About 50% Of Jobs Will Be Displaced By AI Within 3 Years

Fortune Magazine
23 May 202426:25

Summary

TLDRIn a thought-provoking conversation, Kai-Fu Lee discusses the future of AI and its impact on humanity. He shares his insights on the potential of AI to surpass human intelligence and the implications for understanding our own cognitive processes. Lee also delves into the current state of AI, touching on the differences between the US and China in innovation and execution. He highlights his startup's mission to democratize AI and the importance of infrastructure in making AI accessible and profitable. Furthermore, Lee addresses the challenges and opportunities AI presents, including job displacement and the need to foster human qualities like empathy and trust in a world increasingly influenced by machines.

Takeaways

  • 🧠 The speaker believes AI is a significant breakthrough for humanity and could lead to building something more powerful than the human brain, although it might not provide as much insight into how the brain works.
  • 🚀 The concept of AGI (Artificial General Intelligence) is seen as a narrow-minded view since AI can surpass human intelligence without mimicking human brain functions.
  • 🌐 There's a distinction between the capabilities of AI and humans; while AI can perform many tasks better than humans, it may lack certain human qualities like empathy and compassion.
  • 💡 The speaker's startup aims to make AI accessible and open, in contrast to some major players in the field who are becoming more closed, and to prove China's prowess in execution.
  • 💼 The company's strategy involves a small AI modeling team and a large infrastructure team to save on costs and make informed decisions before investing heavily in GPU resources.
  • 📈 The speaker predicts that generative AI will be a game-changer, similar to the impact of electricity, the internet, and mobile technology, and that it could lead to significant wealth creation.
  • 🏁 The speaker warns of the potential for job displacement due to AI advancements and emphasizes the need for society to prepare for these changes.
  • 🌟 Open AI is praised for its execution and is predicted to become a trillion-dollar company in the near future, despite concerns about its closed nature.
  • 🛣️ The speaker discusses the divergence between US and Chinese AI development, suggesting that they are in 'parallel universes' with separate innovations and challenges.
  • 👨‍👧‍👦 For future generations, the speaker advises embracing AI as a tool and focusing on uniquely human qualities such as empathy, compassion, and the ability to form trust-based relationships.

Q & A

  • What was the main topic of discussion between Kai-Fu Lee and Alison during the conversation?

    -The main topic of discussion was the future of AI and its impact on humanity, including the potential of AI to surpass human intelligence and the implications of such advancements.

  • What did Kai-Fu Lee believe in 1983 regarding AI and understanding ourselves?

    -In 1983, Kai-Fu Lee believed that once we build an AI, we would understand how we think, which was his motivation to pursue the field of AI at that time.

  • What does Kai-Fu Lee think about the concept of AGI (Artificial General Intelligence)?

    -Kai-Fu Lee believes that AGI, defined as a superset of human intelligence, is a narrow-minded and narcissistic view because it assumes AI must operate exactly like the human brain to be superior.

  • What is Kai-Fu Lee's perspective on the future capabilities of AI compared to humans?

    -Kai-Fu Lee predicts that within the next two to three years, AI will be capable of more tasks than humans, but it may still lack certain human qualities such as awareness, love, empathy, and compassion.

  • What is the name of Kai-Fu Lee's startup, and what is its mission in the AI world?

    -The startup is called '01'. Its mission is to ensure that China is not left out of the AI revolution, promote an open approach to AI, and make AI accessible to everyone.

  • Why did Kai-Fu Lee decide to found '01' instead of investing in another company?

    -Kai-Fu Lee decided to found '01' because he felt that China should not be left out of the AI revolution and that it was time for China to prove its prowess in execution. He also wanted to counteract the trend towards closed AI development.

  • What is Kai-Fu Lee's view on the current state of openness in the AI industry?

    -Kai-Fu Lee believes that despite the namesake 'open AI', the industry is moving towards more closed practices. He criticizes companies like Google, Meta, and OpenAI for not being open enough and advocates for an open approach to AI development.

  • What is Kai-Fu Lee's strategy for '01' in terms of team structure and resource management?

    -Kai-Fu Lee's strategy for '01' includes having a small AI modeling team and a large infrastructure team. He emphasizes being very focused, making key decisions, and being diligent about resource management to save on GPU costs.

  • How does Kai-Fu Lee see the future of AI and its impact on job displacement and wealth creation?

    -Kai-Fu Lee is concerned about the accelerated pace of job displacement and the challenges it poses for smaller entrepreneurs and researchers. He believes that AI will create wealth but also warns of the potential for a few big players to dominate the industry.

  • What advice does Kai-Fu Lee give for preparing children to live alongside AI and machines?

