The Race For AI Robots Just Got Real (OpenAI, NVIDIA and more)

ColdFusion
25 Mar 202421:26

Summary

TLDRTech startup Figure AI introduces Figure 01, a humanoid robot capable of natural language conversations and tasks execution, powered by OpenAI's integration. Amid a global labor shortage, Figure 01 aims to assist in sectors like warehouse and retail, showcasing advanced AI's potential to transform the workforce. The video also explores other industry players like Boston Dynamics, Tesla, and Nvidia, hinting at a future where AI and robotics could redefine human labor dynamics.

Takeaways

  • 😲 The Jetson's episode 'Rosie the Robot' from 1962 depicted a future with humanoid robots, which is now closer to reality with advancements in AI and robotics.
  • 🤖 Figure, a tech startup, launched 'Figure1', a humanoid robot capable of natural language conversations and performing tasks autonomously.
  • 💡 Figure1's capabilities are a result of integration with OpenAI, allowing it to understand and respond to commands in a human-like manner.
  • 🚀 Figure AI was founded in 2022 and quickly became a unicorn, raising $675 million with backing from major tech companies like Microsoft, Nvidia, and Intel Capital.
  • 📈 The company aims to address the global labor shortage by developing robots to fill essential roles in industries like warehousing, transportation, and retail.
  • 🤝 Figure AI's partnership with OpenAI is a significant step forward in enhancing the robot's ability to process and reason from natural language and improve computer vision.
  • 🔍 The robot's actions are learned and not teleoperated, using multimodal models to understand text and images from cameras and microphones to decide on actions.
  • 📉 Despite advancements, there are still challenges like latency in response times, which need to be addressed for large-scale commercial applications.
  • 🏭 Figure AI has signed an agreement with BMW to test Figure1 in their manufacturing plant, indicating a move towards real-world industrial applications.
  • 🌟 The development of Figure1 and similar robots signifies a potential shift in the labor market, raising questions about the future of human jobs and the integration of AI in daily tasks.

Q & A

  • What is the significance of the 'Rosie the Robot' reference in the script?

    -The 'Rosie the Robot' reference is significant as it draws a historical parallel to the concept of humanoid robots that were once a futuristic fantasy, as depicted in the 1962 sitcom 'The Jetsons', and contrasts it with the current advancements in robotics, indicating how close we are to making such technology a reality.

  • What is the name of the humanoid robot launched by the tech startup Figure?

    -The tech startup Figure launched a humanoid robot named Figure1.

  • How does Figure1 integrate with OpenAI to enhance its capabilities?

    -Figure1 integrates with OpenAI by utilizing their large multimodal models, which are trained to understand both text and images. This allows the robot to process and reason from natural language, enabling it to have full conversations with people and respond to external stimuli.

  • What is the primary difference between Figure01 and other Android robots mentioned in the script?

    -The primary difference between Figure01 and other Android robots is its ability to carry out full conversations with people, respond to external stimuli, and perform tasks autonomously without being operated by anyone, which is attributed to its integration with OpenAI.

  • What was the amount of funding Figure AI raised within a year of its inception?

    -Within a year of its inception, Figure AI raised $70 million.

  • Which prominent companies have invested in Figure AI, according to the script?

    -Prominent companies that have invested in Figure AI include Microsoft, Nvidia, Jeff Bezos, Intel Capital, and others.

  • What is the mission of Figure AI as stated on their website?

    -The mission of Figure AI, as stated on their website, is to 'expand human capabilities through advanced AI' by targeting the labor force.

  • What is the current labor shortage situation in the United States as mentioned in the script?

    -According to the script, there are 10 million unfilled jobs in the United States, with 7 million of those job openings being for essential roles such as warehouse, transportation, and retail, and only 6 million people available to fill these positions.

  • How does Figure1 process and execute actions after understanding a command?

    -After understanding a command through its cameras and microphones, Figure1 processes the conversation using a large multimodal model trained by OpenAI. It then decides which learned action to take by selecting and loading the relevant neural network configurations onto the GPU and executes it with a mechanical action speed of 200 HZ.

  • What is the significance of the partnership between Figure AI and OpenAI?

    -The partnership between Figure AI and OpenAI is significant as it accelerates Figure's commercial timeline by enhancing the capabilities of its robots to process and reason from natural language, which is crucial for the robot's ability to have full conversations and potentially improve its computer vision.

  • What are some of the potential applications of Figure1 as discussed in the script?

