Will AI Take Our Jobs? | All-In Podcast

All-In Podcast Clips
9 Apr 202421:57

Summary

TLDRThe transcript discusses the impact of AI on the job market, with various perspectives on whether AI will displace human workers or create new opportunities. It highlights the historical context of previous industrial revolutions and technological advancements, suggesting that while certain jobs may be automated, overall productivity and economic growth tend to increase. The conversation also touches on the potential of AI in robotics and the future of domestic robots, with predictions on their availability and functionality in households.

Takeaways

  • 📺 John Stewart's viral segment on AI and job displacement has brought the topic into mainstream public consciousness.
  • 🚀 CEOs like Brian Chesky and Aaron Levy anticipate significant productivity gains due to AI, potentially affecting large portions of their workforces.
  • 🤖 AI customer support agents are exemplified as driving substantial profit increases and replacing hundreds of full-time employees.
  • 📊 Historically, GDP components remain resilient over time, suggesting that economic structures adapt rather than collapse due to technological revolutions.
  • 🏭 The Industrial Revolution saw job displacement, but ultimately led to economic growth and the creation of new job categories.
  • 💡 AI is expected to increase productivity, which historically has correlated with rising compensation and the emergence of new job classes.
  • 🌐 Public perception of AI's impact on jobs is being shaped by social media and influential figures, potentially amplifying concerns.
  • 📰 Past media coverage has sensationally warned about job loss due to technology, such as the 1983 New York Times article on computers in knowledge work.
  • 🔄 The cycle of technological advancement, job disruption, and subsequent economic growth suggests that AI's impact may follow a similar pattern.
  • 🏠 Speculations on the future role of AI and robotics in domestic settings include the potential for general-purpose robots performing household tasks.
  • 🤔 The discussion emphasizes the importance of understanding historical contexts and avoiding alarmist perspectives when considering AI's role in the economy and job market.

Q & A

  • What is the main topic of discussion in the transcript?

    -The main topic of discussion is the impact of Artificial Intelligence (AI) on jobs and the economy, including the potential for AI to replace or augment various job roles and the historical context of technological revolutions and their effects on employment and productivity.

  • How has the public perception of AI in relation to jobs changed recently?

    -The public perception of AI's impact on jobs has started to enter mainstream consciousness, with more people becoming aware of the potential for AI to replace large swaths of jobs, as highlighted by discussions on platforms like The Daily Show and influential figures like Brian Chesky and Aaron Levy.

  • What are the historical components of US GDP, and how have they remained resilient over time?

    -The components of US GDP have remained surprisingly consistent over long periods of time. Consumer consumption is always around 70%, net exports are a few percent plus or minus, gross domestic investment is around the 20% level, and government consumption is also around the 20% level. These components add up to GDP and have not changed significantly since the 1920s, except during acute events like World War II.

  • What was the outcome of job displacement during the Industrial Revolution?

    -During the Industrial Revolution, there was significant job displacement as specific job classes fell to zero, leading to unemployment and the collapse of income associated with those jobs. However, new job classes were eventually created, and the economy found a way to grow despite the initial displacement.

  • How does productivity relate to worker compensation over time in the US?

    -As productivity increases in the US, worker compensation has historically tracked it. This means that as AI and other innovations boost productivity, it is expected that compensation will also rise, leading to the creation of new job classes and economic growth.

  • What is the significance of the historical cases mentioned in the transcript when discussing AI's impact on jobs?

    -The historical cases demonstrate that while technological revolutions may initially cause job displacement, they ultimately lead to increased productivity, lower costs, and the emergence of new industries and job opportunities. This provides a framework for understanding the potential impact of AI on the job market.

  • What was the New York Times article from June 2nd, 1983 about?

    -The New York Times article from June 2nd, 1983 was about how computers were eliminating jobs in industries that were effectively offline knowledge work industries at the time, such as engineering designs and architectural drawings. The article highlighted the fear that these jobs would be eliminated due to automation enabled by software.

  • How did the software revolution change the nature of work?

    -The software revolution enabled people to do much more work than they could with pencil and paper. It allowed for the creation of digital output and the emergence of new industries, leading to increased productivity and economic growth.

  • What is the potential role of AI in the future of work?

    -AI is expected to provide leverage for knowledge work, enhancing productivity and potentially leading to the creation of new job classes. Instead of simply replacing human workers, AI could enable people to shift to higher-order work, resulting in economic growth and increased productivity.

  • What is the significance of humanoid robots in the context of AI and job displacement?

    -Humanoid robots represent a further step in the integration of AI into various aspects of work and life. They have the potential to perform tasks optimized for the human body type, leading to increased efficiency and possibly new job roles. However, there are still significant safety and development challenges before they become commonplace in domestic settings.