    -Kai-Fu Lee advises embracing AI and using it as a tool to enhance capabilities. He also emphasizes the importance of human qualities such as trust, authenticity, teamwork, and high EQ, which he believes will remain unique to humans.

Outlines

00:00

🤖 AI's Role in Humanity's Future and Self-Understanding

The speaker, Kai-Fu Lee, discusses the potential of AI to be a significant breakthrough for humanity, comparing it to the final step in understanding ourselves. He reflects on his early belief that creating AI would reveal how human thought processes work, but now acknowledges that AI can be more powerful than humans without mirroring our brain's structure. Lee emphasizes the importance of collaboration between cognitive scientists and AI developers, hinting at the possibility of surpassing human intelligence without fully understanding our own cognitive processes.

05:01

🌟 The Vision and Strategy of 01AI Startup

Kai-Fu Lee, CEO and founder of 01AI, shares his motivation for starting the company, which includes providing China with access to AI technology and proving China's execution capabilities. He criticizes the closed nature of companies like Open AI and advocates for an open approach, making their models available on platforms like Hugging Face. Lee outlines 01AI's unique strategy focusing on infrastructure and applications, aiming for a stack that includes data, models, and infrastructure to drive profitability and sustainability, unlike the research-focused approach of other companies.

10:03

💡 Necessity Breeds Innovation in AI Infrastructure

Lee highlights the importance of infrastructure in AI, explaining the challenges of GPU reliability and the need for efficient utilization of these resources. He contrasts 01AI's approach with that of larger companies like Google, emphasizing the need for a smaller AI modeling team and a larger infrastructure team to manage costs and make informed decisions before investing in expensive GPU resources. He also discusses the importance of model flop utilization (MFU) and how 01AI has achieved a higher MFU than the industry average, showcasing their innovative approach under financial constraints.

15:05

🔮 Predictions on the Future of AI and Tech Giants

Kai-Fu Lee provides his insights into the future of AI, discussing the divergence of American and Chinese AI markets and the challenges of regulation in a parallel universe scenario. He reflects on his previous book 'AI Superpowers' and the accuracy of his predictions regarding data's value and the execution capabilities of China. Lee also shares his thoughts on the current state of tech giants like Microsoft, Apple, Google, and Open AI, expressing his bullish outlook on their future, especially praising Open AI's execution and predicting its potential to become a trillion-dollar company.

20:08

🚀 Opportunities and Challenges in the New AI Era

Lee discusses the immense opportunities presented by AI, likening it to previous technological revolutions but on a much larger scale. He suggests that both large companies and startups could see significant growth, with AI technologies advancing rapidly. However, he also raises concerns about the increasing dominance of a few big players and the challenges faced by smaller entrepreneurs and researchers due to the high costs of entry and the risk of job displacement. Lee emphasizes the need for accessibility in AI to help bridge the gap and create a more equitable landscape.

25:11

🛠 Preparing for an AI-Dominated Future

In the final paragraph, Lee addresses the future impact of AI on jobs and the importance of preparing the next generation. He argues against the notion that using AI tools like ChatGPT for schoolwork is cheating, advocating for embracing AI as a tool for enhancing productivity and creativity. Lee stresses the importance of human qualities such as empathy, compassion, and the ability to form trust, which he believes will remain uniquely human and invaluable in the workplace. He encourages nurturing these qualities in children and leveraging AI to complement, rather than replace, human skills.

Mindmap

Keywords

💡AI (Artificial Intelligence)

AI refers to the simulation of human intelligence in machines that are programmed to think and learn. In the video, AI is described as a revolutionary technology that has the potential to surpass human capabilities in various tasks. The speaker discusses the impact of AI on understanding human cognition and its role in future technological advancements.

💡AGI (Artificial General Intelligence)

AGI is a type of AI that possesses the ability to understand, learn, and apply knowledge across a wide range of tasks, comparable to human intelligence. The video mentions that AGI is often viewed as the ultimate goal of AI development, but the speaker suggests that this is a narrow perspective and that AI can excel in areas without needing to replicate human cognition entirely.

💡Cognitive Science

Cognitive science is the interdisciplinary study of the mind and its processes, including how people think, learn, and remember. The speaker in the video indicates that despite AI's advancements, understanding human cognition still heavily relies on cognitive scientists and brain researchers, as AI does not perfectly mimic human brain functions.

💡GPU (Graphics Processing Unit)

GPUs are specialized processors designed to accelerate the rendering of images and complex computations, essential for training AI models. The speaker highlights the importance of GPUs in developing AI technologies, noting their high demand and critical role in executing large-scale AI computations efficiently.

💡Execution

Execution refers to the practical implementation and operationalization of technological innovations. In the video, the speaker contrasts the US and China, stating that while the US excels in innovation, China is superior in execution, meaning China can effectively implement and scale AI technologies more efficiently.