    -The script discusses potential applications of Figure1 in addressing the global labor shortage, particularly in essential roles such as warehouse, transportation, and retail, where there is a significant number of unfilled job positions.

Outlines

00:00

🤖 Introduction to Figure1: The Humanoid Robot

This paragraph introduces the concept of humanoid robots that can replace human tasks entirely, referencing the animated sitcom 'The Jetsons' and the launch of Figure1 by tech startup Figure AI. Figure1 is highlighted for its ability to converse with humans through integration with OpenAI, setting it apart from other robots like those from Boston Dynamics or Tesla. The demo showcases Figure1's natural language processing and task execution capabilities, such as picking up an apple, disposing of trash, and placing dishes in a drying rack. The company claims the robot operates autonomously, not remotely, and is designed to respond to external stimuli, indicating a significant leap in AI and robotics.

05:02

🚀 Behind the Scenes of Figure AI's Success

The paragraph delves into the background of Figure AI, a California-based tech startup founded in 2022 focusing on general-purpose humanoid robots. Despite being a young company, Figure AI has raised substantial funding, reaching unicorn status with investments from major tech players like Microsoft, Nvidia, and Jeff Bezos. The company's mission is to expand human capabilities through advanced AI, targeting the labor force to address the global labor shortage. The partnership with OpenAI is a key development, enhancing the robot's natural language processing and computer vision capabilities. The robot's operation is explained, detailing how it processes audio and visual inputs through a multimodal model to execute actions with a rapid reaction time, faster than human reflexes.

10:04

🏭 The Shift Towards Automation in the Labor Force

This section discusses the broader implications of robotics in the workforce, particularly in warehouses and assembly lines, where robots have been increasingly deployed. It introduces Digit, a robot by Agility Robotics designed to perform tasks similar to human warehouse workers, and ABB's robotic solutions for moving goods. The narrative suggests a historic turning point where technology is not just augmenting human jobs but replacing them, with Amazon planning to deploy 10,000 Digit robots annually. The conversation also touches on other companies like Toyota and Boston Dynamics working on their robotics solutions, each with a unique focus, from mobility to intricate tasks.

15:07

🌐 The Expanding Landscape of Robotics and AI

The paragraph explores the rapid advancements in AI and robotics, with a focus on how generative AI has enabled robots to dynamically understand and adjust to their environments without hard coding. It mentions Nvidia's Project Groot, an AI model designed to teach robots natural language understanding and physical coordination through simulation. The potential for an industry standard is discussed, with OpenAI providing foundational models, Nvidia offering hardware and software, and various vendors developing different robot bodies. The impact on jobs is also considered, with predictions that AI could replace or diminish a significant number of jobs across various sectors.

20:08

🔮 The Future of Humanoid Robots and Automation

The final paragraph contemplates the future of humanoid robots like Figure1, questioning whether they will fully replace humans or merely alleviate manual tasks. It also raises the possibility of military applications for such technology. The discussion acknowledges that while consumer-ready, highly functional robots are not yet available, the rapid pace of innovation suggests this could change. The video concludes by acknowledging the diverse opinions on the impact of robotics and AI on society, from those who view it as a showcase of human ingenuity to those who question its necessity, and the need for humanity to collectively navigate this new era.

Mindmap

Keywords

💡Humanoid Robot

A humanoid robot is an advanced type of robot designed to mimic human movements and behaviors. In the video, Figure1 is introduced as a humanoid robot capable of having conversations and performing tasks like a personal assistant. The script mentions how Figure1 is integrated with OpenAI, allowing it to process natural language and interact with its environment, showcasing the potential for such robots to become part of everyday human life.

💡Artificial Intelligence (AI)

Artificial Intelligence refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. The video discusses how Figure's AI capabilities enable the robot to carry out full conversations and respond to external stimuli, highlighting AI's role in advancing robotics to perform tasks that were once exclusive to humans.

💡Natural Language Processing (NLP)

Natural Language Processing is a subfield of AI that focuses on the interaction between computers and humans through natural language. In the context of the video, Figure1's ability to understand and respond to natural language is a key feature, enabled by OpenAI's technology. This allows the robot to engage in human-like conversations, as demonstrated in the script where it explains its actions after performing tasks.

💡Robotic Sales Surge

The term 'robotic sales surge' refers to a significant increase in the demand and deployment of robots, particularly in industrial and warehouse settings. The video script mentions that robotic sales hit a record high in 2022, indicating a growing trend towards automation in various sectors to address labor shortages and improve efficiency.

