00:00:02
[Music] all right welcome back to another
00:00:07
episode of the waveform podcast we're your hosts i'm marquez and i'm andrew and today i think we're gonna i think
00:00:12
we're gonna talk about the tesla bot in every detail possible there's a lot of there's a lot of versions of how you
00:00:19
can talk about the tesla bot but just for those who haven't heard the overall announcement that we're
00:00:24
addressing is tesla had an announcement at their ai day where they announced a humanoid robot
00:00:32
and now andrew is if you aren't watching the video you can't see but he's wearing a a helmet that makes you
00:00:39
makes you look like a humanoid robot oh god yeah i feel like the daft punk
00:00:44
combo with the with the humanoid robot is that's something it's something the one helmet the gold that bunk robot
00:00:51
looks pretty similar to what it's much better dress yeah yeah so i
00:00:57
guess let's uh let's just break down i'm gonna start with like how we got here yeah how we got here and what it is so
00:01:02
tesla makes cars tesla's been making electric cars with air quotes self-driving and driver
00:01:08
assist for a long time um and suddenly i don't know where they unveil that they're going to be making
00:01:14
this like humanoid robot and they they show this like sleek slender like black
00:01:19
and white robot with like a head and shoulders and fingers and and feet and everything
00:01:25
and we're like how did we get here how is tesla suddenly going to be a robot company so if you think about it if you go back
00:01:31
a little bit elon basically said on stage
00:01:36
tesla is essentially one of the largest if not the largest robotics companies in the world if you want to frame it like
00:01:41
this yeah and of course this is ai day and this is you know a recruiting event and obviously they're talking a lot
00:01:46
about the way they do things in computer vision and ai but
00:01:52
i can kind of see it because if you describe a tesla as a semi-sentient
00:01:57
robot on wheels navigating around the world uh using its sensors mostly
00:02:02
cameras yeah it kind of makes sense okay it's one version self-driving cars is one version of applied robotics and ai
00:02:10
in computer vision so okay cool so they make a lot of they make a bunch of these robots and they happen to be in the forms of what we
00:02:16
call cars but um this next form factor of a robot is
00:02:22
going straight to humanoid robot and they really landed hard on this one
00:02:30
sentence i kept hearing over and over again which is that this robot will be able to do things that are dangerous
00:02:36
repetitive or boring to humans and lots of other questions were asked by the way of them on stage like elon
00:02:43
and some of the fellow engineers on stage talked about the dojo chip and talked about the way tesla's compute
00:02:48
with the self-driving computer and the way they're going to be doing super computers and learning in the future and a lot of they had the robot behind
00:02:55
them on stage the whole time so a lot of the questions were like a model of it yeah yeah a lot of the questions were like hey by the way this robot's
00:03:01
interesting why does have to have ten fingers like what's the biggest challenge about this humanoid form and like do you think this is something that
00:03:07
people are gonna really wanna deal with is there an emotional component to making it human shaped and they really
00:03:14
can find their answers to well we'd like this to be able to do things that are dangerous repetitive and boring to
00:03:20
humans that's kind of all they said and so it the rest was left up to us to
00:03:25
infer and look i'll be the first one to say i'm not a robotics expert a lot of people are like
00:03:30
sorry marquez you're way out of your lane talking about robotics i don't know i feel like i'm i'm in the tech space and this is a tech thing and i have
00:03:36
feelings about it and on top of all this we very recently did a retro tech episode retro tech season 2
00:03:44
all about humanoid robots yeah and i learned a lot about i mean obviously people have been
00:03:50
around longer than me and have learned plenty about this but about like the history of pop culture and people's obsessions with humanoid robots just how
00:03:57
many human robots you you see in like from everything from star wars to like the jetsons to like just everything
00:04:05
future facing seems to include some sort of human shaped robot into like why we haven't gotten
00:04:11
human-shaped robots today and a lot of what we found in all this research and all the interviewing i did and what
00:04:17
eventually the episode sort of summarizes is that it turns out the human shape for a robot
00:04:24
isn't necessarily the best shape for a robot to actually do things and this is this is probably where the
00:04:31
i'm going to say debate in air quotes comes up about form factor so i'm curious by the way you've been quiet
00:04:37
what do you think about tesla unveiling a human-shaped robot and saying by the
00:04:42
way there's probably going to be a prototype late next year how does that make you feel i mean you just basically asked the question that this whole
00:04:48
episode's going to be about yeah i have a lot of thoughts on it um i i've tried to to turn this episode
00:04:55
into it is going to be a rant ramble episode um i hope you're already it is only going to be about teslabot today
00:05:01
because we have a lot of thoughts on it um it's hard to kind of break it all down because there's so much that they
00:05:08
announced by not announcing very much yeah um like it's i think the first thing i'm going to say is the
00:05:14
announcement seems extremely optimistic and i am very pessimistic about what they said i
00:05:21
think um but we're going to try and go through both parts of here fairly as fairly as we can and again
00:05:27
we're not experts um this is a lot of just our thoughts as tech people but we have gotten to like you said you did the
00:05:33
humanoid robot episode we went and saw boston dynamics and spot and they were that whole
00:05:38
day was fascinating they were talking to all of the people at boston dynamics is cool they've done all this atlas stuff
00:05:43
recently i hope we get to go back at some point and see that um but let's just start with let's do the
00:05:49
specs of what yeah human or uh tesla bot is yeah so okay there it's again a very
00:05:55
preliminary thing they've given us this like silhouette of like a sleek human form factor or whatever maybe it looks like that maybe not but they're saying
00:06:01
it'll be about five foot eight 125 pounds so actually pretty lightweight considering all the motors and batteries
00:06:07
and things you need to have in a robot it seems super light it seems like it'd have to be some very lightweight materials to make a computer like that
00:06:14
that big and that powerful 125 pounds but that's about a human dimension 5 8 125 sure
00:06:21
uh it's all sleek white and it's got this like black mesh as a head which
00:06:26
kind of reminds me of like a like a home pod like as if it had like microphones and speakers at the top but they're
00:06:32
hidden yeah i mean i think the best way to describe it is it looks like a person in a white you know those white spandex
00:06:38
body suits kind of like the green screen suits and then it has almost like the daft punk helmet it's like full black
00:06:43
directly over its face kind of down to your shoulders i mean it looks so much like it that they had this really cringy
00:06:49
moment at one point where they brought this like guy dressed like it in a white bodysuit and a black head out there's a
00:06:55
lot more than it's around it was a long time it was way too long to the point where like even elon was kind of elon
00:07:00
the king of terrible memes on twitter was like cringing about it he's like all right all right we're done uh no but
00:07:06
it's a it's definitely a human shape it has 10 fingers it has a 5 mile an hour top speed
00:07:14
to which elon commented you should be able to run away from it to which my response was how long can it
00:07:19
go five miles per hour because five miles an hour on a treadmill is pretty slow but five miles an hour chasing you
00:07:26
for hours with unlimited endurance is is a different story but he makes the point
00:07:31
that you should be able to easily overpower it or run away from it it's a friendly robot
00:07:36
great we'll just go with that sure uh it can lift 10 pounds with his arms extended or it can deadlift 150 pounds
00:07:42
so it's got some lifting power for boring so essentially it's humid arms aren't very strong but it's grip
00:07:47
strength and leg strength are pretty good sure yeah if you can deadlift 150 at 125 pounds not bad that's really soft
00:07:55
uh it's got the full self-driving computer inside it's got a screen up at the front and it is it is just the
00:08:02
closest thing to a humanoid robot that anyone's announced that they'll be able to do now when they announce this it's funny
00:08:08
you see this like cool sleek looking robot and who knows if it actually looks like that but when i saw this two different things came to mind
00:08:15
one is boston dynamics atlas like that version of a humanoid robot which is like this very industrial
00:08:22
incredible looking but very powerful and impressive super stocky it's only five feet tall but it is yeah thick it has
00:08:28
like a huge backpack on it's like it has no fingers it's just like
00:08:34
balls like fists for hand i guess sorry yeah it's kind of like rubber yeah so i pictured that but then on the other side
00:08:39
of the spectrum i picture like those sort of uh dinky like ces you know how you go to ces and you see like a
00:08:45
humanoid robot that like its only job is like to hand you like yeah like
00:08:50