💡Open Source

Open source refers to software with source code that anyone can inspect, modify, and enhance. The speaker criticizes some leading AI companies for being closed off and emphasizes their company's commitment to open-source models, aiming to make AI technologies accessible to a broader audience and foster collaborative development.

💡Infrastructure

Infrastructure in the context of AI refers to the foundational hardware and software systems required to support AI operations. The speaker discusses the importance of having a robust infrastructure team to maintain and optimize GPU clusters, ensuring efficient AI model training and minimizing operational downtime.

💡Job Displacement

Job displacement refers to the loss of jobs due to technological advancements, such as automation and AI. The speaker predicts a significant impact on the job market, particularly in white-collar professions, and emphasizes the need for societal and governmental measures to address this challenge.

💡Compassion and Empathy

Compassion and empathy are human emotions that involve understanding and sharing the feelings of others. The video highlights that while AI can simulate these emotions, it cannot genuinely experience them. This distinction underscores the unique aspects of human intelligence that AI cannot replicate.

💡Economic Impact

The economic impact refers to the effects of AI on the global economy, including wealth creation, market dynamics, and job markets. The speaker discusses the potential for AI to drive unprecedented economic growth and create significant wealth, but also warns of the risks of increased inequality and the need for careful management.

Highlights

Discussion on the future of AI and its impact on humanity.

AI as the biggest breakthrough for humanity and its role in self-understanding.

The possibility of building AI more powerful than humans without mimicking the human brain.

The distinction between AGI (Artificial General Intelligence) and superhuman AI capabilities.

China's potential to prove its prowess in AI execution despite being blocked from accessing OpenAI.

The importance of open-source in AI development and accessibility.

The startup's billion-dollar valuation and its place in the AI world.

The strategy of focusing on infrastructure to reduce computational costs in AI.

The comparison between AI and human intelligence, emphasizing their different operational mechanisms.

Predictions on the future of AI and the competition between US and Chinese AI markets.

The potential for AI to replace 40-50% of jobs within the next 10-15 years.

The importance of embracing AI tools for future generations.

The unique aspects of human empathy, compassion, and EQ in contrast to AI capabilities.

The role of trust and authenticity in human success, which AI cannot replicate.

The future trajectory of wealth creation in the AI industry and its impact on capitalism.

The challenges faced by startups and researchers in the era of large AI companies.

The potential for generative AI to revolutionize data usage and problem-solving.