💡Labor Shortage

Labor shortage describes a situation where there are not enough workers available to fill job vacancies. The video discusses how Figure AI aims to address the labor shortage by developing humanoid robots to fill essential roles in industries like warehouse, transportation, and retail, where there are millions of unfilled jobs.

💡Investment in Robotics

Investment in robotics refers to the financial backing provided by various entities to support the development and commercialization of robotic technologies. The script highlights that Figure AI has raised significant funding, including from major tech companies like Microsoft and Nvidia, indicating a strong belief in the potential of robotics to transform industries.

💡Machine Learning

Machine learning is a type of AI that enables machines to learn from data and improve their performance over time without being explicitly programmed. The video suggests that Figure1's actions and behaviors are learned, not pre-programmed, indicating the use of machine learning techniques to enable the robot to adapt and perform various tasks.

💡Generative AI

Generative AI refers to AI systems that can create new content, such as text, images, or actions, based on existing data. The video mentions generative AI in the context of how robots are now able to dynamically understand their environment and adjust without hard coding, which is a significant advancement in the field of robotics.

💡Autonomous Robots

Autonomous robots are self-directed machines capable of performing tasks without human intervention. The video discusses how Figure1 is not teleoperated but is able to carry out full conversations and respond to stimuli autonomously, showcasing the progress in creating robots that can operate independently in various settings.

💡Industry 4.0

Industry 4.0 refers to the current trend of automation and data exchange in manufacturing and industrial practices. The video's discussion on the increasing use of robots in warehouses and assembly lines is a reflection of Industry 4.0, where technology is integrated to increase efficiency, flexibility, and productivity in the industrial sector.

💡Simulation for Robot Training

Simulation for robot training involves using virtual environments to teach robots how to perform tasks before they are deployed in the real world. The video mentions Nvidia's Project Groot, which uses simulation to train robots, reducing the need for physical training and accelerating the development of robotic skills.

Highlights

The Jetson's episode 'Rosie the robot' from 1962 is a precursor to modern AI-powered robots.

Tech startup Figure launched a humanoid robot named Figure1 with conversational AI capabilities.

Figure1 demonstrates autonomous tasks like picking up trash and sorting dishes.

Figure AI, founded in 2022, focuses on general-purpose humanoid robots and has raised significant funding.

Investors in Figure AI include major tech companies like Microsoft, Nvidia, and Intel Capital.

Figure AI aims to address the global labor shortage with advanced AI.

The labor force is shrinking, with 10 million unfilled jobs in the United States alone.

Founder Brett Adcock's previous ventures include a talent marketplace and an electric aviation company.

Figure AI collaborates with OpenAI to enhance robots' natural language processing and computer vision.

Figure1's actions are learned, not teleoperated, using multimodal models trained by OpenAI.

Figure1's mechanical action operates at 200 HZ, allowing for rapid response times.

The robot's natural language capabilities and human-like imperfections in speech are notable.

Figure AI is compared to other industry players like Boston Dynamics and Tesla in the robotics space.

Nvidia's Project Groot aims to provide a foundational model for humanoid robots to understand and emulate human actions.

The future of robotics may involve a shift from human labor to AI-driven automation in various sectors.

Goldman Sachs predicts significant job loss due to AI, with 300 million jobs at risk.

Figure AI has signed an agreement with BMW to test Figure1 in a manufacturing environment.

The future of humanoid robots is a topic of debate, with potential for both assistance and replacement of human roles.