brochures or something yeah it's like like an hors d'oeuvres robot or like yeah there's one that remember there was
00:08:55
one that like pulled out chairs or like moved chairs around something like scrabble or something like that yeah
00:09:00
they're all just super clunky and the only reason they look like a human is because they put like eyes and
00:09:06
a face on it exactly and that that is the interesting part to me is because robots like that
00:09:13
become unnecessarily complicated and difficult when they are human shaped
00:09:19
now if i'm an optimist i'm thinking okay an engineering company a robotics company a software company like tesla
00:09:26
should be able to make a really good general-purpose human-shaped robot but as of right now there's way more
00:09:32
examples of the dinky ces robot that can barely stand up that can push in chairs it's like why don't you just put on
00:09:38
wheels versus like boston dynamics atlas which is about as advanced and capable
00:09:43
as i've seen any humanoid robot so far now here's the uh the thing that can
00:09:48
separate tesla just before we get into the form factor discussion because i know that's the that's the conversation we probably want to have yeah but the
00:09:54
the advantage to something like uh what tesla's doing is in their ai and machine learning and training it yes so um right
00:10:02
now tesla's cars are all mostly uh capable of self-driving and
00:10:08
the way they do this is they have cameras and sensors around the car and they're taking in all this stimuli from
00:10:13
their environment and it recognizes a lot of it because it's been trained and it has this database of information and
00:10:20
what it can do is navigate through that environment so it sees the road it sees the lines on the road it knows where to
00:10:26
be but it also sees pedestrians and motorcycles and trucks and cones and
00:10:32
traffic lights and stop signs and it uses all this information and makes thousands of real-time decisions on the
00:10:38
fly and is able to navigate through its environment all that decision making and its results
00:10:44
are uploaded to the point where it that information can be used to train the next sets of teslas that go through the
00:10:50
same intersection yeah so tesla making a humanoid robot with a
00:10:56
general purpose uh application in theory we'll be able to
00:11:02
navigate and train itself and continue to be trained
00:11:08
the more of them that are out in the world so if this robot's going through home depot
00:11:13
the first time it might get it kind of right but once it uploads that information it's going to be able to learn from that and train it and that's
00:11:20
going to come back down to the next robot that goes to home depot and it's going to be better at that same task that's the that's the general idea to
00:11:26
simplify yeah and i think tesla you know tesla's ai is clearly fantastic when it comes to
00:11:32
all the road stuff now the biggest advantage to that is the way it can gather all the
00:11:38
information on the roads is because a car is a thing that a person controls to start so they sold these cars that
00:11:44
people are controlling while starting to gather the information on that which then they can continue to build on
00:11:51
build on build on while people are driving the cars and then as it gets further and further and further it's
00:11:56
collecting more and more information yeah roads are also like there are tons of variables on roads but generally a
00:12:02
road is lined it has very clear markers for things like ideas that you have to follow anyways because that's just how a
00:12:08
road works so now it has these very specific set of objects that it is going to react off of yeah there's like rules
00:12:14
exactly there's laws and there's there's like rules where if you tell something to go from point a to point b your task
00:12:20
is go from this place to this place and your set of rules is your guidelines exactly it literally has guidelines on
00:12:27
how to make it to there yeah so you know obviously the self-driving is one version of applied robotics and computer
00:12:34
vision but i could see it you can now say tesla is a robotics company and this ai
00:12:42
humane robot thing is understandable as an announcement from tesla the question i've had is
00:12:49
is it a good idea to make a humanoid robot and are they are they so optimistic that they can make an
00:12:55
efficient useful robot in a human form to the point where that's a better idea
00:13:01
than what i think is actually a better idea which is application specific robots
00:13:06
dedicated yeah dedicated for each task and that's my that's been my toughest
00:13:11
thing i've seen a lot of really interesting comments actually in the comments section of the videos and on twitter obviously i mentioned like maybe i'm
00:13:17
being short-sighted here but i'm open to hearing different versions of people's feedback and what they think it would be
00:13:22
useful for grocery shopping was the number one uh well that's a good example
00:13:28
it was the example given on stage you want to unpack that first and then maybe later we'll go we'll take a couple other like uh yeah
00:13:34
suggestions yes so uh it's an example of a task that is best achieved by the human form yep so the couple examples i
00:13:40
gave in the video were um vacuuming for example like a boring repetitive task yep
00:13:47
instead of making a human-shaped robot to push around a vacuum we turned the vacuum into a robot right
00:13:55
we have the roomba same thing with the dishwasher i just you know obviously you can have humans put things into
00:14:01
dishwashers but instead of making a human-shaped dishwasher
00:14:07
it's just the dishwasher is the robot and so the more of these examples of tasks i
00:14:14
would want to be taken over by robots that i could think of the more i thought that thing should just be the robot and
00:14:21
that's actually a better more efficient way of doing it because a human-shaped dishwasher is constrained by the limits
00:14:28
of the human form versus just making a machine specifically dedicated to cleaning
00:14:34
dishes and that's all it does it'll do it better than any machine would ever do cleaning dishes so
00:14:39
go to the grocery store is an example where we're like all right we have this whole world around us that is like
00:14:45
designed by humans for humans we've got doors and their handles are at like arm
00:14:51
height and we've got like parking lots and we've got ramps and stairs shopping carts shopping carts are like this
00:14:57
perfect application it's just built for a human to use and then that's it so maybe going to the grocery store and
00:15:03
going shopping for food for you is an example where okay you want humanoid robot to be able
00:15:09
to do that and maybe you can do some other things pretty well but like that's one thing it does well and the more i thought about that the
00:15:15
more i'm like maybe you would just get a dedicated like set
00:15:21
of robots to achieve that for you before a general purpose robot and obviously
00:15:26
that's me being pessimistic about a general-purpose humanoid-shaped robot the way it has to balance on two feet
00:15:32
and walk around and like be waterproof in case it rains and like all these other form factors it has to walk into
00:15:38
the store identify the things you want how many times you go to the store and buy something you didn't even realize you were going to get just based on your
00:15:44
own instinct like that has to happen with the robot there's a lot of other factors but in theory yeah it could walk
00:15:49
right in push a shopping cart around it's designed for the human form grab the things you want put them in
00:15:55
check out drive home whatever it's got to do and it could achieve that task in theory
00:16:01
in theory yeah pretty cool um i don't know do you picture humanoid robots amongst people at shoprite no um
00:16:09
i think like i mean so my initial thought on that was yes i believe there's a way to automate
00:16:14
that you know with a dedicated specific grocery shopping robot i think it makes way more sense to have something that
00:16:20
was on wheels that could go pretty much shelf to shelf grab whatever product is on the shelf i can see a robot having an
00:16:26
issue with i mean it's 5a it's not the tallest robot in the world i mean that's pretty average height but anything
00:16:33
within the top two shelves that might be pushed back i can't imagine it reaching back there yep you've got a ton of other
00:16:39
issues in the sense of if something's out of stock what does it do does it just continue to look for it the whole time or like gotta
00:16:47
figure out what a human would do yeah so my first thought was we have amazon warehouses that are extremely automated
00:16:53
at this point that know whether things are in stock beforehand that know where to go pick stuff up and bring it to a
00:16:58
central area where it gets boxed together and then after thinking about that someone had mentioned this tom scott video which
00:17:05
again how many times can we reference this guy he makes the best videos in the world they're so informative uh
00:17:11
if you ever want to be on the podcast we'd love you but he has this really really cool video about this completely automated grocery
00:17:17
store it's essentially this giant warehouse and if you can just imagine the center floor being like a chess
00:17:23
board just squares this huge grid yeah and essentially all of the groceries are underneath the
00:17:29
grid and there's specific squares and these robots go on all the tracks and are it looks
00:17:34
like this giant beehive almost so now they're going to all the different um i guess you aisles i can't
00:17:42
i don't know what the pods maybe yeah there's like pods yeah yeah goes above it reaches down grabs how
00:17:47