Transcripts

play00:00

hu and I have been talking backstage and

play00:02

we were deciding should we give you guys

play00:04

some sweet dreams tonight or some

play00:05

nightmares and we decided maybe a little

play00:08

bit of both uh we have a lot to discuss

play00:10

what the future looks like um what we're

play00:13

building right now and how it will

play00:15

impact us in the future um so really

play00:17

excited for this conversation K thank

play00:19

you so much for being here thank you

play00:21

Alison so you've said AI is the biggest

play00:23

breakthrough for Humanity um it is also

play00:26

the final step to to understanding

play00:28

ourselves so I'm curious Kyu why do you

play00:30

think that and what exactly does that

play00:31

mean to understanding ourselves I think

play00:33

I wrote that in my PhD application in

play00:37

1983 and yes I at the time I thought uh

play00:42

once we figure out how once we build an

play00:45

AI we'll know how we think and and that

play00:49

was what drove me to go into AI at the

play00:52

time uh was basically AI winter but what

play00:55

I found out today is I think we can

play00:57

build something much more powerful than

play00:59

human not necessarily AGI but much more

play01:03

powerful than human and it doesn't have

play01:05

to follow exactly our brain so so the

play01:08

good news is we can build something much

play01:10

more powerful than I thought the bad

play01:12

news it is that it may not give us as

play01:15

much insight about how the brain works

play01:18

we still have to rely on the cognitive

play01:20

scientist and the Brain uh scientist in

play01:22

the world so wait are you saying that

play01:24

there's something Beyond AGI I thought

play01:26

artificial general intelligence was like

play01:28

the end all be all end point

play01:30

when are we going to reach that and

play01:31

what's beyond yeah AGI is defined to be

play01:34

a superet of human intelligence

play01:35

everything the human can do AI can do I

play01:38

think it's a very narrow-minded and

play01:40

narcissistic view because we are humans

play01:43

so when you watch movies you want aliens

play01:44

to be aliens want to become humans pets

play01:47

be want to become humans monkeys want to

play01:49

become humans but they don't you know AI

play01:53

is this giant machine that gets better

play01:56

with more gpus and computes and data it

play01:59

can do so so so many more tasks so much

play02:03

better than people but it doesn't mean

play02:05

it does everything people do because

play02:07

their brain operates differently from

play02:09

our brain you know we don't compete we

play02:11

don't compare um you know marathon

play02:15

runners with cars right um and I don't

play02:18

think we need to compare AI with us even

play02:21

though it is exhibiting a lot of

play02:23

intelligence so I'm certain so if we

play02:26

look at what can humans do and what can

play02:28

AI do uh five years ago humans can do a

play02:32

lot AI can do a bit there's some overlap

play02:35

um and today what AI can do is gotten

play02:38

larger than what humans can do but but

play02:40

it's not a complete superet I think in

play02:43

the next two to three years uh if the

play02:45

human circle of what human capabilities

play02:47

is this big uh you know the AI will be

play02:50

the size of Earth but it still may not

play02:52

do what everything we can do it it may

play02:54

not have awareness love um empathy

play02:58

compassion um or other skills so

play03:01

something super human greater than us

play03:03

all that has no compassion it sounds

play03:04

like a great great world we're building

play03:06

great and it can fake compassion oh god

play03:09

oh boy okay so we have a lot to discuss

play03:12

you have startups you have a startup

play03:14

you're running you are a VC have been

play03:17

investing for years you've also worked

play03:18

at just about every major tech company

play03:20

Under the Sun um who's a leader in AI so

play03:23

lots and lots to dig in there into there

play03:26

but first I want to talk about 01 a okay

play03:29

uh your startup which you're the CEO and

play03:30

the founder spun up about exactly a year

play03:32

ago you had no team today your a billion

play03:35

dollar valuation no Revenue some Revenue

play03:38

some Revenue some Revenue some revenue

play03:40

on track for Revenue more Revenue um so

play03:42

what is it and what is its place in this

play03:44

AI world uh yeah I I felt that um you

play03:50

know people in China cannot access

play03:52

chpt um actually CH many of you may not

play03:55

know this but um open AI blocks China

play03:58

from ability to access

play04:00

so I feel that China shouldn't be left

play04:02

out of this Revolution and um I felt

play04:06

that also I strongly believe as I stated

play04:08

in the AI superpowers that uh us will

play04:11

lead in breakthrough Innovations but

play04:13

China is better at execution and I

play04:16

thought a year ago the time has come for

play04:19

China to prove its prowess in execution

play04:22

and rather than investing in someone I

play04:24

would do it myself this time so that was

play04:27

the initial attempt uh the other is kind

play04:30

of a frustration that the field is

play04:33

moving toward more closed despite its

play04:35

namesake open AI is not open at all it's

play04:38

probably the most closed company in the

play04:40

world even compared with Google and uh

play04:44

Gro and meta and and and others so I

play04:48

felt you know uh we need to work with

play04:50

academics we want to open with work with

play04:52

open source community and entrepreneurs

play04:55

and people like yourselves and that we

play04:57

should not only make AI work great and

play05:01

make it create value but also make it

play05:03

accessible to everyone and if the best

play05:06