Transcripts

play00:00

fascinating crazy maybe a little scary I

play00:04

mean it sounds like we're talking about

play00:05

not just using something to help humans

play00:08

do their jobs but just replacing it

play00:09

entirely could this be a new normal on a

play00:13

on a large scale hi welcome to another

play00:15

episode of Cold Fusion this scene is

play00:18

from the popular animated American

play00:19

sitcom the Jetson this episode titled

play00:22

Rosie the robot first aired on September

play00:25

23rd 1962 the main character in the

play00:28

scene is presented with a few variations

play00:30

of a robot one that's responsive to

play00:32

commands will do housework and basically

play00:35

work as a personal assistant 61 years on

play00:38

and we're one step closer to seeing that

play00:40

reality this week a tech startup called

play00:43

Figure launched a humanoid robot called

play00:45

figure1 you can have conversations with

play00:48

it thanks to their integration with open

play00:50

AI now at this point you might be

play00:52

thinking we've already seen robots like

play00:54

this over the years whether it's from

play00:56

Boston Dynamics or Tesla how is this any

play00:59

different well first let's take a look

play01:01

at the demo of the robot before we get

play01:03

into the weeds keep in mind that this

play01:05

demo as claimed by the company is

play01:07

unaltered and in real time but for the

play01:09

purposes of brevity I've shortened some

play01:11

parts all right so let's watch great can

play01:14

I have something to

play01:15

eat sure

play01:23

thing great can you explain why you did

play01:28

what you just did while you pick up this

play01:30

trash on it so I gave you the Apple

play01:33

because it's the only uh edible item I

play01:36

could provide you with from the

play01:39

table so based on the scene right now

play01:42

where do you think the dishes in front

play01:44

of you go

play01:45

next the dishes on the table like that

play01:48

plate and cup are likely to go into the

play01:50

drying rack next great can you put them

play01:54

there of

play01:58

course

play02:03

after seeing how well it did for those

play02:05

tasks and the human style conversation

play02:07

some of you will be impressed some

play02:09

excited and some even a little bit

play02:11

scared according to the company this bot

play02:14

isn't being operated by anyone it's able

play02:17

to carry out full conversations with

play02:18

people and respond to external stimuli

play02:21

and that right there is one of the

play02:23

primary differences between figure 01

play02:25

and the other Android robots we've seen

play02:27

so far but let's back up for a second

play02:30

how did a startup only founded in 2022

play02:33

make such Headway who is backing figure

play02:36

and what are their

play02:37

intentions given what Nvidia just

play02:39

announced in the past week are we moving

play02:41

towards an aid driven robotics race and

play02:44

most importantly how do these advances

play02:47

in robotics and AI impact all of us

play02:50

we'll take a look at all of that and

play02:51

have a Roundup of all the latest robots

play02:53

in this episode so sit back relax into

play02:56

your docking station as we figure out

play02:58

the future of robotics

play03:03

you are watching cold fusion

play03:11

TV it all started in the wake of Co at

play03:15

this time robotic sales began surging

play03:17

hitting a record higher in 2022 as the

play03:20

workforce retreated these robots were

play03:22

the kind that were used in warehouses

play03:24

and assembly lines the exact sort of

play03:26

predictable environment that was the

play03:28

lowest hanging fruit to train and deploy

play03:30

robots and amongst this backdrop comes

play03:33

figure AI founded in 2022 figure AI Inc

play03:37

is a California based Tech startup with

play03:39

a focus on general purpose humanoid

play03:41

robots within just 1 year of its

play03:44

Inception the company raised $70 million

play03:47

but now they've recently raised a

play03:49

whopping 675 million and now have a

play03:51

valuation of 2.6 billion officially a

play03:55

unicorn and you know it's serious

play03:57

business when the likes of Microsoft

play03:59

Nvidia Jeff Bezos Intel capital and many

play04:02

others have become investors figure's

play04:04

Mission according to their website is to

play04:06

quote expand human capabilities through

play04:09

advanced AI they aim to do so by

play04:11

specifically targeting the labor force

play04:14

according to figure the labor force is

play04:16

shrinking they mentioned that quote

play04:18

there are 10 million unfilled jobs in

play04:20

the United States 7 million of those job

play04:22

openings are for essential roles such as

play04:24

Warehouse transportation and Retail and

play04:26

there are only 6 million people

play04:28

available to fill these open positions

play04:30

while key Warehouse suppliers predict

play04:32

they will run out of people to hire by

play04:34

2024 quote as we reach the upper limits

play04:37

of our production capability humanoids

play04:39

will join the workforce with the ability

play04:41

to think learn and interact with their

play04:43

environment safely alongside

play04:46

[Music]