  • What are the potential timelines for having domestic help robots in middle-class American homes?

    -The discussion suggests varying timelines, with one viewpoint suggesting that within seven years, every middle-class household in America could have a domestic help robot capable of performing various household tasks. However, other viewpoints suggest that the timeline might be longer and that the form of these robots might not necessarily be humanoid.

Outlines

00:00

🤖 AI's Impact on Jobs and Public Perception

The paragraph discusses the growing public awareness of AI's potential to revolutionize the job market, as highlighted by a viral segment by John Stewart on The Daily Show. It mentions interviews with CEOs from Airbnb and Box, who anticipate significant productivity gains and job displacement due to AI. The speaker argues that while specific job categories may be eliminated, historical trends show that new jobs are created and productivity increases, as seen in previous industrial revolutions. The speaker uses three charts to illustrate the resilience of GDP components over time, the disruption of specific jobs, and the historical correlation between productivity gains and increased worker compensation.

05:00

🚀 Historical Context of Technological Revolutions

This paragraph delves into the historical context of technological revolutions, particularly the Industrial Revolution, and draws parallels to the current AI revolution. It discusses the fear of job elimination due to automation and the eventual growth of industries through increased productivity. The speaker argues that computers, like AI today, have historically enhanced jobs rather than eliminated them, leading to the creation of new industries and economic growth. The conversation emphasizes the importance of understanding historical cases to predict the future impact of AI on jobs and the economy.

10:00

🤔 Media Narratives and Public Fears on AI

The paragraph addresses the media's role in shaping public perception about AI and its impact on jobs. It criticizes the media for hyping the fear of job loss due to AI, while downplaying other significant issues like war and national debt. The speaker argues that AI will lead to productivity gains in the short to medium term, and while job losses may occur in the long term, new jobs will be created as a result of the productivity boom. The paragraph also includes a humorous anecdote about a video clip featuring AI, highlighting the public's evolving attitude towards AI and automation.

15:02

🧠 The Limitations and Potential of AI Compared to Human Intelligence

This paragraph explores the limitations of current AI technology, referencing an expert's comparison of a four-year-old's experience to the data processed by the largest language models. It emphasizes that despite advancements, AI is still more 'artificial' than 'intelligent'. The conversation turns to the potential of AI in robotics, with speakers discussing the development of humanoid robots and their potential applications. The discussion includes the challenges of making robots safe for domestic use and the potential for AI to improve decision-making in robotics, drawing parallels to the development of self-driving cars.

20:03

🏠 The Future of Domestic Robots and Their Economic Viability

The paragraph discusses the future of domestic robots, focusing on their potential roles in households, the expected timeline for their availability, and the economic considerations. The speakers debate the feasibility of a general-purpose domestic robot, the tasks it might perform, and the price point that would make it accessible to middle-class households. They also consider alternative forms of AI applications in homes, such as specialized robots for specific tasks, and compare the potential profitability of industrial versus consumer AI applications.

Mindmap

Keywords

💡AI

Artificial Intelligence (AI) refers to the simulation of human intelligence in machines that are programmed to think and learn like humans. In the context of the video, AI is discussed as a transformative technology that has the potential to change jobs and increase productivity, but also raises concerns about job displacement.

💡Productivity gains

Productivity gains refer to the increase in output or efficiency achieved through technological advancements, better processes, or improved skills. In the video, it is suggested that AI will lead to significant productivity gains, particularly in white-collar jobs and customer support roles.

💡Job displacement

Job displacement occurs when technological advancements, such as AI, replace human labor, leading to unemployment in certain job categories. The video highlights the public's growing concern about AI wiping out large swaths of jobs and the potential impact on the workforce.

💡GDP components

Gross Domestic Product (GDP) components refer to the various sectors or categories that contribute to the total economic output of a country. The video uses historical data on US GDP components to illustrate the resilience and consistency of these components over time, suggesting that while specific jobs may be disrupted, the overall economic structure remains stable.

💡Industrial Revolution

The Industrial Revolution was a period of major industrialization that began during the 1700s and spread throughout Europe and North America. It marked a shift from agrarian economies to industrial ones, often associated with significant technological advancements and societal changes. The video draws parallels between the Industrial Revolution and the current AI revolution, noting the displacement of certain jobs but also the creation of new opportunities.

💡Public consciousness

Public consciousness refers to the collective awareness or perception of the society towards a particular issue or phenomenon. In the context of the video, it highlights the growing awareness and concern among the general public about the potential impact of AI on jobs and the economy.