many items it needs out of it brings it up into the robot and now it has everything and can bring it to a
00:17:52
centralized spot to then package all of it i think they said by the time i order you order it takes about five hours but
00:18:00
i mean maybe that's kind of long if you need groceries right away but that seems pretty efficient if you're booking stuff
00:18:06
in advance and everything yeah it just seems to make way way more sense to have a dedicated bot doing all that than like
00:18:13
you said if a humanoid robot does that it has to drive itself there maybe it gets into a car that automatically drives but we're
00:18:19
not at that point either yet well yeah that was my that was sorry to interrupt that was number one thing is like
00:18:24
tesla is already one of the companies that has made application specific robots if the goal is to have a robot
00:18:32
drive you around i didn't build a human-shaped robot to sit in the car that is clearly built for humans and
00:18:38
hold the steering wheel and drive no the car is the robot you get in the car whatever you want to put in the car
00:18:45
the car itself is what's going to drive you from pointing to zombie it has the sensors it has the capabilities and that
00:18:50
is the robot so that's why i was thinking like yeah that's way better than having a human like try to fold
00:18:56
itself into every different type of shaped car that exists in the world yes they are generally shaped the same and
00:19:01
they're shaped for humans but yeah that that's what really got me on the question of like all right
00:19:06
how useful is a general human-shaped robot and uh
00:19:11
there's also like the emotional aspect of it which someone else so asked which was like okay you've made it
00:19:17
human-shaped is this to elicit some sort of compassion or make people less scared of it you know if you were in shoprite
00:19:24
and you saw a human-shaped robot looking at cans of peaches versus if you saw like
00:19:29
a spot grabbing peaches like would that be better claw that yeah like the clock arm
00:19:35
claw yeah yeah or just something something eight feet tall so i can reach the top shelf like would that be too much for people would they be more
00:19:41
accepting of a human robot i think that goes and that's an argument i haven't even really thought of i i will be
00:19:48
optimistic in the aspect that it is created as a humanoid because of the world we live in is based on human
00:19:54
ergonomics and i i totally like i'll i'll stay with there the emotional aspect i don't think that's i'm ready to
00:20:00
unpack something like that okay i think it has to be way more real to me to like
00:20:05
elicit enough of a response there i don't think we have a humanoid robot that's real enough for me to wonder
00:20:12
like that thing is a robot but looks like me yeah how do i feel about this well
00:20:18
here's another one how do you feel when you see uh i don't you probably haven't seen this in person we saw spot when we got
00:20:23
to boston dynamics and we saw videos of spot like walking around new york city
00:20:28
if you saw spot walking around new york city is that too much is that weird
00:20:35
i want to use a different example than new york city because it was used by the nypd and i don't i don't appreciate
00:20:40
really um so what if cause it's the whole point is it's supposed to be doing tasks that are like
00:20:46
replacing either humans or other people who would do that labor so i'm trying to think of like another robot so it was
00:20:51
interesting because most of the spot things were were in situations where humans wouldn't really be i mean maybe
00:20:57
the construction site ones where it's just going around and taking uh the same service every day surveying um but a lot
00:21:04
of it a lot of the applications have to do with remote work where it could be controlled from hundreds of miles away
00:21:10
or dangerous work where humans wouldn't be something like nuclear power plant that maybe had a leak somewhere or i
00:21:16
don't know if they have leaks but you know what i mean like something where a human would be in danger at i think the
00:21:22
remote thing was a nice bonus because it was like all right there's a construction site going up and it's in
00:21:27
some remote city you're managing the site instead of driving up there once every week and taking pictures of the
00:21:33
same stuff to see how progress has gone you just sit at home with the controller
00:21:38
and tell the robot to do it again and you can watch it go around it's like it's remote just because you can you can
00:21:44
just have the robot there and you never have to send the robot home so yeah there's just uh there's a there's a
00:21:49
lot of advantages to this sort of robots replacing typical labor stuff it's just a question
00:21:55
of what's the best form factor for it i want to take a quick break we're going to come back and talk about the human
00:22:01
shape because i think this is where it gets the most interesting and the most juicy and we can have the most uh fun discussion so we'll be right back this
00:22:08
episode of waveform is brought to you by netsuite so slow is just right if you're on vacation or a sloth or describing
00:22:14
quickbooks more like slow books and you know me i despise slow software so you can have the most bleeding edge hardware
00:22:20
and if the software doesn't run well it's basically pointless but now is the time to make the switch to netsuite by
00:22:26
oracle the number one financial system because netsuite gives you visibility and control of your financials inventory
00:22:32
hr e-commerce and more it's everything you need to grow all in one place with netsuite you can automate your processes
00:22:38
and close your books in no time no matter how big your business grows and you never know how big your business
00:22:43
might get so in fact we're hiring again at the mkbhd studios and i always make sure my important software doesn't lag
00:22:50
behind 93 percent of surveyed businesses increase their visibility and control since switching to netsuite so special
00:22:57
financing is back netsuite is offering a one-of-a-kind financing program only for those ready to switch today so head to
00:23:02
netsuite.com waveform right now that's special financing at n-e-t
00:23:09
s-u-i-t-e dot com slash waveform netsuite.com waveform all
00:23:14
right we're back so let's talk let's debate
00:23:20
human-shaped robots versus non-human-shaped robots so this is the number one comment i i i was trying to
00:23:26
like figure out from people because my my short-sightedness is is always coming to bite me i i can't
00:23:34
for one i can only see smartphones as like the end form factor when they started folding in half i was like mind
00:23:40
blown like i can i can't see them being injected into our arms or our brains i just can't i i'm so short-sighted so i
00:23:46
only see the now so the other half of that is robots like
00:23:52
i get that we want to make human shaped robots and they seem really cool but like they don't seem ready anytime soon no
00:23:59
not at all and so when i know that we have really really good application specific robots and they're easier to
00:24:05
design and create and code than a general purpose amazing humanoid robot i think we should do everything that way
00:24:12
that's just my my intuition because of my lack of robotics degree and my short-sightedness but a lot of
00:24:18
interesting comments going the other way so i wanted to talk about that um i actually want to start there was a pretty solid
00:24:24
video that was tweeted at me of someone breaking down like a couple examples of why we should do humanoid robots so i'm
00:24:30
gonna toss it to adam you can play that video hi guys i want to do a very quick video
00:24:36
i've just seen um marcus brownlee's video where he's you know basically saying he doesn't agree with the
00:24:42
approach of making the the tesla bot humanoid now first of all marcus is fantastic one of the greatest tech
00:24:48
communicators on the planet so this isn't an anti-marcus video but i do totally disagree with him on this i'm on
00:24:55
elon's side on this one surprise surprise so marcus's uh argument was
00:25:01
that you build application specific robots okay um and and that could be true if you're building from the ground
00:25:07
up if you're building a car factory for example absolutely because everything it's a flat surface you've got an empty
00:25:12
warehouse and you need to fill it with things but we are talking about um integrating a robot as possibly a human
00:25:18
assistant into a world that has been designed over thousands of years to accommodate the
00:25:24
human form okay so the example he gave was he said the dishwasher is the robot that's been designed well that's only
00:25:31
half true because the dishwasher still needs a human to interact with it okay it'll be very difficult you'd have to
00:25:37
design some kind of machine that could take the dishes out off the side or whatever put them in the dishwasher in
00:25:43
the right place so clearly most practically the space between your counter and your dishwasher and all these things it would need to have the
00:25:49
dimensions of a human because your kitchen was designed for human um same with the washing machine this is
00:25:56
first principle stuff elon musk has said what's the best way to solve the most problems in one go it's not to build an
00:26:03
individual robot for every single tech application you have in your home to replace it so you you have to replace
00:26:09
every single piece of tech or what you do is you work with what we've got and you build a robot in the form that is
00:26:15
most compatible with your home because if you had wheels on the bottom of it it could go up and down stairs for example
00:26:21
and yes it does have to fit in a car which means it has to be able to bend like a human to fit in a car so the