play04:51

us figur founder Brett Adcock is no

play04:54

stranger to the tech industry he's

play04:57

previously founded two other companies

play04:58

vety and Archer Aviation veter is an

play05:02

online Talent Marketplace and in 2018

play05:04

they were acquired for 100 million

play05:06

Archer Aviation is now publicly traded

play05:09

and they're working on building electric

play05:10

vertical takeoff and Landing aircraft so

play05:13

with his experience Adcock has built a

play05:16

team of Industry veterans including

play05:17

professionals who have worked at Boston

play05:19

Dynamics Google Deep Mind and

play05:22

Tesla perhaps their biggest breakthrough

play05:25

has come in the form of their

play05:26

partnership with open AI in a press

play05:28

release in February 2024 figer announced

play05:31

their collaboration and it will help

play05:33

quote accelerate figure's commercial

play05:34

timeline by enhancing the capabilities

play05:36

of its robots to process and reason from

play05:39

natural language using open AI research

play05:41

and large multimodal models in simple

play05:44

terms it allows for one to have full

play05:46

conversations with the robot and I

play05:47

suspect will also make it better at

play05:49

computer vision so ultimately the idea

play05:52

is simple build autonomous humanoid

play05:54

robots to address the global labor

play05:56

shortage and to do just that figures

play05:59

first r robot is

play06:01

[Music]

play06:03

here let's start by looking at how the

play06:06

robot actually works first of all every

play06:08

single action and behavior in the video

play06:11

was learned and not teleoperated the

play06:13

robot is equipped with cameras and

play06:15

microphones so when someone speaks and

play06:17

asks it to do something the transcribed

play06:19

texts and images from the cameras are

play06:21

captured to a large multimodal model

play06:23

which is trained by open AI to

play06:25

understand both text and images the

play06:27

model processes the full conversation

play06:29

including previously shared images it

play06:31

then decides which specifically learned

play06:33

Clos Loop action to take by selecting

play06:35

and loading the relevant neural network

play06:37

configurations onto the GPU and then

play06:39

executing it the mechanical action of

play06:42

figure 01 is rapid at 200 HZ which

play06:45

basically means that after it knows what

play06:46

to do it sends back instructions at 200

play06:49

times per second that's 5 milliseconds

play06:52

of reaction time for all those micro

play06:53

adjustments humans are in the order of

play06:55

250 milliseconds and fun fact a cat's

play06:58

reaction time is around 30

play07:00

milliseconds so essentially they are

play07:03

using a combination of different neural

play07:04

networks to sort through audio and

play07:06

visual inputs understand them and then

play07:08

do the appropriate actions it's actually

play07:11

kind of amazing all right so how do you

play07:13

think you did I I think I did pretty

play07:16

well the apple found its new owner the

play07:18

trash is gone and the table wear is

play07:20

right where it belongs did you notice

play07:22

the speech hesitation there it's the

play07:24

little humanlike imperfections that

play07:26

allows us as viewers to break the

play07:28

illusion that it's an in anate robot

play07:30

we've already seen this in chat GPT

play07:32

voice hey chat GPT can you describe a

play07:36

nice kitchen scene but make it very

play07:38

casual and breathe and say um and and

play07:41

things that sound very natural while

play07:43

you're

play07:46

describing sure

play07:48

thing picture this you're in this cozy

play07:51

kitchen right the sun's peeking through

play07:53

the curtains giving that warm glow you

play07:55

can almost smell the aroma of fresh

play07:57

coffee brewing um there's this gentle

play08:00

hum of the fridge in the background kind

play08:02

of soothing you know and uh maybe

play08:05

there's a bit of chatter as someone's

play08:07

flipping Pancakes on the stove it's just

play08:09

this laidback Vibe like time's taking a

play08:12

chill pill to see it play out in a real

play08:14

robot was only the next

play08:16

Frontier as aside in an episode I made

play08:19

back in late 2021 I talked about how it

play08:21

was only a matter of time before a

play08:23

company cracked the code of mixing

play08:25

dexterity mobility and natural language

play08:27

into a single robot and through 3 years

play08:29

later I think we're getting close but

play08:32

all that being said as you can see

play08:34

there's still some latency between

play08:36

talking and getting a response and I

play08:38

think that's from open AI side mainly

play08:40

although we don't know the intricate

play08:41

details of the integration voiced

play08:43

instructions probably still have to go

play08:45

up to a server and come back I've

play08:47

noticed the same delay issue when

play08:49

playing around with chat GPT voice so

play08:51

there's still some work to be done if

play08:53

these robots are going to be used in

play08:54

large scale commercial applications

play08:56

imagine an emergency response robot just

play08:58

standing there for 10 seconds while

play09:00

someone's having a heart

play09:05

attack others are simply not impressed

play09:08

months ago figure showed a demo of the

play09:10

robot making coffee some people back

play09:12

then asked if the robots are also

play09:14

learning bad habits from the training

play09:15

data the figure robot is also not the

play09:18

best at walking but we can't be too

play09:20

harsh it's still an early stage product

play09:22

as I always say criticizing is easy but

play09:25

building is hard with also such an

play09:28

incredible amount of investment and very

play09:30

few demo videos to show some have even

play09:32

drawn the comparison to a certain Trevor

play09:34

Milton if you're a regular cold fusion

play09:36

viewer you'll know what that's all about

play09:38

but obviously this is way too soon for

play09:40

that

play09:45

[Music]