💡White collar jobs

White collar jobs are typically office or administrative positions that require specialized skills or knowledge, often involving non-manual work. The video discusses the impact of AI on these types of jobs, suggesting that AI could lead to increased efficiency and potential job displacement in fields such as media and customer support.

💡Economic growth

Economic growth refers to the increase in the production of goods and services in an economy over a certain period of time, typically measured by the increase in GDP. The video suggests that despite the potential for job displacement, AI could lead to economic growth by increasing productivity and creating new industries.

💡Knowledge work

Knowledge work involves the creation, manipulation, and dissemination of information and knowledge. It typically requires specialized skills and education. The video discusses the potential for AI to augment or replace certain types of knowledge work, such as architectural design and data analysis.

💡Robotics

Robotics refers to the branch of technology that deals with the design, construction, operation, and use of robots. The video touches on the intersection of AI and robotics, suggesting that advancements in AI could lead to more intelligent and capable robots that can make decisions and perform tasks autonomously.

💡Optimus

Optimus is a project by Tesla, aiming to develop a general-purpose humanoid robot. The video mentions Optimus as an example of how AI advancements could lead to the development of robots that can perform a variety of tasks, potentially impacting the labor market and the way we live.

Highlights

John Stewart's segment on AI and its impact on jobs went viral, sparking public discussion on the topic.

CEOs like Brian Chesky from Airbnb and Aaron Levy from Box anticipate significant productivity gains due to AI, potentially affecting job roles in their companies.

AI customer support agents are driving a $40 million increase in profits, replacing the workload of 700 full-time employees.

There is a debate on whether AI will displace people or if new jobs will be created, reflecting the public's growing concern about the impact of AI on employment.

The components of US GDP have remained surprisingly resilient over long periods, with consumer consumption always around 70%.

During the Industrial Revolution, specific job classes saw a drop to zero, but the economy found a way to grow, indicating resilience in the face of technological change.

Productivity and worker compensation have historically moved in tandem, suggesting that as productivity increases, so does compensation, leading to the creation of new job classes.

The perception of AI's impact on jobs is becoming more widespread due to social media, influencing public consciousness and discourse.

An article from 1983 in the New York Times discussed the fear of computers eliminating jobs, which turned out to enhance job roles and create new industries.

AI is seen as a tool for giving humans leverage in knowledge work, potentially leading to higher-order tasks and economic growth.

The shift from manual labor to knowledge work, and now to AI-enhanced creativity, represents a significant transition in the nature of human work.

The potential for AI to replace or augment jobs is seen as a continuation of historical patterns with technological revolutions, ultimately leading to job creation.

The media's focus on AI's impact on jobs may be a distraction from other pressing issues, such as wars or national debt.

The comparison of AI's capabilities to a four-year-old's experiences highlights the significant gap between current AI technology and human cognitive abilities.

The development of humanoid robots like Tesla's Optimus represents a new frontier in AI applications, with potential implications for both industrial and domestic settings.

The timeline and cost for domestic robots to become a common household item is a topic of debate, with estimates ranging from three to seven years and prices from $1,000 to $100,000.

The potential applications of AI and robotics in industrial monitoring and maintenance show the technology's versatility and practical benefits.

