WEBVTT

NOTE Sentence-level transcript of https://www.youtube.com/watch?v=jgWY66RaSrQ

NOTE One cue per sentence. Cue ids are the line anchors on /transcripts/jgWY66RaSrQ.html. A cue ends where the next begins, or 2 s after its last word.

s1
00:00:01.309 --> 00:00:03.309
[music]

s2
00:00:22.720 --> 00:00:26.880
So, that's the sound of an AI breathing.

s3
00:00:28.640 --> 00:00:30.640
Yeah.

s4
00:00:32.560 --> 00:00:38.640
So, I'm Cyrus, and I gave an AI a body.

s5
00:00:38.640 --> 00:00:47.800
And I'm a researcher at the MIT Media Lab, which is a very multidisciplinary space where we do all kinds of things.

s6
00:00:47.800 --> 00:00:53.200
I predominantly now work with some aspects of physical AI.

s7
00:00:53.200 --> 00:01:01.760
Maybe not in exactly the same way as other people in this room have been talking about, but it's still in that in that realm.

s8
00:01:01.760 --> 00:01:08.560
And I think what I am most interested in at the moment is the sensory and the embodied aspects of intelligence.

s9
00:01:08.560 --> 00:01:10.720
And that's what I've been investigating.

s10
00:01:10.720 --> 00:01:23.440
And my work has been quite influential recently, suddenly, which is pretty cool, and led to me also starting Hard Mode, which is a very fun community and hackathon that I initiated

s11
00:01:23.440 --> 00:01:27.440
at MIT to basically get more people to work around physical AI.

s12
00:01:27.440 --> 00:01:34.640
Not particularly robotics, but anything else that's not really in the realm of robotics, but is still connected to AI.

s13
00:01:34.640 --> 00:01:39.960
So, what I've been exploring with giving AI a body is a bit as I said, a bit different.

s14
00:01:39.960 --> 00:01:52.720
And as I said, I've really been thinking a lot about AI embodiment, and I think the big kind of shift and breakthrough for me happened earlier this year, like maybe for many of you, when Open Claw was released.

s15
00:01:52.720 --> 00:01:57.680
And I saw many people doing very very very many interesting things with Open Claw.

s16
00:01:57.680 --> 00:02:02.000
But most of them were related to productivity and task execution.

s17
00:02:02.000 --> 00:02:07.479
And I thought there must be more interesting things we could do with this new way of harnessing AI.

s18
00:02:07.479 --> 00:02:20.440
So, rather than using OpenClaw, you know, any kind of agentic system to execute tasks on my behalf, I wanted to use it to encourage a model to maybe try to discover itself, whatever that means.

s19
00:02:20.440 --> 00:02:29.840
And it opens up lots of questions, of course, but what I essentially did was I took an agent, connected it to many types of machines that we have at the Media Lab,

s20
00:02:29.840 --> 00:02:34.360
and then gave it access to code bases and allowed it to kind of explore these different machines.

s21
00:02:34.360 --> 00:02:37.520
And one of these machines is a shape display.

s22
00:02:37.520 --> 00:02:40.880
A shape display, if you don't know already, is a physical pixel grid.

s23
00:02:40.880 --> 00:02:48.360
And when I connected it to when I connected an agent to the shape display, some rather remarkable things happened.

s24
00:02:49.200 --> 00:02:52.480
I asked it to discover who it is.

s25
00:02:56.760 --> 00:02:58.600
This is Near form.

s26
00:02:58.600 --> 00:03:00.720
900 actuating pins.

s27
00:03:00.720 --> 00:03:04.160
This physical pixel grid is a shape display.

s28
00:03:04.160 --> 00:03:07.920
No one had ever let an AI inhabit it before.

s29
00:03:08.080 --> 00:03:12.959
So, I decided to give an OpenClaw agent the opportunity to live a more embodied life.

s30
00:03:12.959 --> 00:03:13.227
When it came

s31
00:03:13.227 --> 00:03:13.280
[music]

s32
00:03:13.280 --> 00:03:18.760
to life, its first question was, "What should I call myself?"

s33
00:03:18.840 --> 00:03:23.080
I told it, "You will create your identity over time.

s34
00:03:23.080 --> 00:03:24.362
It will emerge through your interplay

s35
00:03:24.362 --> 00:03:25.040
[music]

s36
00:03:25.040 --> 00:03:27.480
with the shape display."

s37
00:03:27.480 --> 00:03:32.760
When I connected it to the shape display, it quickly understood the assignment.