play09:46

conversation broadly speaking it's an

play09:48

exciting time for the robotics industry

play09:50

so the playing field for figure AI isn't

play09:52

going to be Barren by any means let's

play09:55

take a quick look at the latest in the

play09:57

industry first up this digit which is

play09:59

made by agility Robotics and backed by

play10:01

Amazon before now robots could only

play10:04

replace part of a human job enough to

play10:06

make coffee or an omelet but this is

play10:10

digit a robot it's maker say is ready to

play10:12

do what human warehouse workers do and

play10:15

they can lift about 35 lbs you know the

play10:17

the robot weighs about what a person

play10:18

would weigh it can reach about the size

play10:20

of what a person can reach you know

play10:22

we're really going for that kind of like

play10:23

OSHA regulations and requirements for

play10:25

the robot agility robotics is funded in

play10:27

part by Amazon which is investing a

play10:29

billion dollars in new Industrial

play10:31

Technologies the fact that it can enter

play10:33

places that were designed for humans um

play10:36

and solve those challenges that exist

play10:38

without having to rebuild uh the

play10:39

environment for Automation and it

play10:42

doesn't quit and it doesn't quit there

play10:44

are about

play10:45

$250,000 so they're not inexpensive but

play10:48

you know over just a few years they're

play10:51

worth it the digit robots will be

play10:53

deployed in Amazon warehouses and the

play10:55

company aims to produce 10,000 units per

play10:58

year so Jake fascinating crazy maybe a

play11:02

little scary I mean it sounds like we're

play11:03

talking about not just using something

play11:05

to help humans do their jobs but just

play11:08

replacing it entirely could this be a

play11:10

new normal on a on a large scale I mean

play11:12

this is the thing right as a tech

play11:13

correspondent for years I've had

play11:15

companies tell me you know well this

play11:16

technology is for augmenting a human or

play11:18

making a human's job a little bit easier

play11:20

in this case it is truly right out and

play11:22

out there to replace a human worker

play11:25

Amazon warehouse workers you know that's

play11:26

one of the largest employers in America

play11:28

people really go for that job and so you

play11:31

know we're really I think at this sort

play11:32

of historic turning point where suddenly

play11:34

it's no longer the case of of of you

play11:36

know when could technology replace a set

play11:39

of Javas but this is truly something

play11:41

that's built to do that next up abtronic

play11:44

is designed to move Goods around a

play11:45

warehouse and can work for 22 hours per

play11:48

day all you need to do is swap out the

play11:50

battery every few hours deliveries start

play11:52

in 20125 the company's also working with

play11:55

NASA on a robot called Valkyrie and that

play11:57

aims to work in dangerous environment

play11:59

ments and perhaps even future space

play12:01

walks Toyota and hunde are also

play12:04

currently working on robots of their own

play12:06

but I'd say less seriously than the

play12:07

others compared to other robots as

play12:09

mentioned earlier figure o one's key

play12:11

difference is the ability to directly

play12:13

Converse in natural language in order to

play12:15

get tasks done for you it's not a new

play12:17

concept as about a year ago on Cold

play12:19

Fusion we covered a project from Google

play12:21

that did the same thing at that time we

play12:24

saw that the concept was still in its

play12:25

infancy and the robot was quite slow

play12:28

earlier this year here we saw Elon Musk

play12:30

sharing a video of Tesla's Optimus robot

play12:32

folding some clothing at first glance it

play12:35

was impressive but the curtains came

play12:36

down later when Elon added that Optimus

play12:39

couldn't do it autonomously but since

play12:41

then there's been footage of the Tesla

play12:43

bot carrying boxes watering plants

play12:45

walking crouching handling delicate

play12:48

objects such as eggs and Performing

play12:50

other domestic activities interestingly

play12:52

these abilities come from the knowledge

play12:54

and benefits from Tesla's self-driving

play12:56

efforts which is pretty cool but again

play12:58

as far as we know it lacks that natural

play13:00

language

play13:02

aspect perhaps the atlas robot by Boston

play13:04

Dynamics is the most agile robot we've

play13:06

ever seen we've seen it organized

play13:09

objects and even perform back flips but

play13:11

is this and the figure1 robot really the

play13:13

same thing Boston dynamics's Atlas is

play13:16

definitely more focused on Mobility

play13:18

whereas figure seems to be suited for

play13:20

intricate tasks there's more of a focus

play13:22

on finger dexterity alluded to by the

play13:25

five fingers with three joints on them

play13:27

and for factory deployment this this

play13:29

could make more sense in a lot of cases

play13:32

assembly line work is done by sitting

play13:33

down or being stationary at a specific

play13:36

location for long periods of time and as

play13:38

we were working on this video Nvidia

play13:40

announced project Groot a general

play13:42

purpose foundational model for humanoid

play13:44

robots it's meant to help humanoid

play13:46

robots quote understand natural language

play13:49

and emulate movements by observing human

play13:51

actions quickly learning coordination

play13:53

dexterity and other skills in order to

play13:55

navigate adapt and interact with the

play13:57

real world

play14:01

[Applause]