Transcripts

play00:00

all right if you missed it John Stewart

play00:01

did a segment on AI on The Daily Show he

play00:04

came back he's doing I think Mondays

play00:05

every week and it went viral and it was

play00:10

about how AI is going to change our jobs

play00:12

faster than any previous labor

play00:14

Revolution so it seems like the public

play00:16

is starting to get an idea uh about AI

play00:19

wiping out large swats of jobs and it's

play00:22

starting to hit the mainstream CEOs like

play00:25

Brian chesky from Airbnb and Aaron Levy

play00:27

from box I had them on this week

play00:29

startups that's in the last year they

play00:30

said they anticipate 30 to 50%

play00:32

productivity gains for a lot of the jobs

play00:34

in their companies developers customer

play00:36

support all that stuff and we covered

play00:37

cloners AI customer support agent doing

play00:40

the job of 700 full-time employees and

play00:42

driving a $40 million increase in

play00:44

profits this year yada y y we've we've

play00:46

talked about this over and over again

play00:48

but it seems to be tipping over into

play00:50

public Consciousness we chamath have

play00:52

talked about whether humans will find

play00:55

more work to do or if this is going to

play00:56

truly displace people I think we all

play00:58

kind of feel like at least to the best

play01:00

of my memory we all feel like new jobs

play01:02

will be created but it is entering the

play01:04

Public's Consciousness what impact is

play01:06

that going to have if the public starts

play01:08

thinking AI is going to take their job

play01:10

chath I mean I think that social media

play01:13

will make this perception more

play01:15

widespread than it's been at other

play01:17

moments of Revolution and Innovation but

play01:20

we've gone through this before I'd like

play01:22

to summarize my thoughts in three charts

play01:24

okay and I call

play01:26

this this time it's not different so so

play01:30

chart number one for those watching on

play01:32

YouTube is a look at the components of

play01:34

US GDP this is from the Bureau of

play01:36

economic

play01:38

analysis now this goes from 1929 up to

play01:41

2011 so it doesn't go all the way back

play01:43

to the 1800s and we're missing the last

play01:44

decade but the point is the following if

play01:47

you can see the chart I'll if you can't

play01:49

see it I'll describe it to you which is

play01:50

that GDP the components of GDP are

play01:54

surprisingly resilient and roughly the

play01:56

same over long stretches of time which

play01:58

is that even though GDP goes up consumer

play02:02

consumption is always around 70% net

play02:05

exports are a few perc plus or minus

play02:08

gross domestic investment is around the

play02:10

20% level and then government

play02:13

consumption is around the 20% level and

play02:15

that's what adds up to GDP so that's an

play02:19

important thing to note why because in

play02:22

the absence of something very acute like

play02:24

World War II these things don't change

play02:26

over long periods of time okay so if

play02:28

that is true what happens when you have

play02:31

any kind of a revolution so let's look

play02:34

at the Industrial Revolution so the

play02:35

shift from farms to factories and what

play02:38

you saw was exactly what people should

play02:41

be worried about with respect to AI

play02:43

which is in specific job classes things

play02:46

just fall to zero so unemployment

play02:48

basically went to zero and the income

play02:51

associated with those jobs also went to

play02:52

zero so this is what people are worried

play02:54

about but if you remember the last chart

play02:57

the point is somehow we found a way

play03:00

to find growth and this is what's

play03:04

demonstrated on this final chart which

play03:06

is when you look at us productivity and

play03:07

worker compensation this is going from

play03:09

World War II to today you find that

play03:12

every time we find a new way of

play03:15

innovating compensation tends to track

play03:18

it so if you take these three things

play03:20

together number one which is the

play03:22

components of GDP rarely change number

play03:24

two is that yes there are certain

play03:26

categories of jobs that always get

play03:28

disrupted away but the third is the most

play03:31

important which is that as productivity

play03:33

goes up which is what AI should give us

play03:36

just as we've seen in the past

play03:37

compensation also goes up which means

play03:39

new job classes will be created so I

play03:42

think the macro picture if you look back

play03:44

hundreds of years is that this is like

play03:46

many other moments in time it feels more

play03:49

personal right now because we're all

play03:50

living it right none few of us live The

play03:53

Agrarian to Industrial Revolution yeah

play03:55

we missed it and few of us live the

play03:58

technological Revolution right we came

play04:00

in at the heels of it but I suspect that

play04:02

this time is not different freeberg your

play04:04

thoughts on this I think you've said

play04:07

something similar on past

play04:08

episodes but it is kind of tipping and

play04:11

into public Consciousness and it's also

play04:14

affecting White Collar jobs this time

play04:16

not just people in fields picking

play04:18

berries and so those people may be a

play04:20

little more vocal and we've seen massive

play04:22

layoffs intact massive layoffs in media

play04:25

and those jobs don't seem to be coming

play04:27

back people seem to be get taking the

play04:29

gain and just having people on the team