company in the bunch is closing all of

play05:09

its Technologies never open source don't

play05:12

publish then we can't engage and bring

play05:15

in all the brilliant people in the world

play05:17

so we decided uh we would also take an

play05:19

open approach which is a bit unusual

play05:22

because you know uh us is known to be

play05:25

you American companies are known to be

play05:27

more open than Chinese companies but

play05:28

we've taken an open approach uh to date

play05:31

every best model we have built from tax

play05:34

model to multimodal model we've made

play05:36

available in open source their on

play05:38

hugging face and other sites because we

play05:41

felt we wanted to um change the way

play05:44

people think and make it accessible to

play05:46

more people and then maybe lastly what's

play05:49

unique is that uh I feel that this I'm

play05:53

I'm kind of

play05:54

seeing deu that what happened with um PC

play06:00

and the phone I'm seeing again with AI I

play06:03

know we're people love to talk about AGI

play06:05

but practically speaking about making

play06:07

money it's very analogous right PC was

play06:12

uh creating a stack from CPU to uh to to

play06:17

operating system to applications and

play06:20

then to servers and to cloud and then to

play06:23

the end user and to business and to C

play06:25

and to be basically that whole stack

play06:27

made Microsoft the amazing company

play06:30

and made them incredible amount of money

play06:33

and the phone more or less did that for

play06:36

uh Google and apple and I think the same

play06:39

opportunity exists but but I feel a lot

play06:42

of the llm companies on out there are

play06:46

run by researchers who care only about

play06:49

making a great model and I think that

play06:52

science fair phase needs to end um no

play06:56

matter how much Brilliance and great

play06:58

demos you make uh at some point there is

play07:01

a uh a point of Reckoning when investors

play07:05

are going to say what do you have to

play07:07

show for yourself what's your p&l what's

play07:09

your Revenue what's your growth when do

play07:11

you break even and I I've learned a lot

play07:14

of that myself having been a researcher

play07:16

from the beginning but now that I've

play07:18

been a VC for 14 years I feel like um I

play07:21

can make this thing make money so we're

play07:23

not spending a lot of time on it but

play07:25

we're building the company with the

play07:27

ambition that this company will have

play07:29

have infrastructure model uh

play07:33

applications and also data working as a

play07:37

stack and we've come up with a number of

play07:39

ways which we think in the future will

play07:41

make money and we don't do it because we

play07:44

want the money but we do it because we

play07:46

see the need to continue to raise money

play07:48

to afford our gpus and we want to do it

play07:52

by not creating a bubble or promising

play07:55

the moon um uh or through demos we want

play07:58

to do it through through actual Revenue

play08:00

growth and profit over time sounds very

play08:03

expensive sounds like something Venture

play08:05

Capital alone cannot fund I mean you

play08:07

have Google with gazillions of dollars

play08:09

at their disposal meta as well I how do

play08:11

you even compete in that World um to

play08:14

give you an example um when you think

play08:17

about gpus and models most of you

play08:20

probably just think okay he's they've

play08:21

got so many gpus then they can do great

play08:24

but actually Google has over 2,000

play08:26

people just in its um deep Division and

play08:31

those 2,000 people are competing for

play08:33

gpus and resources and Google's uh

play08:36

approach is let you know 100 flowers

play08:38

bloom and that's why open AI leaped

play08:41

ahead by betting on one approach uh with

play08:44

as much less GPU than Google had years

play08:47

ago now they have both have a lot so um

play08:51

I I'm think I I think the approach we

play08:53

think is we want a very small AI

play08:57

modeling team and a very large

play08:59

infrastructure team because if you have

play09:02

too many researchers and a culture where

play09:04

everybody can try ideas you'll quickly

play09:07

run out of money as a startup as you

play09:09

said but if we're very focused we want

play09:12

these researchers to read every paper to

play09:15

understand every technology to discuss

play09:18

things um theoretically argue from first

play09:21

principles and then build some

play09:23

experiments to fight them fight it out

play09:26

and then before we really spend the

play09:28

expensive GPU we will have either

play09:30

reached consensus or I would have to

play09:32

make a decision so this is a very

play09:34

different culture than what Google and

play09:36

other companies try to do we're

play09:38

basically trying to gather lots of data

play09:40

and make um key decision so we don't

play09:43

take as much GPU as an example you may

play09:46

have read a paper by Steven Levy about

play09:48

Huawei and the polar code I won't go

play09:51

into details here but basically Huawei

play09:53

asked this Engineers to read all the

play09:55

papers and they found this Turkish

play09:57

Professor who invented something called

play09:59

the polar code and that made all the

play10:01

difference and gave huawei's leadership

play10:03

position in 5G we're taking that same

play10:05

approach to be very very diligent to

play10:07

save GPU then the infrastructure is one

play10:10

of the least appreciated but most

play10:12

important Technologies because these

play10:16

these gpus don't work very well they

play10:18

break 4% 4% of the time each GPU breaks

play10:22

so if you have a cluster of 10,000 gpus

play10:24

half of your month is gone because the

play10:26

cluster every time a GPU goes down the