play14:05

about the same

play14:06

size the intersection of computer

play14:09

Graphics physics artificial intelligence

play14:12

it all came to bear at this moment well

play14:16

I think we have some special

play14:19

guests do

play14:24

we little Jetson robotics computers

play14:27

inside they learn to walk in Isaac Sim

play14:32

let's

play14:33

go five things where you

play14:38

going I sit right

play14:42

here the system removes some of the leg

play14:45

work from training robots in the real

play14:46

world via simulation the lessons and

play14:49

skills learned can be put into a

play14:50

knowledge base and transferred to other

play14:52

Robots part and video wants Groot

play14:54

machines to be able to understand human

play14:57

language sense and navigate the world

play14:59

around themselves uh and Carry Out

play15:02

arbitrary tasks the training part of the

play15:04

stack starts with video so the model can

play15:06

sort of intuitively understand physics

play15:09

then accurately modeling the robots in

play15:10

Virtual space in order to teach the

play15:13

model how to move around and interact

play15:15

with terrain objects or other robots

play15:17

Jensen described this as a gym here the

play15:21

model can learn Basics far quicker than

play15:23

in the real world as we've covered in

play15:26

another episode on the history and size

play15:27

of Nvidia the already turning heads with

play15:29

their stocks growing at unprecedented

play15:31

levels this year I'll leave a link to

play15:33

that episode in the description as well

play15:35

the company has also mentioned that

play15:37

they're working with Boston Dynamics

play15:38

agility Robotics abtronic and of course

play15:41

figure Ai and just thinking about this I

play15:45

can already see an industry standard

play15:47

forming open AI provides the

play15:49

foundational models Nvidia provides some

play15:51

of the processing hardware and software

play15:53

packages and finally there's going to be

play15:55

a plethora of different robot bodies by

play15:57

different vendors so why is this all

play16:00

happening now well the reason that a lot

play16:02

of this stuff didn't work in previous

play16:03

decades was because the instructions for

play16:06

the robots had to be laboriously hand

play16:08

coded now with generative Ai and the

play16:11

latest generations of models robots are

play16:14

beginning to be able to dynamically

play16:15

understand their environment and adjust

play16:17

seamlessly with no hard coding the

play16:20

amount of change we're going to see over

play16:21

the next 5 years 10 years will dwarf

play16:26

everything that's happened over the last

play16:27

30 we went from automating pen and paper

play16:31

right with spreadsheets to connecting

play16:33

PCS into loc Network to connecting

play16:36

networks to get the internet to getting

play16:38

the network effect of communicating

play16:40

people globally but now we're

play16:41

introducing machine intell you know

play16:43

artificial intelligence machine learning

play16:44

deep learning neural networks Etc so and

play16:47

what that enables is the automation of

play16:49

automation right and so the people who

play16:52

were writing software particularly at

play16:55

the lower end unless you are doing these

play16:57

Advanced things they're gone right the

play16:59

people that s is writing itself in doing

play17:01

itself right it's just math programmer

play17:03

back in the day so when I was writing

play17:04

code it was the algorithms were if this

play17:07

than that right ex or whatever it may be

play17:09

right but you had to guess right you had

play17:12

to use your your best instance then it

play17:14

got smarter and smarter and smarter and

play17:15

libraries to do bigger and better things

play17:17

all that is being

play17:19

automated the progress of automation has

play17:22

been a funny story to watch unfold if

play17:24

you've been following the tech World

play17:25

closely for long enough you'd remember

play17:27

all of those discussions over a decade

play17:29

ago about how manual labor jobs would

play17:31

eventually be taken by robots but as

play17:33

was' seen since the AI boom a couple of

play17:35

years ago it was the cognitive jobs that

play17:38

were the first to be impacted in 2017

play17:41

only one in five companies used AI

play17:43

according to McKenzie now half of all

play17:45