00:26:27
scenario may be that the robot hops into your tesla one day the tesla a robot itself drives a robot to supermarket the
00:26:33
robot has to be able to get in and out of the car otherwise you have to redesign cars so this is the point and this is why i disagree and this is why i
00:26:40
think the humanoid approach is correct thanks guys all right so i love that i love that
00:26:46
we're actually getting constructive feedback this is the best part of like making youtube videos is you immediately get like a breadth of different opinions
00:26:52
and people who immediately have a thought and can give an opinion like that okay so i wanted to talk about
00:26:57
these points for example just real quick um the the humanoid
00:27:03
form thing is so interesting because he mentions like a dishwasher while it is a robot that washes the dishes it still
00:27:09
needs a human to like load the dishwasher and i thought i was like okay that makes a lot of sense but what would be easier
00:27:15
designing a robot or designing a dishwasher for that matter with like an arm inside that can reach out and like
00:27:21
grab additions to put it in or designing an entire humanoid robot that can do a ton of stuff
00:27:27
including loading the dishwasher that's my question i think quick
00:27:33
if i wanted to use the wash the laundry version sure i think i could find a solution to that a little quicker the
00:27:39
dishes one is a is kind of interesting and i think what's funny is every time someone brings up the dishwasher
00:27:44
scenario is they don't know where the dishes are getting taken from because it's like well we have to put them from the sink into the dishwasher well who
00:27:51
put them in the sink well then we have to put them from the table to the dishwasher well who set the table like it just like it keeps going back and
00:27:57
back into these more mundane and mundane tasks that i guess kind of like does
00:28:02
make sense for the humanoid robot but i still not fully sold on it but for laundry let's say yeah there's people
00:28:09
are saying the clothes don't need to make it into the laundry machine i think the easiest way for that is like
00:28:15
your hamper at that point now has some sort of shoot that goes straight to the laundry to the the washing machine and
00:28:22
then that washing machine can transfer to the dryer or i mean i could totally see and they
00:28:27
might already have this a two in one washer dryer that's gonna come way before a humanoid robot can like pick my
00:28:33
clothes up off the floor i saw that at ces and at a certain point like you've that's everything right there unless you
00:28:38
want the robot to take my clothes off of me and put them in the laundry i which i will put put my foot down right there
00:28:45
somebody wants to prefer that somebody wants that so i could see maybe there are people who who have trouble to have
00:28:51
disabilities where that could be a thing there's probably an arm connected to your bed or something that could do it
00:28:56
better than that yeah um dishes yeah i agree there's there's a it's a tougher scenario but i
00:29:03
also think in all of these cases we are being extremely optimistic about
00:29:08
being able to have a humanoid robot that's controlling it i think one thing to talk about
00:29:14
while our world is set up for human sized things i still don't always necessarily know if that means
00:29:20
a exact human i think something with potentially four legs has a better balance scenario different types of arms
00:29:27
yeah fingers might not always be the best and uh the best for it i don't
00:29:32
think human always makes the most sense for all those and probably a lot of those could be done and be done better
00:29:39
with something like spot and an armor or something like that yeah lots lots of thoughts on that i mean really the
00:29:44
question might also be there's so many versions of this pokemon oh yeah which side which side do you want to attack the problem at from do you want to
00:29:50
attack it from the end task side or from like the human side but the other thing that he mentions is
00:29:56
uh let's say you want it to go grocery shopping for you it has to drive your car to the grocery store so it has to be
00:30:02
able to fit in a car so it has to be human shaped to get in the car which is true if you don't have a
00:30:07
self-driving car but if you do have a tesla the tesla drives itself so you don't need a human-shaped robot it can
00:30:14
fit in the trunk and then the car drives itself so anyway there's lots of like back and forth conversation i i would
00:30:20
like to have but obviously sure let's imagine that right there where we've already proved that there is
00:30:26
a way you can do a automated grocery store you don't even need the robot if your car can drive
00:30:32
itself if we're at that if we're assuming it's at that point where it can drive itself already it can just pull up and pop the trunk and then the automated
00:30:38
grocery store is the one loading it into your trunk exactly or you know shipping yeah scenario the grocery store
00:30:44
is the robot or the fedex trucks the robot so there's a bunch of different versions of this but i figured we'd make
00:30:49
it a little more interactive for this so what we're going to do we just actually completed is we opened up a discord
00:30:55
stage event and we had you guys who are in our discord server talk to us about this exact question is
00:31:02
the human form factor for robot a good idea and we got a lot of really interesting responses this is the back
00:31:07
and forth i was after so i'm glad we got to record it just a little asterisk this is on discord so this is discord audio
00:31:13
it's just a bunch of people talking to their computers bear with us on that but also uh it goes really well and there's
00:31:20
some really fun there's questions and answers yeah it went way further than i was ever expecting and it ended up yeah
00:31:27
it got completely different prices level yeah it's fantastic so without any further ado here's the discord stage
00:31:33
event uh jay basco18 i just invited you on stage if you want to talk about uh
00:31:39
your comment
00:31:45
yo how's it going good how are you i'm sorry if my audio sucks
00:31:50
that's no problem how's it going what's your uh what's your idea how you feeling
00:31:58
yeah i was just thinking of about what you said about that you don't see
00:32:04
the robot going to pick up groceries any soon but i was thinking that maybe you know for that old generation our
00:32:12
grandparents or something maybe it could be useful for them to maybe
00:32:17
you know not having them to get up to pick up the mail or you know
00:32:24
simple stuff that you know sometimes it's not as simple
00:32:30
for this old vlog you know yeah yeah so i like i think one of the one of
00:32:37
the interesting versions of this that i heard was uh like going to the grocery store or going
00:32:43
to pick up the mail like you said and i do think almost all of those tasks
00:32:48
could be accomplished with not quite a human shaped robot but probably more of like a small like box
00:32:55
on wheels with an arm attached to it basically and the question is do we want it to be human-shaped because
00:33:02
it's like friendlier or people will be more accepting of it like are we willing to put in more engineering money and
00:33:08
dollars to make it balance on two legs basically because i think you could get away with grocery shopping with just
00:33:14
like a wally robot right
00:33:20
not even a robot but maybe a person that you can hire you know you don't need that ie technology to do those
00:33:28
those tasks but yeah i don't i don't i don't see why it has to be human state at all
00:33:35
yeah yeah i mean even i think the the argument for a robot would be it's cheaper like you buy the robot once
00:33:40
instead of having to pay a salary you know so i get that people want to automate that or give it to a
00:33:46
robot uh i'm just trying to find a good reason for it to be human-shaped
00:33:52
right even though maybe we don't need them now i guess someone had to do it and
00:33:59
who better than you know elon musk to be the first with the idea you know
00:34:04
that's fair well we'll look forward to it we'll see how it goes thanks for the uh thank you guys appreciate it yeah
00:34:11
see you all right hr seven oh cool what's up dude yeah hey can you hear me
00:34:17
yeah what's going on uh hey i'm fine uh yeah so uh i was
00:34:23
thinking like the entire world like we have an entire world that's created for the human shape like all of
00:34:30
the facilities everything is like uh it's designed for the human aesthetic so if you want robots that can
00:34:37
do anything and everything like we would have to build a world around it so rather if the robot is already in a
00:34:44
humanoid shape we already have so it might be easier for a like
00:34:50
for a wider purpose rather than one specific use case yeah so i keep going back back to like
00:34:57
trying to find good examples right so one example we kept going back to was a grocery store example like
00:35:03
doors are at the perfect height for a human to pull the handle and like a shopping cart is the perfect size thing
00:35:09
for a human to push around on wheels and it's a valid point like do you need to
00:35:15
redesign the whole grocery store or do you design a humanoid robot to fit into
00:35:20
that human-shaped task and i actually think maybe it's a hot
00:35:26
take but i think we'd be it would be easier to make
00:35:31
a robotic grocery store instead of designing an incredible human-shaped robot is that
00:35:37
crazy uh yeah but like okay like for example you mentioned that we don't have a human
00:35:44
robot that does the vacuum you have a robot a vacuum that's hit its surface rover so but suppose you want a rover that's a
00:35:51