play10:28

whole clust cluster goes down you've

play10:30

heard Jensen hang talk about how

play10:32

Blackwell has ways to recover well we

play10:34

can recover today by patching H H 800s

play10:38

that we have in in the company and also

play10:41

uh there's a metric called mfu which

play10:44

measures the uh model flop utilization

play10:47

um most llm companies are around 40% um

play10:51

you know Google Nvidia are slightly

play10:52

better but we're at 63% so we basically

play10:56

you know I think I truly believe that uh

play11:00

necessity is the mother of innovation

play11:03

and the necessity for a poor company

play11:06

like us compared to a Google open AI

play11:09

Force us to be diligent about how the

play11:12

approach to take to be decisive about

play11:15

which approach to take and also to build

play11:18

a large infrastructure team to reduce

play11:20

the cost of computes because we just

play11:22

don't have that many got

play11:24

it okay um well big journey ahead um

play11:29

you're also called the tech Prophet so I

play11:31

want to get some prophecies here from

play11:33

you while I have you but let's look back

play11:35

at you've written two books um both sort

play11:38

of predicting the future in their

play11:39

different ways superpowers um was sort

play11:43

of the US versus China and who will win

play11:46

um who will win uh where are we now

play11:49

where are both and were you

play11:51

right uh the our last last answer to

play11:54

your last question is of course

play11:56

yes uh the prediction I made in my book

play11:59

was was First Data is the new oil that

play12:01

makes all the difference and I think

play12:03

look what generative AI is right take

play12:05

all the data in the world to train using

play12:06

generative as a way to uh solve the

play12:09

objective function problem second

play12:11

prediction in the book was that us was

play12:13

better at Innovation breakthrough China

play12:16

was better at execution and we saw that

play12:19

in the earlier AI days and in the recent

play12:22

Genai I think we're just about to see it

play12:25

I think we'll have to see how this works

play12:27

out but certainly uh taking my company

play12:29

as an example we were eight years behind

play12:32

a year ago now we're probably less than

play12:34

one year behind of the top American

play12:37

company so at least so far we've been

play12:39

executing of course the last one year is

play12:42

the hardest right because think about uh

play12:45

what open AI has achieved in one year so

play12:47

we we don't view it uh in any kind of um

play12:50

um hubris or um uh taking for granted

play12:53

but we have closed the Gap because we

play12:55

did execute better in the past year uh

play12:58

in my company and in other Chinese

play13:00

companies um so I I believe in that

play13:03

still the other point I made was well

play13:05

when can us have an unsalable advantage

play13:08

over China and the point I made in the

play13:10

book was when the technology

play13:14

became invented by some corporate

play13:17

company uh by some corporate entity not

play13:20

academics because academics were

play13:22

published and that corporate entity

play13:24

chooses to stop publishing which we are

play13:27

seeing now so it's definitely plausible

play13:29

that us can extend this leadership

play13:32

because open Ai and to a lesser extent

play13:34

Google have stopped publishing um but

play13:36

but we'll see so are China and the US

play13:39

headed on different paths here will they

play13:41

even be competitive in their a

play13:42

respective AI markets or will there be

play13:44

the China option for the world to engage

play13:46

with on on AI and the US version and

play13:49

then if that's the case how the heck do

play13:50

you regulate this all if they're

play13:51

separate things yeah we're well beyond

play13:53

that and we're in our parallel universe

play13:56

this is not ideal none of us like it I

play13:58

think but it is what it is and given

play14:01

it's a parallel universe I think we're

play14:03

going to uh see a lot of interesting

play14:06

American solutions that won't make it to

play14:07

China and interesting Chinese solution

play14:10

that won't make it to us so as an

play14:12

entrepreneur or VC or just a curious

play14:14

mind it would make sense to look at both

play14:17

worlds to see what's exciting in each

play14:19

one because you'll while you compete in

play14:22

your world whether it's the American

play14:24

world or the Chinese world if you have

play14:26

the advantage of actually having studied

play14:28

the other parallel universe you will

play14:31

have superpowers in the work that you do

play14:34

but we don't have have any unrealistic

play14:36

uh expectation that uh us and Chinese AI

play14:40

companies will actually compete um in

play14:42

the same country unless it's in a small

play14:45

number of countries that are friendly

play14:47

with both

play14:48

countries um when you look at the

play14:51

world's biggest tech companies today uh

play14:53

you've worked for at least half of them

play14:55

Apple Microsoft Google who is still on

play14:59

top in 10 years or even 5 years Who

play15:02

falls and which uh startups kind of

play15:05

overtake them yeah um I think on the US

play15:09

side um I mean right

play15:12

now Microsoft is a darling and um uh

play15:16

Apple's doing stuff and uh Google is uh

play15:20

frustrated and a lot of negative uh

play15:23

comment and then open AI is the upst

play15:26

starter I would say that despite my

play15:28

earlier comments on open AI I am very

play15:30

bullish on their future uh they've

play15:33

really done an admirable and

play15:34

unbelievable job executing um even today

play15:38

GPD 4 is still the gold standard you see

play15:41

Gemini Ultra and cloth 3 and make these

