:30,580 --> 00:09:32,360
    But let's be realistic for a moment.

    273
    00:09:32,540 --> 00:09:32,560
    Okay.

    274
    00:09:32,740 --> 00:09:36,261
    We've seen how technology can
    sometimes exacerbate existing inequalities.

    275
    00:09:36,340 --> 00:09:36,440
    Right.

    276
    00:09:36,800 --> 00:09:37,760
    Even with the best intentions.

    277
    00:09:37,800 --> 00:09:38,800
    Of course.

    278
    00:09:39,220 --> 00:09:43,080
    What about access to these
    advancements within developed countries?

    279
    00:09:43,640 --> 00:09:43,680
    Right.

    280
    00:09:43,880 --> 00:09:48,420
    How do we ensure that everyone benefits
    not just a wealthy elite?

    281
    00:09:48,920 --> 00:09:51,480
    Amodei acknowledges that this is
    a valid concern.

    282
    00:09:51,820 --> 00:09:52,020
    Okay.

    283
    00:09:52,021 --> 00:09:57,620
    He's not naive about the potential for AI
    to be used for profit over people.

    284
    00:09:58,300 --> 00:10:01,280
    But he argues that at
    least in democratic societies.

    285
    00:10:02,040 --> 00:10:05,014
    Market forces combined
    with citizen demand will

    286
    00:10:05,015 --> 00:10:08,220
    drive down costs and
    increase access over time.

    287
    00:10:08,380 --> 00:10:08,600
    Okay.

    288
    00:10:09,140 --> 00:10:10,840
    Think about how quickly cell
    phones went from

    289
    00:10:10,841 --> 00:10:12,920
    luxury items to then
    practically ubiquitous.

    290
    00:10:13,420 --> 00:10:15,680
    That's a good point.
    It seemed like almost overnight

    291
    00:10:15,681 --> 00:10:17,880
    everyone and their grandmother
    had a smartphone.

    292
    00:10:18,560 --> 00:10:18,740
    Yeah.

    293
    00:10:18,880 --> 00:10:21,100
    And it wasn't just about
    the technology becoming cheaper.

    294
    00:10:21,101 --> 00:10:21,600
    Right.

    295
    00:10:21,760 --> 00:10:24,360
    It was also about people
    recognizing the value.

    296
    00:10:24,640 --> 00:10:25,000
    Exactly.

    297
    00:10:25,220 --> 00:10:26,400
    Demanding to be connected.

    298
    00:10:26,840 --> 00:10:26,980
    Yes.

    299
    00:10:27,160 --> 00:10:29,300
    So you're saying a similar
    dynamic could play out?

    300
    00:10:29,600 --> 00:10:29,920
    Yes, right.

    301
    00:10:30,100 --> 00:10:32,180
    With these AI-powered advancements.

    302
    00:10:32,760 --> 00:10:33,760
    Exactly.

    303
    00:10:33,800 --> 00:10:36,860
    As people experience the transformative
    power of these technologies.

    304
    00:10:37,240 --> 00:10:37,340
    Right.

    305
    00:10:37,780 --> 00:10:39,580
    Whether it's a life-saving medical treatment.

    306
    00:10:39,620 --> 00:10:39,800
    Okay.

    307
    00:10:40,320 --> 00:10:44,420
    Or personalized education that unlocks
    a child's full potential.

    308
    00:10:44,960 --> 00:10:45,280
    Mm-hmm.

    309
    00:10:45,540 --> 00:10:48,680
    They'll demand access for themselves
    and their families.

    310
    00:10:48,681 --> 00:10:51,080
    So we're not just talking
    about making AI affordable.

    311
    00:10:51,400 --> 00:10:54,000
    We're talking about like, I don't know,
    a fundamental shift.

    312
    00:10:54,440 --> 00:10:54,500
    Right.

    313
    00:10:54,920 --> 00:10:56,960
    In how we value these advancements.

    314
    00:10:57,500 --> 00:10:57,560
    Right.

    315
    00:10:58,140 --> 00:11:00,520
    Rather than just a select few
    who own the AI.

    316
    00:11:01,000 --> 00:11:04,160
    So Amodei acknowledges that the
    traditional economic models.

    317
    00:11:04,480 --> 00:11:04,640
    Okay.

    318
    00:11:05,000 --> 00:11:07,160
    Based on scarcity and competition.

    319
    00:11:07,700 --> 00:11:07,780
    Right.

    320
    00:11:07,781 --> 00:11:12,800
    Might need a serious rethink in
    this compressed 21st century.

    321
    00:11:12,980 --> 00:11:13,240
    Okay.

    322
    00:11:13,640 --> 00:11:16,700
    He talks about the need for a
    more equitable distribution of resources.