vacuum at a dishwasher and can do your clothes and can tidy up and like a bunch of stuff
00:35:58
so rather than have different robots i can do all of that one single humanoid robot may be able to do that since the
00:36:05
ecosystem is already there yeah i actually think it was it would be
00:36:10
totally fine to have a bunch of robots i think it would be actually probably a little more efficient economically and
00:36:17
physically to have a dedicated like it sounds crazy like wow you now you'd have like 100 robots in your house and that's
00:36:23
crazy to have a hundred robots for 100 tasks but we're already kind of there like you have a dishwasher you have a
00:36:29
you have a washer you have a dryer maybe one day that's one robot but like that's two robots right now you might have a
00:36:35
vacuum robot you might have a microwaving robot you might have you know a tv robot that transfers the
00:36:41
signal into an image like we have a bunch of computers even my thermostat right now is a robot
00:36:46
and uh you know there's probably some tasks out there that could be turned into robot tasks in the future
00:36:53
but those will also be right at that rate but like wouldn't every single app like appliance manufacturer and like any
00:36:59
anyone who makes any device would have to become a robotics company right i mean would they have to detect for
00:37:05
that yeah good question uh i think we're we're actually surprisingly close and i
00:37:12
think we're closer to samsung making a robotic like washer that can actually take the
00:37:18
clothes out of your hands and do the whole job for you than we are to you know
00:37:23
boston dynamics atlas like putting clothes in the dishwasher in the washer
00:37:29
but yeah you know we'll keep an eye on it but that's a that's a good that's a very common comment i saw on like uh on
00:37:36
the youtube comments on twitter like do we really want 500 robots in our houses and i don't know we might be closer than we
00:37:42
think but thanks for a question thank you take care i think there's an interesting
00:37:47
aspect in there of like jack of all trades master of none um yes a single humanoid robot can probably
00:37:54
do a whole mess of things but i do agree with you that like a bunch of tiny individual
00:37:59
i mean we're calling them robots but like we said our house right now has stuff like that it has a dishwasher has a washer and dryer a thermostat like
00:38:05
nobody's downstairs like i said shoveling coal or you know distributing oil and making sure all the temperatures
00:38:11
are correct it's kind of like that already i just don't think we see it because we don't think of it as a specific robot and then
00:38:18
a single human robot trying to do all these different things to me feels like that's a
00:38:25
mighty mighty task to accomplish when we have so many things that do all of those already
00:38:30
yeah tom scott had a really great point he made a video about actually a grocery
00:38:36
store robot a robotic grocery store and it was super interesting but actually one of the best points in that video
00:38:41
besides the fact that oh my god you can make a grocery store robot is the fact that humans don't really see
00:38:48
uh robots until they're roughly human sized and so you might see you know a bunch of
00:38:54
little robot shopping carts rolling around and think wow that's like 150 robots but really the entire grocery
00:39:01
store is one robot with 150 different little arms inside doing work
00:39:06
and so when you think about like uh do i really want a hundred different robots in my house doing a whole bunch of
00:39:11
different things another way of thinking about that or at least just framing that is your house is one robot that is doing
00:39:19
a bunch of things for you already your house you know you think about a thermostat like instead of
00:39:25
a a bunch of different objects coming together and like pressing the buttons on a thermostat to change the
00:39:30
temperature you can have the temperature automatically be changed by a robotic thermostat by doing air quotes
00:39:37
inside the same house where it you know washes the dishes for you and dries them and dries your
00:39:42
clothes and like there's all these things that happen inside the house already so maybe it's uh maybe it's just one big
00:39:48
robotic house instead of a humanoid robot walking around inside it smart house if anyone remembers the old
00:39:54
disney channel original movie that uh this is goals right there cannon update
00:40:00
would you i'm gonna invite you to speak if you want to come on cannot update i can't believe i said that incorrectly it was a very simple
00:40:07
word hey what's up uh pretty good market so my question was
00:40:13
like recently we almost lost the hubble telescope because
00:40:18
uh we did not have we don't have a human program like a space shuttle program that can
00:40:25
take us to any satellite so hubble was designed to be
00:40:31
maintained and upgraded and we don't have the replacement quite up so could these uh robots
00:40:38
something like uh spacex uh
00:40:44
tesla robot be a good option because you could technically
00:40:49
there won't be enough latency so you could technically control it by people on the ground and it would have similar
00:40:55
joints as humans so in that kind of a situation would it
00:41:00
be a good option to have a humanoid essentially instead of spot
00:41:06
doing it that's a really good question i'm going to ask you a follow-up question because i'm not as familiar with this but when
00:41:11
you say we almost lost this the hubble uh telescope is it because there were
00:41:17
things that had to be performed in person by a human
00:41:22
on the telescope so what essentially happened was
00:41:28
a hubble telescope has redundancies in it and the power unit in there is the
00:41:33
one that is active now is the one that is the redundant one and it is fairly common for the power
00:41:40
supply and stuff to fail on there because nothing is designed like most of the things are not designed
00:41:46
to stay in space forever so now we are on the last or the
00:41:53
last fail safe in on the hubble telescope so if somebody were to ship a
00:41:58
power supply up it would have to be linked to the existing circuitry and
00:42:04
then that would become the redundant or the fail-safe for the one that is already in use
00:42:10
if that makes sense no it does that makes a lot of sense and i think that's like the
00:42:15
i guess the general question would be okay now that we know we have this task that we need to complete
00:42:21
um and from what i understand the hubble space telescope is pretty old i mean it's it's been out there for a while but
00:42:28
would it be a better use of resources to modify and improve the telescope itself
00:42:36
or to design a functioning humanoid robot to go up there and continue to maintain
00:42:43
its current design and that's that's not a question i have an answer
00:42:49
to just i'm not as familiar with obviously the telescope i know it's it's been out there for a long time and it's
00:42:55
probably our most important most famous telescope but i wonder like uh you know the the
00:43:01
general question coming down to like do you want to redesign the old thing
00:43:06
to work better in a more robotic way or do you want to design a new robot to
00:43:12
continue using the old thing built to be used by humans so hubble was like a more recent c bias
00:43:20
kind of an example however like once the james webb telescope gets active that is
00:43:25
the telescope that would not be in earth orbit it would be in a lagrange point so
00:43:30
fairways away from the earth and for that you won't be able to just like
00:43:36
and if you want that telescope to be active for like 15 20 years so you need
00:43:41
some with somebody or something being willing to go to the lagrange point
00:43:48
maintain it and come back or not come back kind of a thing so do you think though that the humanoid
00:43:54
form factor is the best to accomplish something like that or we see things like spot which i mean
00:44:00
i don't know the exacts of robotics but it seems like the the two-legged arms and hands form factor right now
00:44:07
generally has like a higher center of gravity and can't lift as many things and like spot seems to be something that
00:44:13
in general use kind of like one of its big um things they sell for is remote work so
00:44:19
somebody far away can accomplish stuff with the cameras and arms on it so it's something
00:44:24
now space adds a whole other aspect to this where maneuvering around space is obviously very different but
00:44:30
would there be something you think non-humanoid that would be more efficient in space being able to you
00:44:36
know maybe maybe almost like a spider robot that can grab onto the walls and everything and still have two arms that
00:44:42
can be used remote to like you said replace this uh power supply and the
00:44:47
wires stuff like that or does a two-legged finger humanoid robot make the most sense here
00:44:53
so the reason why i was thinking that the humanoid would be more easier in the early run would be if
00:45:00
say there is less latency and you can have some an electrician uh on planet earth trying to control it
00:45:08
and then you have one to one finger feedback for initial time being till you don't have smart robots that is that can
00:45:15
do things on its own so that is nearly important right do you think that
00:45:21
sorry sorry i interrupted go ahead quick um so it seems like the the finger
00:45:26
aspect is what you're most interested here and i i totally agree like if it's like you said one to one finger i
00:45:31
believe they have surgery robots that there's something very similar with that where you're essentially putting your hands into
00:45:37
uh yeah like a glove and it's reacting but i mean you can still put fingers and