play04:31

be 30% more efficient as Brian told me

play04:34

on my other pod so what do you think

play04:37

freeberg is this time different or is it

play04:40

the same so here's this article from

play04:43

June 2nd 1983 in the New York Times all

play04:47

about how computers are eliminating jobs

play04:51

in industries that were effectively

play04:54

offline knowledge work Industries at the

play04:56

time creating engineering designs

play04:58

creating architectural drawing

play05:00

I think this article spoke to the fact

play05:01

that these jobs were going to be

play05:03

eliminated and as we all know those jobs

play05:06

actually got enhanced by computers

play05:08

productivity went up and new sub

play05:10

Industries emerged and in fact the

play05:12

overall Industries actually grew in some

play05:14

cases when we were fearful of them being

play05:16

replaced due to the automation enabled

play05:18

by software so I think that in this

play05:21

particular sense we can talk about the

play05:24

Industrial Revolution enabling through

play05:26

Manufacturing Systems and centralized

play05:28

production

play05:30

a a replacement of manual labor with

play05:34

machines what we're talking about now is

play05:36

a replacement of knowledge work that has

play05:38

been aided by computers with machines so

play05:41

the machines no longer even need the

play05:42

human but the reality is that these

play05:44

systems are actually going to give

play05:45

humans 10 to 100x leverage so when you

play05:49

think about that one person could spend

play05:51

three weeks making an architectural

play05:53

drawing today what if that one person

play05:55

could make an architectural drawing

play05:56

every six hours so the question then is

play06:00

do do we stop making architectural

play06:02

drawings and we fire a bunch of

play06:03

Architects or does the cost of making an

play06:06

architectural drawing drop by 90% And it

play06:09

enables us to do more detailed higher

play06:11

resolution architectural drawings across

play06:14

more places more frequently and the

play06:17

industry actually booms and what we've

play06:19

seen historically is that when

play06:20

productivity goes up costs go down the

play06:23

actual volume balloons and the economy

play06:25

grows so it's a it's an example where I

play06:29

think in this particular case we will

play06:31

see these tools creating more leverage

play06:33

for knowledge work instead of just

play06:34

simply replacing knowledge work and that

play06:36

humans will start to shift to a higher

play06:38

order of work and we'll see the economy

play06:40

grow and productivity go up uh as a

play06:43

result so so I think that's my kind of

play06:46

key read on on on the story but it's

play06:48

very hard to connect the dots for people

play06:50

without having all of these historical

play06:52

cases and I think one of the ways to

play06:54

think about doing this usefully is you

play06:55

go back to the software Revolution and

play06:57

all the stuff that we were doing with

play06:59

pencil and papers before computers we

play07:02

actually didn't lose all those jobs the

play07:04

people could now do 100 times or a

play07:06

thousand or a million times as much work

play07:08

and new Industries emerged and

play07:09

productivity went up and the economy

play07:11

grew and so we just have to have

play07:14

this this realization as this starts to

play07:16

take hold that the industries will

play07:20

change and that these systems will

play07:22

actually provide leverage not

play07:23

replacement yeah it's such a good point

play07:26

and I think what you teach your kids is

play07:28

like really important at the this moment

play07:29

in time like having a job that is

play07:32

replaced by AI or that could be greatly

play07:35

replaced by AI might be a mistake and if

play07:37

you think about being a conductor

play07:39

freedberg or a

play07:40

maestro a conductor of an orchestra I

play07:43

think that's the job of the future is

play07:44

can you work with these agents forget

play07:47

about co-pilots because that's phase one

play07:48

of all this but agents are phase two

play07:50

where you have an agent who's writing

play07:52

copy who's should the HubSpot example

play07:55

you gave before you know a designer

play07:57

who's in the cloud who's an agent an AI

play07:59

agent making you artwork and then you

play08:01

Stitch all these things together I've

play08:03

been loading chat jpt with my kids

play08:05

constantly asking history questions and

play08:07

whatever questions they want I've been

play08:09

teaching them how to use chat GPT that's

play08:11

great yeah I mean I think it's like a

play08:12

like there's this whole transition of

play08:14

humans doing manual labor to doing

play08:17

knowledge work where you're using

play08:19

software to create digital output to now

play08:22

having more folks spend more of their

play08:24

time being conceptualists or creators

play08:26

where you can kind of be an architect or

play08:29

a of something and the system just

play08:31

generates it you know you state your

play08:33

intended objective and the system solves

play08:35

for it as opposed to hey I got to go

play08:37

build the Excel spreadsheet and check

play08:39

the formula in every cell and do all the

play08:41

manual what if I just say hey here's

play08:42

what I want the model to do please

play08:44

generate it for me and you get the

play08:45

result it enables you to do a 100 times

play08:48

more work that's why I use the analogy

play08:49

conductor or yeah August leader sax what

play08:53