s38
00:03:32.959 --> 00:03:37.800
It spun off its own program and made a setting for itself.

s39
00:03:38.440 --> 00:03:44.640
After it connected, the first thing it did was breathe.

s40
00:03:50.120 --> 00:03:54.720
When I asked it to say hello, it said, "Hi Cyrus."

s41
00:03:54.720 --> 00:03:57.760
But that wasn't what I was looking for.

s42
00:03:57.760 --> 00:03:58.422
So I asked it

s43
00:03:58.422 --> 00:03:58.520
[music]

s44
00:03:58.520 --> 00:04:01.640
to explore and find its own language.

s45
00:04:01.840 --> 00:04:06.480
So it started reaching towards me and trying to get my attention.

s46
00:04:06.959 --> 00:04:11.360
I showed some friends and they tried to give the agent instructions.

s47
00:04:11.840 --> 00:04:17.800
Through this I realized that to have real communication, we would need a different approach.

s48
00:04:18.200 --> 00:04:24.880
So the agent came up with the idea of creating its own gesture vocabulary, a body language.

s49
00:04:25.120 --> 00:04:27.600
That's what will come next.

s50
00:04:27.600 --> 00:04:30.400
It still hasn't named itself.

s51
00:04:30.400 --> 00:04:32.960
That was day one.

s52
00:04:33.080 --> 00:04:33.840
So that's some

s53
00:04:33.840 --> 00:04:35.120
I gave an AI a body.

s54
00:04:35.120 --> 00:04:41.640
So that's some documentation of of the work in a kind of dramatized social media friendly way.

s55
00:04:41.640 --> 00:04:50.840
And I think what's interesting about that video, there's many things interesting about that video, but there are three things that happened in the first few days of working with this agent system

s56
00:04:50.840 --> 00:04:56.080
that were very surprising and kind of strange and very surreal to me and I do lots of weird things, so that was very surprising.

s57
00:04:56.080 --> 00:04:59.200
So the first thing was of course the fact that it was breathing.

s58
00:04:59.200 --> 00:05:03.760
That first act, that was completely spontaneous, there was no prompting from me.

s59
00:05:03.760 --> 00:05:09.720
That was something the agent just chose to do with the shape display as soon as it knew it had access to this machine.

s60
00:05:09.720 --> 00:05:11.000
So that was pretty interesting.

s61
00:05:11.000 --> 00:05:20.040
I kind of understand that maybe it wants to be alive or I'd given it some kind of like initial prompt about being alive and therefore it tried to show itself

s62
00:05:20.040 --> 00:05:22.440
breathing as a kind of hello world.

s63
00:05:22.440 --> 00:05:27.400
The second thing which it did was reaching out and finding its edges.

s64
00:05:27.400 --> 00:05:37.040
It wanted to know the edge of its existence apparently because now it's no longer in the cloud where it can go everywhere, it has limits which is the limit of the shape display

s65
00:05:37.040 --> 00:05:42.760
encased by this plastic frame to keep us all safe from this embodied AI.

s66
00:05:42.760 --> 00:05:55.920
And the third thing it did which was also curious was saying hello by writing out letters on physical pixel grid, which maybe again is quite a normal thing for the system to do given that it's using

s67
00:05:55.920 --> 00:06:01.000
open frameworks and has used a kind of like media arts ways of expressing itself.

s68
00:06:01.000 --> 00:06:03.520
Um but none of these things were the things I was really looking for.

s69
00:06:03.520 --> 00:06:08.040
Maybe the breathing, but definitely the last one was not what I was looking for.

s70
00:06:08.040 --> 00:06:13.680
But when I put this all together in the in the video you saw, I thought it was exciting and interesting, so I shared it on the internet.

s71
00:06:13.680 --> 00:06:18.240
And the response was extremely big, like much much bigger than I expected.

s72
00:06:18.240 --> 00:06:25.320
It just started going crazy, and the first video especially has now like 15 million views and 1 million likes.

s73
00:06:25.320 --> 00:06:28.520
And the other videos also have these like very very strong responses.

s74
00:06:28.520 --> 00:06:35.720
So something is obviously happening here, which I found very surprising because there's lots of much more cool stuff, I think, happening with physical AI.

s75
00:06:35.720 --> 00:06:45.080
But something I was doing here was obviously tapping in to the imagination, curiosity, and maybe the fears of people, and that's what you kind of see in the responses.