00:45:43
hands onto a robot with four legs or six legs or maybe something that has a better center of gravity can grip onto
00:45:50
different things with multiple other legs better and is would still be considered designed specifically
00:45:55
for that yes so yeah that is essentially where that was the only part where i said that
00:46:00
this would be useful because mostly we've designed like spot as i gave the example spot would be useful in the
00:46:06
longer run once it learns or anything so it was just like in this quick case
00:46:12
scenario for the next five years that seems like could be an option
00:46:17
but i do agree that a spot would be a better option because it is a more refined concept of a
00:46:24
moving structure yeah yeah i think and even in the future
00:46:29
you know as we continue to make new telescopes that go further from earth and are more less likely to want to send
00:46:34
a human up to repair they should probably be designed to be
00:46:39
remote repaired and not repaired by humans which is like a whole another story i'm not about to
00:46:45
sit out here in an armchair designed telescope but that's like uh that's like another
00:46:51
uh general thought is like okay we know these things are built to be operated by
00:46:56
humans and then humans go up and repair them but what if uh what if we just make them so we never have to
00:47:01
do that again would that be better yeah that would be better yeah so that is essentially what i was thinking about
00:47:08
in the future all right real quick let's take a quick break we'll come back and we'll talk more about some robots
00:47:14
this episode of the waveform podcast is brought to you by blackvue have you ever been in a situation where a dashcam
00:47:20
might come in handy i kind of wish i had one that truck sideswiped me a few years ago but before it's too late get the day
00:47:25
and night protection your vehicle deserves with the blackvue dr750x plus series cloud dash cams so whether you're
00:47:31
driving around on a busy road waiting for the traffic light to change or just parked at night blackvue dash cams have your vehicle covered the blackvue dr750x
00:47:39
plus model features an upgraded sony starvis sensor that gives you stunning clarity and vivid colors even in the
00:47:44
most challenging of situations and not only does it give you great looking footage but because of the impressive dynamic
00:47:50
range you can also be sure that you're not going to miss important information like a license plate or something like that just because of harsh lighting
00:47:56
conditions so the front camera also records at a full 1080p 60fps so you
00:48:02
won't risk losing information even at higher speeds and with blackvue cloud you can rest easy knowing you can check
00:48:07
in at your car at any time and even receive notifications on your phone so you can go to blackvue.com waveform and
00:48:13
use promo code waveform to get 10 off any blackvue dashcam free shipping on orders over 200 bucks so that's b l a c
00:48:20
k v u e dot com slash waveform promo code waveform
00:48:26
uh noah i invited you up if you'd like to join hey is that my level good
00:48:32
fantastic great audio quest mike oh thank you it's a blue spark digital
00:48:38
um nice yeah so for hot takes i mean one of the hottest takes for robotics is
00:48:45
you know being able to replace jobs and tesla's doing things like with the tesla semi right it's being able to
00:48:51
replace a driver potentially with its whole ai self-driving and that driving
00:48:56
trucks is one of the largest markets in the usa and uh so if you think about there's
00:49:01
also a lot of jobs that require not too much thinking tasks but a lot of tasks with like using your hands and walking
00:49:08
around like say serving tables at a restaurant or like helping out the disabled people in the hospital right
00:49:15
so for any of those tasks which are quite a few and there's just anyone that doesn't require a lot
00:49:21
of thinking but a lot of you know just general interacting with objects i think
00:49:26
a bot that's designed like a human would be able to fit that role very well since
00:49:32
you know the world is designed for humans it's not designed for a dog yeah so i like this this is interesting
00:49:39
so the one is i kind of like the like waiting tables uh example
00:49:45
where let's say because in theory like all the robot has to do if it's waiting tables is it needed to collect your
00:49:51
orders it needs to carry the plates out and back and clean the tables and stuff like that
00:49:57
and it could be shaped like a human but it could also be shaped like like wall-e or like a box on wheels with
00:50:03
three arms and it could probably accomplish the same thing and i wonder is it human-shaped because
00:50:09
that makes it more easy to understand or it's friendlier for people if you're in a restaurant and you sit down and
00:50:15
there's a human-shaped robot you know that it's taking your order versus if it's a box you don't know what's happening
00:50:21
is that like the main advantage i mean that is one part of it because
00:50:26
we've seen you know when the tesla robot came out we saw people referencing movies like the terminator right oh
00:50:32
here's this evil robot that's gonna come kill us all but you know another famous movie is big hero 6 where you have
00:50:39
a human-shaped robot that's one of the main characters one of the big advantages is that you're able to
00:50:44
empathize with it so i think that is a big factor having a robot that people can you know empathize
00:50:51
with or even feel comfortable having it in a room i think another example would be data from
00:50:56
the old star trek movies or tv shows but uh the other part of it is just that
00:51:02
you know we are in a world designed for humans you know every object from you know the doorknob to the
00:51:08
stairs is made for a human to be able to interact with so i think having a human-shaped robot
00:51:13
while we might not be able to think of any single example having something that
00:51:18
if you wanted a robot that could adapt to any situation when it comes to just general day-to-day
00:51:24
activities then i think a human-shaped robot would be ideal for that yeah i think a key word you used was
00:51:31
adapt so i remember when i was talking to elon in in the interview we did i put the clip
00:51:37
in the video actually um where we were talking about the factory and like what are some things that robots do in the
00:51:43
factory versus like what humans are doing because there's still humans working in the factory yeah and then the example he gave was like oh you know
00:51:49
maybe just connecting two hoses together and so there might be a hose on the left and a hose on the right and then tell the human to connect them and they just
00:51:55
grab one grab the other put them together and a robot's like decent at that once
00:52:00
but then if the next time the hoses are in different places the human just knows to grab them again and the robot needs to like figure it out again
00:52:07
yeah and i wonder like what level of adaptability is possible from these like impressive
00:52:14
machine learning computer vision enabled robots that test this building and can actually be like how hard is it going to
00:52:20
be to make them as good as a real human because i love the idea but it seems like that is
00:52:26
equally as difficult as redesigning parts of the world we live in yeah and i think the solution or the the
00:52:33
answer to that question which none of us are going to be able to say for now uh relies more on the software of the
00:52:40
robot rather than hardware like tesla's building one of the largest super cute computers in the world just
00:52:46
for ai and once they solve self-driving what will they be able to train ai on could you create something that could
00:52:54
say hey google make me a cup of coffee right would we be able to do it i don't know but it will be very interesting to
00:53:00
see that in the future yeah do you think about like how many variables there are in the real rest of
00:53:05
the world versus self-driving because this is one of the things that came up in our talks was like all right driving a car there's a road there's
00:53:12
lines on the road and then there's maybe like a dozen things you're going to encounter a bunch of times like there's
00:53:19
other cars there's trucks motorcycles bikes pedestrians stop signs street
00:53:25
lights and then you kind of like you've kind of done it you solve self-driving
00:53:31
but make me a coffee you know there's there's one task you recognize the
00:53:36
coffee and there's a bunch of things there but then there's like hey grab my headphones downstairs doesn't know what headphones are does it need models of
00:53:43
that and then it's hey go grocery shopping and there's just this massive list of
00:53:48
way more stuff to learn and you know if i can do it it's probably tesla but i just i'm curious
00:53:54
that's a factor yeah and i think the thing about will be able to identify my products
00:54:01
i think that's something that we're already seeing these programs being able to solve because if you go into like if you look
00:54:08
at what ai is doing when you go into google or anything you can say give me a picture of a cat and there's ai that
00:54:13
will actually be able to give you pictures of cats that have not been taken ever you could say here's a
00:54:19
picture of a sunny beach so to say that there's an ai that could say hey get my headphones and it would be able to know what your headphones are
00:54:26
and be able to identify them i think that's something that we already have
00:54:32
but on the matter of you know how hard is it to do day-to-day activities versus how hard is it to drive a car
00:54:40
i mean i kind of wish that i knew more about ai just to answer that question because it's like