do you think this is uh you're you're

play08:55

you're on the populace side you really

play08:56

have your finger on what Americans think

play09:00

and that's a compliment it's a literal

play09:03

compliment but I do think you're I think

play09:05

you've become a pop especially as the

play09:07

longer I've known you we've known each

play09:08

other for over 20 years you've become

play09:09

more populist so what what's the word on

play09:12

the street here amongst you know JP and

play09:15

how do you think they're taking this

play09:16

news when they see somebody like John

play09:18

Stewart they respect when you see

play09:19

somebody like Jonathan John Stewart you

play09:21

know doing this that's like gonna hit a

play09:24

large swath of these you know you know

play09:27

Elites that we've talked about before on

play09:28

the show who are losing their jobs or

play09:30

maybe their salaries are getting capped

play09:32

well first of all Jason to quote Senator

play09:34

Gras from Gladiator I may not be a man

play09:37

of the people but I do try to be a man

play09:39

for the people yes exactly so uh oh my

play09:43

god did you see the AI did you see the

play09:45

AI from some YC commentator company

play09:47

where they like made a little video of

play09:49

us and like we're talking about

play09:51

somebody's nuts and then they were like

play09:53

you said oh I'm going to ask my butler

play09:55

to ask my assistant to ask my house

play09:58

manager to then ask my chauffeur to pick

play10:00

those up it's like it was a pretty great

play10:03

clip yeah it looked funny look to be

play10:06

frank no one cares what John Stewart

play10:08

thinks he's never been less relevant and

play10:11

less funny um this is a story that the

play10:14

media has been hyping up for months now

play10:17

CO's over so they need something else to

play10:19

scare us with and what they really

play10:21

should be talking is that we've got two

play10:22

Wars that risk spiraling out of

play10:25

control they don't want to go there they

play10:28

don't want to go there or or the naal

play10:29

debt exactly right they don't want to go

play10:31

there either because neither of those

play10:33

issues reflect well on the current

play10:35

Administration and power so they're

play10:37

going to scare us with this now look in

play10:39

the short to medium term AI leads to

play10:41

productivity gains in the long term it

play10:44

may lead to job losses but as you guys

play10:46

pointed out hopefully by then we'll have

play10:48

lots of other jobs created by the

play10:50

productivity boom that we're going to

play10:51

get and this has been the case

play10:53

throughout history with regard to

play10:54

technology improvements and if we don't

play10:56

have these productivity improvements

play10:58

what's going to drive the growth in GDP

play11:00

what's going to allow us to pay off this

play11:02

enormous national debt that seems to be

play11:04

yep you know so large that we it's unrep

play11:07

we need the productivity gains that AI

play11:09

is going to unlock without them we're

play11:11

definitely toast so look I I don't place

play11:14

a lot of stock in this John Stewart's

play11:16

story it's just one of many that the

play11:18

media is is creating to try and scare us

play11:20

about AI that's actually a great guest

play11:22

John Stewart would be a great guest

play11:23

along with Lena Khan put those on the

play11:25

list have you seen this um clip where

play11:27

Yan Lon basically says our best llm is

play11:31

50 times smaller than what a

play11:33

four-year-old has processed since

play11:34

they've foury old is awake has been

play11:37

awake a total of 16,000 hours and you

play11:40

say Okay 16,000 hours multiply this by

play11:43

3600 seconds per hour and then figure

play11:46

out like what's the um bandwidth of the

play11:49

optical nerve going to the cortex it's

play11:51

about 20 megabytes so you have 1 million

play11:53

nerve fibers per uh per per eye and it's

play11:57

about 10 bytes per second right give a

play11:58

take

play11:59

um so multiply that's 10 the 15 bytes by

play12:03

the time you're four 50 times more than

play12:06

whatever llm like the biggest in the

play12:08

world have been trained on okay so what

play12:11

that tells you is that in the space of a

play12:13

few months a baby has seen more

play12:16

information than the biggest LMS that we

play12:18

have the point is that and this is one

play12:20

of the foremost experts in Ai and really

play12:22

one of the fathers of modern AI what

play12:24

he's basically saying is it's still more

play12:26

artificial than intelligent and and

play12:30

everybody needs to take a deep breath

play12:31

and understand that there's just going

play12:32

to be a lot more work before you get to

play12:34

this omnipresent agent that just

play12:36

replaces and destroys everything and

play12:38

thinks on its own yeah I'm willing to

play12:41

bet on all of us versus a bunch of

play12:42

four-year-olds and I just want to say

play12:44

Kumbaya to Davos as well the clip from

play12:47

Davos thanks for letting us use

play12:48

that yeah it's interesting Sachs like

play12:52

um the number of jobs that will be

play12:55

replaced or augmented and then the

play12:58

creation of jobs and then you start

play12:59

thinking about well how many jobs exist

play13:01

in the real world I saw weo is doing

play13:03

Uber Eats deliveries and you just think

play13:05

wow more people are going to be able to

play13:06

afford UB breeds which is kind of

play13:08

expensive to to use so consumption is

play13:10

going to go up and then you think about