play15:43

claims but if you use these models GPD 4

play15:46

and gbd4 turbo is unbelievably good and

play15:49

a great balance for uh performance and

play15:52

um uh cost so I I I I would say um you

play15:57

know open AI will like be a trillion

play15:59

dollar company in the not too distant

play16:01

future how distant or how not so distant

play16:04

two three years trillion dollars in two

play16:06

or three years I think that's the likely

play16:08

outcome of course they could Mis execute

play16:11

or some some other company could do a

play16:13

great thing but I I despite my concerns

play16:16

about their lack of openness I have

play16:18

great admiration for them so I would if

play16:20

I could invest in any of them which I

play16:22

can't but if I could I would do open a

play16:24

ey uh Nvidia is another one that's a

play16:27

safe bet it's obviously extremely

play16:29

expensive but the processors they have

play16:32

and the Cuda and the um Technologies

play16:36

they have on um um the the ecosystem and

play16:39

the libraries it's very hard for a

play16:41

competitor to dislodge them they're

play16:44

obviously very expensive but um you know

play16:47

most of you know in secondary stock you

play16:50

uh you buy high and sell higher rather

play16:54

than Buy Low you none of these companies

play16:55

are going to go low Microsoft is exe

play16:58

exec it brilliantly um I am a little bit

play17:01

disappointed at the Microsoft co-pilot

play17:03

because it is

play17:04

gluing gen to an existing product that

play17:09

is a dinosaur that should be thrown out

play17:11

uh but I understand why they did what

play17:13

they did but I would rather see a new

play17:16

product that throws away Microsoft

play17:18

Office and that has AI do most of the

play17:21

writing and human just kind of sit in

play17:24

the back seat twe tweaking it a little

play17:26

bit uh but nevertheless I think my

play17:28

Microsoft's done an amazing job um on

play17:31

selecting open a and partnering with

play17:33

them and I think Sati has demonstrated

play17:36

unbelievable leadership in trying to

play17:39

bring in uh Sam Alman and then when that

play17:41

couldn't gracefully letting him stay and

play17:43

now getting um Mustafa it's just amazing

play17:47

to see uh SAA has been the most

play17:50

phenomenal uh CEO uh Google I'm still

play17:53

somewhat bullish uh despite all the

play17:55

issues we all know um it is still has

play17:58

currently has the largest density of AI

play18:02

talent in the world by far uh more than

play18:05

open AI more than Microsoft more than

play18:07

the others so I guess the question is

play18:09

can they start to execute better so

play18:13

um be that as they may the incumbents

play18:16

are certainly in great position each of

play18:18

them um with their War chest of cash but

play18:20

it's a new era for entrepreneurs as well

play18:22

yourself included um if you were well

play18:25

you are an entrepreneur where do you

play18:26

start um if you want to get into this AI

play18:29

world and what kind of wealth creation

play18:30

are we talking about here I mean wealth

play18:33

in equity is already so vast you're

play18:36

talking about this company open AI can

play18:37

be worth trillions almost overnight a

play18:40

couple years that's very fast what is

play18:43

the trajectory here um for wealth

play18:46

creation for capitalism in general for

play18:49

the Haves and the Have Nots well it's

play18:52

this is by far the most advanced and

play18:54

most amazing technology compared to

play18:57

anything by a factor of 10 right so if

play19:00

we Look Backwards electricity internet

play19:03

PC Mobile is nothing compared to this so

play19:06

if you subscribe to that view right then

play19:09

um then there's no reason why any of the

play19:12

companies we mentioned large or small

play19:14

couldn't go up by 10x right including

play19:17

ours hopefully even more because we're

play19:19

cheaper but the IM mature companies U I

play19:22

don't see why they couldn't go go by 10x

play19:25

uh because I think um there this is just

play19:29

so and also many of the problems that

play19:31

plague these systems like hallucination

play19:34

we we I can predict uh that they will be

play19:36

more or less solved in a year and a half

play19:38

or so so um yes they hold things back

play19:42

but I'm very bullish on these

play19:44

Technologies moving forward um of course

play19:46

on your other question I'm also worried

play19:49

that uh a few big players will dominate

play19:52

more than ever before um and um and that

play19:57

if if one company dominates it would be

play19:59

a really horrible thing if five or 10

play20:02

companies are doing better that's a lot

play20:04

better but even in that case uh it does

play20:08

create an accelerated pace of um job

play20:12

displacement and accelerated um set of

play20:15

difficulties for smaller entrepreneurs

play20:17

if you could only raise five or10

play20:19

million and you decide to do an app um

play20:23

take a look at Jasper right they built a

play20:25

great app at the time but the foundation

play20:28

model sucked in all the learnings and

play20:30

the app became of much more questionable

play20:32

value and that is not even by evil uh

play20:36

Planning by the platform provider is

play20:39

just the natural power of the

play20:41

fundamental models to suck in all that

play20:44

you feed it so that creates um I think

play20:47

uh serious challenges for Have Nots the

play20:50

and also researchers uh if you're

play20:53

talking about lack of gpus I mean we I I

play20:56

try to cry to you how much how few GP

play20:58

CPUs we have but you know we have 100

play21:01

times more than 99.9% of universities so

play21:05

what will professors do so the halves

play21:07

and Have Nots have become at the

play21:09