00:54:45
you know every one of us humans is able to do human things easily yet we
00:54:51
need to train and get a license to drive a car so does that mean that driving a car is more difficult than walking down
00:54:58
a set of stairs or does that just mean that we're really good at walking down sets of stairs and i'm pretty terrible
00:55:03
relatively at driving cars and i don't know but i'm really excited to see
00:55:09
what happens because you know ai wasn't really a thing that
00:55:15
normal people could access what 10 20 years ago and now we already have things that are able to generate images to find
00:55:22
objects and self-drive your car so you know 15 years from now it's going to be pretty swell
00:55:28
yeah i like the optimism i think that's the best the best way to look at it is like we've gotten this far and it's
00:55:34
funny like there's even still some cases where i'll like ask google lens what type of sneaker i'm looking at and it's
00:55:40
like that's a blender and i'm like no not not at all but yeah it is really inspiring yeah
00:55:45
it's inspiring when it gets it right so i like i like the optimism noah thank you and uh i like your videos i
00:55:52
must say thank you appreciate that thanks for coming up yeah no worries have a great day don't
00:55:58
forget to drink water
00:56:11
i think we have one one more all right i have one more somebody in here said
00:56:17
let's see i have mind-blowing thing you should hear no one has ever thought about for a humanoid robot so i think i
00:56:23
can't not invite this person up to talk so let's uh we're gonna do our last one here i think because we need to this is
00:56:29
all going to go into the podcast um and we talked about a lot of stuff before and we will talk about a lot of stuff
00:56:35
after so we can't do too too many uh but specified knoll i'm gonna invite you up
00:56:40
is these are big shoes to fill no one's ever heard about it so let's see uh hello thanks for inviting me
00:56:49
uh you can hear me is it clear um it's good thanks um so i had a really great
00:56:54
idea that i have been thinking about for a while and that is okay so you know for tesla
00:56:59
cars they replace humans to drive so you can use the car you don't have to do any input and you
00:57:06
can drive it it's like a tool for humans now humanoid robot um okay let me put it
00:57:12
this way our bodies are kind of a transport for our heads
00:57:18
basically we live in our heads so now we can uh you know transplant arms limbic
00:57:23
arms or legs we can transplant hearts but not yet entire bodies so i would say
00:57:29
this is a great experiment that if this robot can walk properly can navigate can
00:57:34
do everything that we can do and then at some point let's say we get old we get 90 years old our body is old
00:57:41
but not our brain our brain still works if we could somehow do like
00:57:47
a surgery that removes our head without us dying and put it on a robot
00:57:52
that can you know pump blood to our brain give us food
00:57:58
so we don't have to eat actually we don't have to drink so the robot does everything we just charge it and give it like supplies so
00:58:05
it supplies food and water to our brain and then we can walk we can navigate the world and we can change our bodies just
00:58:12
like the movies but like this is required just like the cars they
00:58:17
need to drive a lot around the cities and town to learn we need a robot learn how to navigate
00:58:23
the world we need a robot that can do that and then we can transplant yes this is for the future like
00:58:28
maybe the next hundred i don't know 10 20 50 years that we can actually get bodies for our brains because basically
00:58:35
our body is a trans transport for our heads like we eat to supply food to it and then we have legs
00:58:42
to walk we have hands to do things basically we only need our brain
00:58:47
man eat that that is a very valid point our our body is a transport vehicle for our
00:58:54
brain our brain is the most important organ according to the brain anyway um
00:59:00
no i'm with you okay so here's another question then so you need to somehow connect
00:59:06
the brain and all of the systems the the breathing system the
00:59:11
the the blood and the oxygen going to the brain and everything so that all you need to do is just charge it
00:59:18
um what did you forget to charge well okay so basically that's where
00:59:24
neuralink comes in place neurolink helps the brain to connect to computers so you can you know control your arms
00:59:31
your legs even you can replace your eyes have high resolution eyes you can see further you can see anything you can see
00:59:37
your emails you can you don't even need a phone they can like literally connect to the world and for the foot part
00:59:42
basically the robot needs two supplies one is charging for the batteries the other supply is like imagine like having
00:59:49
like a small sack of bowls that one of them is oxygen the other is food so you
00:59:55
just put it in and then it will have a supply of oxygen for months for your brain and you can choose what you want
01:00:01
to eat for your brain today you can like insert something that will pump that food and
01:00:07
oxygen and supplies and nutritions to the blood that is a small stream of blood it just goes to the brain and
01:00:14
recycles it and cleans so you don't actually probably need to breathe anymore i guess
01:00:19
what about uh what about a nervous system because i i'm with you on like okay attach my brain to a humanoid robot
01:00:25
what about like feeling things like i can move my hands and i can pick things up but can you still uh are there enough
01:00:32
sensors out there is the tech good enough to replace a nervous system well at the moment probably not we don't know
01:00:38
much about the brain but i'm sure elon what he's doing with the neural link if he can successfully connect the
01:00:44
entire brain interface to the body including controls and sensation feelings
01:00:50
um vision hearing everything we probably just could take the brain out not the
01:00:56
head just the brain in a safe container like in the head or it could be anywhere
01:01:02
in the human robot maybe in the chest and we can go anywhere because we are a
01:01:07
brain that is connected to you know our bodies to take us go around we
01:01:13
probably could do that okay so you've heard of the you've heard the simulation argument where like
01:01:20
if you advance technology sufficiently far enough into the future our simulations get so
01:01:26
good that they're indistinguishable from reality and we may or may not be in a simulation already
01:01:31
do you ever think about this as a uh as a future argument for this
01:01:37
like are we already humanoid robots are we already there we do do do we do
01:01:42
it already we could think about this all day there could be another world they could be we
01:01:48
could be in a simulation or not but doesn't matter what matters is the reality we are in either we are in the
01:01:53
simulation or not and there's no way probably to prove it but that's okay what matters is we look in the future
01:01:59
and what we can do for the humanity right what we can do to improve and what is
01:02:04
the most important thing is our lifetime and it's limited right people have thousands of years have been trying to
01:02:11
you know extend human life but nobody could have done that until today we have the technology it's just a
01:02:17
little bit we're you know we're almost there we just need to understand the brain and be able to connect it to our
01:02:23
bodies and at that point it's important like you said it's important that we go to the right direction we don't go to a
01:02:30
matrix and stack all the people into a computer we have to build a better world and that's where it starts today i
01:02:36
believe i agree we are so far off topic and i love it it's great
01:02:44
uh i thank you thank you for the for the thousand foot view i think that was that was a useful way to probably
01:02:50
end this i'm gonna know everything i thought it was useful yeah i'm gonna be thinking a lot about
01:02:56
whether or not i would like my brain to be connected someday to a sufficiently advanced robot with the right technology
01:03:04
yeah sure because you know our minds probably only see the current situation we see the robots as they are we need a
01:03:10
vacuum robot we need robots for grocery shopping but what we don't see is the far future the next
01:03:16
10 uh 20 at 50 years from now or a decade so that's the future we're
01:03:21
building towards and we have to be able to you know imagine or visualize what's going to be the world in hundred years
01:03:28
another another use case scenario just remember is like for as of right now before we get our olympic bodies uh
01:03:35
uh robotic bodies we can use these humanoid robots uh to learn to navigate uh you know disasters
01:03:42
earthquakes um hazards that no robot could you can't send a vacuum robot there to save people but this
01:03:49
human robot can it can drive a fire truck immediately like we need help all
01:03:54
the people are down everybody needs help so it can drive cars it can take you the car not all cars are self-driving
01:04:01
not all hospitals uh have equipment that can be controlled by computers we need those people to
01:04:08
take care of people at hospitals or people that people need care you know when we need people old elderly they
01:04:15
need care like it's not just all robots can do all their needs for them we need someone that can care for them help them
01:04:22
in case of emergency they need to go to hospital all right sure let's let's go pick up the car keys i'm gonna drive you
01:04:28
to hospital that that robot's gonna do for you gonna save your life yeah i think that the one thing i see in
01:04:35