    323
    00:11:17,040 --> 00:11:17,180
    Okay.

    324
    00:11:17,420 --> 00:11:19,580
    Perhaps through something like a
    global wealth tax.

    325
    00:11:19,920 --> 00:11:20,260
    Mmm-hmm.

    326
    00:11:20,261 --> 00:11:25,781
    Or a system where AI itself helps to manage
    and distribute the abundance it creates.

    327
    00:11:26,200 --> 00:11:28,540
    So it's not just about
    humans redistributing wealth.

    328
    00:11:29,100 --> 00:11:29,160
    Right.

    329
    00:11:29,260 --> 00:11:31,900
    It's about designing AI systems that
    are inherently more equitable.

    330
    00:11:32,640 --> 00:11:34,980
    That prioritize the well-being
    of all people.

    331
    00:11:35,260 --> 00:11:36,300
    Not just profit margins.

    332
    00:11:36,301 --> 00:11:36,840
    Right.

    333
    00:11:37,160 --> 00:11:40,360
    It's a huge challenge but
    also an incredible opportunity.

    334
    00:11:40,860 --> 00:11:41,460
    It really is.

    335
    00:11:41,520 --> 00:11:43,400
    To rethink our relationship with work.

    336
    00:11:43,780 --> 00:11:43,980
    Yeah.

    337
    00:11:44,100 --> 00:11:44,400
    With wealth.

    338
    00:11:45,080 --> 00:11:45,540
    With each other.

    339
    00:11:45,720 --> 00:11:46,100
    Exactly.

    340
    00:11:46,420 --> 00:11:52,500
    And that's what makes Amodei's vision of
    a compressed 21st century so compelling.

    341
    00:11:53,080 --> 00:11:53,180
    Yeah.

    342
    00:11:53,420 --> 00:11:54,800
    It's not some utopian fantasy.

    343
    00:11:55,360 --> 00:11:55,540
    Right.

    344
    00:11:55,720 --> 00:11:58,340
    Where technology magically solves
    all our problems.

    345
    00:11:58,520 --> 00:11:58,740
    Of course.

    346
    00:11:59,080 --> 00:12:00,080
    It's a call to action.

    347
    00:12:00,280 --> 00:12:00,440
    Okay.

    348
    00:12:00,680 --> 00:12:04,640
    A challenge to step up and
    actively shape this future.

    349
    00:12:04,920 --> 00:12:05,160
    Right.

    350
    00:12:05,161 --> 00:12:10,420
    To ensure that it reflects our values,
    our aspirations for a better world.

    351
    00:12:10,640 --> 00:12:12,921
    It's a future that's both exhilarating
    and a bit daunting.

    352
    00:12:13,380 --> 00:12:13,420
    Yeah.

    353
    00:12:13,520 --> 00:12:15,600
    Full of possibilities and
    potential pitfalls.

    354
    00:12:16,220 --> 00:12:16,440
    For sure.

    355
    00:12:16,640 --> 00:12:19,520
    But as Amodei reminds us,
    it's not predetermined.

    356
    00:12:20,320 --> 00:12:20,540
    It's not.

    357
    00:12:20,700 --> 00:12:23,154
    It's a future that we
    all of us have a hand

    358
    00:12:23,155 --> 00:12:25,500
    in creating through the
    choices we make today.

    359
    00:12:25,880 --> 00:12:26,200
    It's true.

    360
    00:12:26,600 --> 00:12:27,820
    So what role will you play?

    361
    00:12:28,320 --> 00:12:28,800
    Yeah, that's the question.

    362
    00:12:29,280 --> 00:12:31,880
    That's something to ponder as we
    wrap up this deep dive.

    363
    00:12:32,240 --> 00:12:32,440
    Yes.

    364
    00:12:32,441 --> 00:12:35,080
    Into Dario Amodei's bold vision
    of the future.

    365
    00:12:35,440 --> 00:12:36,020
    It's a good one.

    366
    00:12:36,260 --> 00:12:37,960
    We'll leave you with that
    thought and a gentle

    367
    00:12:37,961 --> 00:12:40,500
    reminder to subscribe for
    more mind-expanding

    368
    00:12:40,501 --> 00:12:43,860
    explorations at the cutting edge
    of AI and beyond.

    369
    00:15:44,560 --> 00:15:45,560
    Until next time.

    370
    00:15:45,620 --> 00:15:48,020
    Until next time,
    keep asking those big questions.

    371
    00:15:48,520 --> 00:15:48,820
    Keep learning.

    372
    00:15:49,120 --> 00:15:53,121
    Keep pushing the boundaries of what's
    possible and most importantly, keep learning.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.