Historical um Gap where professors

play21:13

entrepreneurs and people who have um

play21:17

don't have the skill sets and people who

play21:19

are working on White Collar relatively

play21:22

routine jobs I worry a lot about the uh

play21:26

disparities uh for them and that's again

play21:28

why on 01 I chose uh by you know making

play21:32

this great technology accessible as our

play21:35

mission and and the the underlying

play21:38

emphasis on is on the word accessible

play21:41

because I think all of us should do what

play21:43

we can to avoid the case of an extreme

play21:46

halves and half Nots and job

play21:49

displacement I think we all know this is

play21:51

coming um you had said around 2017 you

play21:54

thought in 10 to 15 years about 40 to

play21:57

50% of all jobs would be replaced by AI

play22:00

is that still an accurate timeline in

play22:02

your opinion um what the heck is

play22:05

everyone going to do when they don't

play22:06

have a job in three years if so it's

play22:09

actually uncannily accurate people have

play22:12

criticized me for being too aggressive

play22:16

in the 2017 17 18 19 and I was a little

play22:20

nervous at the time but when J ji came

play22:22

came out I think everybody's on the

play22:24

bandwagon and believing that is the

play22:28

correct Pace um and I think the white

play22:31

job collar jobs will go a lot faster

play22:33

blue collar job maybe a little slower

play22:35

because more people are shifting to the

play22:38

software only

play22:39

displacement and and I think it's a very

play22:42

very significant problem and I think

play22:45

finally some governments are waking up

play22:47

and realize they have to do something

play22:49

about this and in my AI 2041 I outlined

play22:53

a number of um creative maybe not

play22:56

necessarily workable Solutions that will

play22:59

that was intended to get people thinking

play23:01

so get a copy of the

play23:03

book um well uh so we've got a lot ahead

play23:08

somehow we're out of time and we need to

play23:09

let these people eat but the one

play23:11

question I do want to leave everybody

play23:13

with is can we get some hope here um how

play23:15

should we prepare our kids to live

play23:18

alongside machines right um if this is

play23:20

what's coming and it's coming fast you

play23:22

know let alone all of our jobs we need

play23:24

to be thinking about how to work this in

play23:25

and and help employees and help

play23:27

everybody but what do we do for our kids

play23:29

when they say what should I be when I

play23:31

grow up yeah I I think first thing we

play23:33

all have to do and influence all the

play23:35

people around us is to stop this

play23:37

nonsense about kids are using chadd to

play23:40

cheat right this is not cheating any

play23:42

more than using word or photoshop the

play23:45

kids when they go into the workplace are

play23:46

going to be measured based on the final

play23:49

output of their work they're not going

play23:51

to be measured on what did you use chbt

play23:53

did you use Google search so I think we

play23:56

need to en encourage people to to

play23:58

harness Ai and use all the tools so that

play24:02

they can be the best that they can be

play24:04

and also it's a great guide to uh what

play24:07

things they can aspire to and what

play24:09

things are not worth uh following I

play24:12

think that's incredibly important is to

play24:14

harness Embrace Ai and stop trying to

play24:18

catch people cheating this is not

play24:19

cheating this is um producing great

play24:22

output it's not any more cheating than

play24:25

um Fortune journalists using Microsoft

play24:27

Word Spell checking or a um um a a

play24:31

fortune photographer using Photoshop

play24:33

this is chbt is a tool everyone should

play24:36

use it and learn it learn from it uh and

play24:39

embrace it and and the second thing is

play24:42

that um I think there needs to be a

play24:45

belief that there is something unique

play24:46

about our Humanity I continue to hold

play24:49

out that we have a soul the machine

play24:52

every will that we have compassion and

play24:54

empathy we have emotions and the ability

play24:57

to love we have the ability to connect

play24:59

to other people and create trust and win

play25:02

trust and that all of you know as

play25:04

successful people in your companies your

play25:07

success is built more more than anything

play25:10

more than your technical skills more

play25:12

than your business skills the most

play25:14

important skill that I tell any young

play25:16

people for the last 20 years is that

play25:19

winning trust from other people is most

play25:22

important and winning trust is about uh

play25:25

authenticity about teamwork about

play25:27

sharing sharing um and it's about high

play25:30

EQ not just the IQ so I believe in that

play25:35

holds out something that we can all

play25:38

Embrace uh you don't have to be a genius

play25:40

to have a high EQ you don't have to be a

play25:43

genius to to love and be compassionate

play25:46

and in both of my books I talk about

play25:49

that as the essence of our humanity and

play25:52

I continue to believe it uh do I think

play25:55

AI can fake it yes do I think people

play25:58

will accept the faking AI at least for

play26:00

the next 50 years no so that's long

play26:03

enough for your kids to survive and

play26:05

figure out the next step for their kids

play26:08

okay our kids will survive and they'll

play26:09

figure out something else later um thank

play26:12

you so much Kaiu I wish this had been an

play26:14

hour it was such a pleasure chatting

play26:15

with you thank enjoy your meals

Rate This

5.0 / 5 (0 votes)

Related Tags
Artificial IntelligenceFuture PredictionsHumanityTech InnovationAI EthicsCognitive ScienceJob DisplacementEducationEntrepreneurshipGlobal Impact