that is like that sort of makes the humanoid robot a
01:04:40
a middle like a stop gap in between technology of the future like you say obviously hospitals today
01:04:47
aren't all robotic and cars and fire trucks and ambulances today are not self-driving so until they are
01:04:54
we should make an a sufficiently advanced humanoid robot that can fill the gap of things humans can't quite do
01:05:01
until we get to that better day where all the ambulances are self-driving and we can send like a wally shaped robot
01:05:09
into the rubble of a disaster to help save people like we need this this in between and if we get the humanoid robot
01:05:16
right quickly then it can fill a gap until tech of the future is amazing
01:05:23
sure exactly exactly and until then hopefully this robot will be able to walk and navigate the world just like
01:05:29
humans just like tesla cars can drive now they can navigate until that day
01:05:35
probably those robots will be able to do most of the things and then hopefully by then a neural link will be
01:05:41
able to link our brains to them so we can actually walk we could be at home and control a human robot remotely so we
01:05:48
could be anywhere in the world we can connect to any one of them and experience the world just as it is
01:05:54
perfect i might be internet by being controlled from home right now who knows
01:05:59
awesome hey thank you so much for uh thanks for chiming in we appreciate that i think that's true
01:06:04
i appreciate it thanks for having me have a good one you too have a good one take care all right thanks
01:06:10
all right well that was incredible uh i think that's what we're gonna probably cut it that was a nice existential way
01:06:18
to end this i want before everyone leaves this discord room i want everyone right now wherever you're sitting
01:06:24
to picture that you are just a remote controlled human-shaped robot being
01:06:31
being controlled by someone else somewhere you are ready player one i
01:06:36
just want everyone to have that that vision right now look down at your own hands is it real what's even real
01:06:42
anymore all right well that went better than expected didn't no i don't
01:06:47
know if i expected to look down in my own hands questioning reality but i ended up doing that
01:06:53
because maybe i am a humanoid robot after all um wow human after all is a song by daft daftbunk which is the
01:06:59
robots dang that is deep whoa there's no such thing as a coincidence whoa okay
01:07:05
well anyway i think just to put a bow on this the uh the last thing that made
01:07:10
headlines from tesla's ai day when they talked about this robot we did get elon saying uh there might be a prototype by
01:07:17
the end of next year i think he said that's what he said 2022 yeah um now for
01:07:23
anyone who's ever heard elon announce a date on stage we already are just gonna go ahead and assume that that's not
01:07:29
happening not at all uh but you know seeing seeing what we know about these ces robots and what we know
01:07:36
about boston dynamics atlas if you were betting when do you think you would see
01:07:42
realistically a tesla built humanoid robot that learns its way around the
01:07:47
world and is actually somewhat useful and also like doesn't fall over that's like that
01:07:53
people are using to actually accomplish tasks and it's doing things by itself like you ask it to do something and it
01:07:59
does something yeah because you know i mean even at the base level i'll be dead by the time that yeah i mean we don't know what coming out means like i i
01:08:07
assume so like that could literally be anything sure yeah it could just be a prototype
01:08:12
that have somewhere like a model the tesla roast is a prototype too like whatever but like eventually it's gonna
01:08:17
come out and does that mean it's gonna be a retail purchase for humans like you'll be able to buy one for your own house i have no idea where we're going
01:08:24
with this particular robot but when do you think we'll see a functional functional humanoid robot that does some
01:08:31
of this stuff if i'm taking so like can we agree we've said it
01:08:37
already atlas is the closest thing we've seen to you and i have seen right it's the most advanced humanoid shaped robot i've
01:08:43
seen yeah and it is not even a fraction of what
01:08:48
this tesla bot seems to want to accomplish it's not as graceful it is not really it doesn't have fingers
01:08:55
it's much shorter it's much stockier like that would not fit into a car it's much louder yeah it's it's basically
01:09:02
mostly running around itself i mean even boston dynamics i don't think is called anything other than research and
01:09:07
development like as far as i know they don't have a specific plan for it yeah it's not coming out and that is
01:09:13
you know they've been in the game way longer than tesla has in terms of the actual building of robots we all agree
01:09:19
tesla's ai and software is fantastic and like totally makes sense for something like this but if i were betting like
01:09:26
like i said i'll be dead by the time teslabot comes out and is doing
01:09:31
you know if it goes for sale for sale and people are actually buying it i mean
01:09:37
yeah i can't imagine this is something that has to fully rely on it tesla cars when they release full self-driving and
01:09:44
stuff and people are paying for it the car still works as a car in that process this robot can't be
01:09:50
used for anything unless it's working in that scenario so it has to be fully
01:09:55
fleshed out at that point i at least 50 years i think we're going to
01:10:01
still alive i think we land on mars before we get the robots working to an acceptable level but i don't think that's 50 years out i think that's close
01:10:07
so i think this is a kind of an interesting thing you mentioned about stop cap robots is we're so far out of
01:10:13
this do we have the technology for it to do to be a stopgap robot or are we further
01:10:20
along in all the other technologies to be have these more individual automations before
01:10:26
yeah the the humanoid robot is doing all of them i think that's the that's the best thing that i got out of this with all of our
01:10:32
conversations we've had about the robots and with all of our back and forth and the discord stage room and the comments on twitter and the videos i think the
01:10:39
best thing that i see from these robots is if we can make a pretty good
01:10:47
general purpose robot whether it's human shaped or not let's ignore the human shape just a pretty good general purpose
01:10:53
robot i think that robot would be pretty useful as a stop gap between
01:11:00
when we actually get everything else to be robots so we were talking about like obviously you want to be able to drive a
01:11:06
fire truck into a forest fire and be able to like put something out from the middle and you don't want to send a human in there but since forest since um
01:11:14
fire trucks aren't self-driving you need something else to drive the truck like before before that is a self-driving
01:11:21
fire truck maybe there is a humanoid robot driving it and that's like self-driving self-driving fire truck way before a
01:11:27
humanoid robot can do it because we have self-driving i mean that's true we basically have self-driving cars already
01:11:33
that's right and then for a fire truck listen that i have a lot of family members who are firefighters it's it's
01:11:38
very hard but like in general you could probably have a fire truck that has a hose attached to it that's that's controlled also by the same sensors and
01:11:45
stuff there's way more to it than that but i still just don't think the humanoid robot is coming and doing the
01:11:52
job of a firefighter or driving a fire truck yeah stuff like that i think i think the aspiration is really admirable
01:11:59
like oh yeah people picturing it as like waiting tables at a restaurant or like
01:12:04
just doing simple things for the elderly like collecting the mail and bringing them the milk from the delivery outside
01:12:11
like very simple you can't talk about humanoid robots and also still talk about getting milk delivered
01:12:17
simple stuff like different worlds i'm just saying you just in this world where we wanted to do very simple basic things
01:12:23
for sure having one robot just to go get the mail is maybe
01:12:28
less desirable than having one general friendly human-shaped robot that does a
01:12:33
bunch of these everyday tasks and so i think if that is the aspiration
01:12:38
of something like teslabot cool i hope i'm optimistic i at least i want it to
01:12:45
go well i'm not necessarily betting on it going well but i'm hopeful that we do get something like that
01:12:51
uh to actually be friendly and useful but maybe when we're old we we will revisit this
01:12:57
if it comes out and you and i are elderly and need help getting our laundry picked up and and mail taken and
01:13:02
we'll we'll come back we'll do another podcast episode if the tesla bot comes out before
01:13:09
2030 then i will i will let it slap me across the face
01:13:15
yeah same okay same i will tell it to slap me across the face to its face that's about it that's where we're gonna
01:13:22
end it this is a wild episode if you're this far kudos to you thank you for listening
01:13:27
appreciate you uh yeah that was a that was a lot it's a fun one we'll be back next week i'm sure there's a lot more tech coming out i mean it's that time of
01:13:33
year it's going to happen so uh yeah stay tuned for that see you guys in the next episode definitely subscribe to the
01:13:39
studio channel to the clips channel and to the main mkbhd channel while you're at it we'll catch you guys
01:13:44
in the next one peace waveform was produced by adam molino we are partnering with studio 71 and our intro
01:13:49
outro music was created by vain silk
01:14:01
you