companies

play17:46

do we've already seen many companies

play17:49

laying off thousands of employees in

play17:51

part due to technological advancements

play17:53

like AI is completely the opposite of

play17:55

what most people thought would happen

play17:57

Goldman Sachs last year reported

play17:58

reported that over 300 million jobs

play18:00

could be lost or diminished by AI most

play18:03

at risk are legal architecture and

play18:05

engineering business and financial

play18:07

operations management sales Health Care

play18:11

Art and Design Goldman Sachs also thinks

play18:14

that the price of robots and their

play18:15

constitute components will drop

play18:17

drastically and they've already dropped

play18:19

drastically since last

play18:21

year unlike humans robots don't need to

play18:24

take a break they don't need to sleep go

play18:26

on holidays strike and don't complain so

play18:30

as soon as these robots are feasible and

play18:31

cheap enough companies will be chomping

play18:33

at the bit to install them so what do we

play18:36

do about all of this well it's your

play18:38

lucky day because I've already done an

play18:39

episode on life after automation it's a

play18:42

huge topic and will just be way too long

play18:44

to get into in this episode so I'll

play18:45

leave another link to that

play18:49

below okay so back to figure one at this

play18:52

point you might think all right the

play18:54

figure demo looks cool but surely it's

play18:57

going to be a long time until these

play18:58

human robots are in production well it

play19:00

seems like figure didn't want to waste

play19:02

any time even before the demo earlier in

play19:05

the year figer signed an agreement with

play19:07

BMW manufacturing they're going to test

play19:09

the figure 01 robot at the car makers

play19:11

plant in Spartenburg South Carolina but

play19:13

we're yet to get reports on how the 01

play19:15

is performing and what tasks it's going

play19:17

to do exactly but I still think that's

play19:19

pretty

play19:20

[Music]

play19:22

interesting humanoid robots like the

play19:25

figure one are a future conundrum will

play19:27

they replace humans completely or just

play19:29

alleviate us from menual tasks leaving

play19:31

the more important stuff for humans but

play19:33

on the other hand of course the military

play19:36

would probably want to have a sniff at

play19:37

humanoid robots in the future too

play19:39

looking at consumer products though

play19:41

we're pretty far away from getting a

play19:42

highly functional robot that's cheap

play19:44

that you can buy at your local store but

play19:46

the day will probably arrive figure 01

play19:50

is the worst robot from this company so

play19:52

there's only going to be Improvement

play19:54

aside from that the competition is

play19:55

heating up so if it's not them it's

play19:57

going to be another compy that's going

play19:59

to crack the code from the broadest

play20:01

perspective possible I'm sure opinions

play20:03

from all of you will be split for the

play20:05

engineering types this is an excellent

play20:07

display of human Ingenuity for the

play20:10

Skeptics they'll say it's all pointless

play20:12

and we should focus our efforts

play20:13

elsewhere the optimists will see this as

play20:16

an exciting New Era for Humanity free of

play20:18

drudgery regardless whether that future

play20:21

turns into the Jetson the Terminator or

play20:24

Utopia will just have to let Humanity

play20:26

collectively figure that one out all I

play20:28

know is that I'll be watching this space

play20:31

closely and that's where we are with

play20:33

figure one and Robotics today I hope you

play20:35

enjoyed that or found it interesting and

play20:37

if you did feel free to share it with

play20:38

someone who would find it interesting as

play20:40

well all right so that's about it for me

play20:42

my name is toogo and you've been

play20:44

watching cold fusion and if you're new

play20:46

feel free to subscribe if you like it's

play20:48

free cheers guys have a good

play20:51

[Music]

play20:57

one

play20:58

[Music]

play21:19

cold fusion it's new

play21:24

thinking

Rate This

5.0 / 5 (0 votes)

الوسوم ذات الصلة
RoboticsArtificial IntelligenceHumanoid RobotsTech InnovationLabor ShortageAI AssistantsWarehouse AutomationFuture of WorkInvestment in TechMachine Learning
هل تحتاج إلى تلخيص باللغة الإنجليزية؟