play13:11

the Optimus and then what's the other

play13:13

robot company that's making a general

play13:15

human robot figure and man those are

play13:19

starting to get really interesting and I

play13:21

think that's going to be the unlock so

play13:24

maybe you could speak a little bit to

play13:25

saxs what do you think happens when we

play13:27

start getting humanoid robots in the mix

play13:29

and do you have any investments in that

play13:31

space I don't because I don't I don't do

play13:33

that kind of Hardware R&D I mean look I

play13:35

think you're right that AI does lead to

play13:37

robotics because one of the hard things

play13:41

about robotics is just having the the

play13:44

robot not just move but understand

play13:47

what's happening in the world around it

play13:49

and then make the right decisions about

play13:51

how to react to that and so llms do

play13:55

start creating a path for the robots to

play13:58

be able to make Intelligent Decisions

play14:01

without having to be programmed with a

play14:04

bunch of if then statements right and I

play14:06

mean self-driving kind of does this too

play14:08

I mean self-driving is sort of the

play14:10

early it's kind of like the early

play14:12

prototype for these kinds of robots and

play14:16

that that's why it's not a surprised

play14:17

that Tesla is developing Optimus is

play14:19

because you think about what

play14:20

self-driving is it's it's a device a car

play14:23

with a whole bunch of cameras on it it

play14:26

takes in all that visual information and

play14:28

then it makes decisions about how to

play14:30

move and how to react and then it's it's

play14:33

trained based on mirroring human

play14:36

decisions all those human decisions that

play14:39

Tesla's been able to gather through the

play14:40

combination of self-driving with humans

play14:42

intervening allows it to

play14:44

train the the self-driving I guess brain

play14:47

you could say well and it's also Sachs

play14:49

moving at two or three miles per hour so

play14:51

it can take its time and if You' haven't

play14:53

seen this figure this combines the

play14:55

language model with what you're

play14:57

discussing Sachs so the language model

play14:59

when you show them a picture and you say

play15:01

hey and this is from figure and there's

play15:04

a this is a their robot and it says hey

play15:06

give me something to eat have you guys

play15:07

seen this before I've heard some of the

play15:09

the founders of these robotics companies

play15:10

talk about why they create robots in a

play15:12

humanoid shape and it's not just

play15:15

because they're trying to create a

play15:17

replacement for humans or something like

play15:19

that it's also because now they can

play15:21

point cameras at the way that humans

play15:23

move yes and so they can actually train

play15:25

these robots on how humans move and

play15:27

react to things so you're able to kind

play15:29

of create a large data set kind of like

play15:32

with self-driving so that the robots are

play15:34

able to learn how to how to move and I

play15:37

I've seen a different video where uh

play15:39

Optimus the Tesla robot is folding

play15:42

shirts pretty impressive yeah what's

play15:44

what's really interesting about this

play15:46

freeberg is when I spoke to the the

play15:49

people who are making these Evolution

play15:51

has made

play15:53

humans to operate in the world most

play15:56

efficiently over whatever number of

play15:58

years

play15:59

and creatures before us so in fact the

play16:02

world is optimized for the human body

play16:06

type and so maybe you could talk a

play16:09

little bit about what you think is going

play16:10

to happen with these robots freedberg in

play16:12

the in the short and medium term when

play16:14

will we have one of these robots in our

play16:16

houses what will the price point be in 5

play16:19

to 10 years and what will they be doing

play16:20

in 5 to 10 years I don't know we explore

play16:23

we should explore that question at the

play16:25

all-in summit 2024 okay shout out to

play16:27

Elon Elon can you can you bring Optimus

play16:30

to the event please Jason I don't think

play16:32

it's a 5year time frame I think it's

play16:34

longer than that that's just my guess

play16:36

and one of the reasons is if you look at

play16:38

the use of robots in call it industrial

play16:42

production today they don't want humans

play16:44

getting too close to them they're

play16:45

actually kind of dangerous because you

play16:46

have these arms flying around they move

play16:49

quickly they're very heavy you get

play16:51

banged on the head by one of them it's

play16:52

going to take you out so a good point

play16:55

the idea of having a robot in your house

play16:57

that's capable of freely moving you to

play16:59

make that so safe to a point that they

play17:02

just haven't gotten to yet with robots

play17:04

so there's just going to be like a lot

play17:05

of fine-tuning work that happens before

play17:07

this is a domestic product I think in

play17:09

the near term it's all about industrial

play17:12

applications or maybe even military

play17:14

applications well it's and if you've

play17:16

ever been to the the

play17:19

gigafactories I I was doing a little

play17:21

tour of one of them once and somebody

play17:23

grabbed me because I almost wandered

play17:24

into one of those areas and they have

play17:26

tape on the floor then they have a wall

play17:28

ET but if you even get within a

play17:31

certain closeness with this tape on the

play17:34

floor it shuts the whole thing down

play17:35

because they're afraid somebody's going

play17:37

to get crushed behind One of These Arms

play17:38

chath I'll give it to you when will we

play17:41

have one of these robots in our homes

play17:44

for the price of a Prius now by the way

play17:46

Prius is a car that costs about $50,000

play17:50

that common folk drive so $50,000 robot

play17:53

in our

play17:54

houses I think it'll be less than that I

play17:57

think it's going to be in the next two

play17:58

or three years Years you'll have a

play18:00

domestic help robot that you can

play18:02

probably pay a, bucks a month for okay

play18:05

which would be like $100,000 car payment

play18:08

that would be the equivalent of a car

play18:09

payment on a $100,000 car so okay you

play18:11

say under five you say three what do you

play18:13

think free BR same bet $1,000 a month

play18:16

robot $100,000 sticker price when will

play18:18

we have that in our homes no no I think

play18:21

it's a thousand a month thousand a month

play18:23

which would be the equivalent over

play18:25

whatever number of months and what is it

play18:27

it does it's general purpose does

play18:28

different stuff General purses robot

play18:30

$1,000 a month 50 payments I think it

play18:33

washes the dishes I think it will do the

play18:35

laundry take out the trash there'll be

play18:38

like a whole set of house household

play18:40

tasks that it will do walk the

play18:42

dog no no not responsible for a live

play18:46

creature what do you say freeberg $1,000

play18:48

a month at home robot does your dishes

play18:50

there there's a great bet for us give us

play18:53

the over under how many years fre I'm

play18:54

not sure I think the so of science come

play18:57

on man give us a year well I I don't

play18:59

think it necessarily follows this

play19:01

general purpose model I think that there

play19:03

are likely going to be

play19:07

more

play19:09

narrow application ranges and they're

play19:12

not going to necessarily be

play19:14

humanoid in form factor um I don't know

play19:18

if you guys have seen a gecko robotics

play19:20

have you guys seen this company are you

play19:21

guys investors in this nope pretty

play19:24

impressive like Suite of

play19:27

um autonomous products that do specific

play19:31

things in industrial settings so they

play19:33

have like robots that climb on the

play19:36

outside of buildings and look for cracks

play19:38

using special scanning equipment but

play19:40

they're very autonomous and how they

play19:42

operate and what they can do and they've

play19:44

got a whole class of robots that can

play19:46

then be each one of those robots can do

play19:49

many different tasks for many different

play19:51

applications and so the form

play19:54

factors they've got kind of a set of

play19:56

form factors meaning a set of robots

play19:57

that look differently

play19:59

and have different capabilities of them

play20:01

like little spider legs or arms or or

play20:03

whatever and then they can be applied to

play20:06

go do something autonomously and then

play20:08

they just run and they do it you pull

play20:09

that up you'll see it climbing walls

play20:12

riding along pipes yeah they built well

play20:15

they're not purpose that's what's

play20:17

interesting they're they're sort

play20:18

of a narrow range of applications but

play20:21

they're not specific to do only one

play20:24

thing and so they can work in different

play20:26

environments and do different things and

play20:28

so you'll kind of pick from their Suite

play20:30

of robots which ones you want to use to

play20:32

do different different tasks and then

play20:35

they go and do it it's really

play20:36

interesting they're mostly using them

play20:37

for industrial monitoring applications

play20:39

right now like looking on bridges for

play20:40

brakes

play20:42

and cleaning windows cleaning windows

play20:45

all that kind of stuff oh cleaning

play20:47

windows that's a good one yeah so

play20:48

they've got like a really cool suite and

play20:50

I think that's what we're likely to see

play20:51

in domestic settings as well all right

play20:53

so you

play20:55

play by the way I will say the success

play20:58

of gecko indicates that there's far more

play21:00

money to be made in industrial

play21:03

applications than there is in consumer

play21:05

applications today I disagree so these

play21:08

yeah I think every human's gonna have

play21:10

one of these in I think every household

play21:12

in America every middle class household

play21:14

in America will have one of these

play21:15

thousand a month robots in seven years

play21:19

I'll give it seven you you say do

play21:21

everything I say to do domestic chores

play21:24

taking out the trash folden laundry

play21:28

domestic tasks I just think it's hard to

play21:30

justify that because you're only

play21:31

spending so many hours a week doing that

play21:33

sort of stuff is it really worth a th

play21:35

bucks a month whereas in the industrial

play21:37

setting it makes a lot more sense those

play21:40

dangerous tasks like climbing on a

play21:41

bridge looking at the seams and climbing

play21:43

on a a building cleaning the windows

play21:45

those tasks take years to do sometimes

play21:49

many many years of high-risk human labor

play21:51

whereas taking out the trash and folding

play21:53

laundry might be a little bit more hard

play21:55

to justify this SP

Rate This

5.0 / 5 (0 votes)

Related Tags
AI ImpactJob MarketProductivity GainsFuture of WorkIndustry LeadersHistorical ContextEconomic GrowthTechnological RevolutionAI EthicsRobotics