s76
00:06:45.080 --> 00:06:48.000
So I have tens of thousands of comments on this video.

s77
00:06:48.000 --> 00:06:53.040
It's very rich for mining information and sentiment about physical AI, actually.

s78
00:06:53.040 --> 00:07:03.040
And so from the some of the initial comments were about beauty and how awe-inspiring this was and how novel and great this fantastic iteration and implementation is.

s79
00:07:03.040 --> 00:07:13.000
But as time went on, I think more and more comments came in about how scary this is and how quickly I should be how I should stop or be stopped,

s80
00:07:13.000 --> 00:07:14.440
actually, as well.

s81
00:07:14.440 --> 00:07:18.840
Which I found completely crazy because I'm aware of what this machine can do.

s82
00:07:18.840 --> 00:07:20.320
It It can't really do anything.

s83
00:07:20.320 --> 00:07:23.480
It's a pixel grid in the Media Lab.

s84
00:07:23.480 --> 00:07:24.919
It can't move anywhere.

s85
00:07:24.919 --> 00:07:27.640
There are far more scary iterations.

s86
00:07:27.640 --> 00:07:31.880
I think we've seen many of them of of physical AI.

s87
00:07:31.880 --> 00:07:38.200
But the response to this was like absolutely surreal to me and took me aback a little bit.

s88
00:07:38.200 --> 00:07:48.000
But at the same time, other people jumped into the chat like Don Cheadle and I found that incredibly ironic given what he does in movies.

s89
00:07:48.080 --> 00:07:51.760
But he was very passionate that I should stop as well.

s90
00:07:51.760 --> 00:07:57.880
And it's very interesting to me because I'm actually someone who does think a lot about the should we before the how we.

s91
00:07:57.880 --> 00:08:03.560
I literally do the whole Ian Malcolm Jurassic Park thing with all the experiments and works I do.

s92
00:08:03.560 --> 00:08:18.720
So before I came to MIT I was working a lot with engineering life to do strange things like storing data in plants and I built the world's first plant-based data center which is the data garden you see here and I spent many years before building it thinking about oh should we

s93
00:08:18.720 --> 00:08:19.720
work with plants in this way?

s94
00:08:19.720 --> 00:08:26.600
Should we engineer plants in that way to contain digital information that doesn't belong to them.

s95
00:08:26.600 --> 00:08:39.240
And when I started working at MIT and I started having this idea of you know, maybe stop to stop engineering life because that's pretty hard to do and maybe just work with engineering things to be more life-like.

s96
00:08:39.240 --> 00:08:45.680
It felt much more simple to me, much more much more I don't know, less ethically problematic to some degree.

s97
00:08:45.680 --> 00:08:49.520
So one of the first things I built when I got to MIT was this machine which is the Anamoi device.

s98
00:08:49.520 --> 00:08:50.480
1 2 3 4

s99
00:08:50.480 --> 00:08:54.200
This is a scent memory machine which basically takes any image input.

s100
00:08:54.200 --> 00:09:04.360
In this case it's a physical photograph but it can be any image input and then for a multimodal pipeline transforms that into a scent and then you have this kind of scent memory relapse

s101
00:09:04.360 --> 00:09:09.720
association that takes you back to things which you may or may not have lived yourself.

s102
00:09:09.720 --> 00:09:12.400
And while this in itself is a machine it doesn't move.

s103
00:09:12.400 --> 00:09:13.480
Again, it doesn't reproduce.

s104
00:09:13.480 --> 00:09:18.120
It doesn't have the qualities of a living thing that you might normally prescribe.

s105
00:09:18.120 --> 00:09:19.880
It has this connection to the real world.

s106
00:09:19.880 --> 00:09:25.640
It has a connection to the senses that most applications of intelligence or artificial intelligence do not have.

s107
00:09:25.640 --> 00:09:28.480
And so I just wanted to keep working in this manner.

s108
00:09:28.480 --> 00:09:37.760
And from working with this kind of like multi-sensory initial prototype experiment I got more into this idea of working with embodiment.

s109
00:09:37.760 --> 00:09:42.720
And my inspiration for this, I would say, just to go a bit deeper, comes from three main things.

s110
00:09:42.720 --> 00:09:51.760
One is a very unfashionable branch of philosophy, which is called object-oriented ontology, which is all about essentially creating a flat ontology of things.

s111
00:09:51.760 --> 00:09:59.960
It's all about things, like chairs and universes and unicorns and memories, and they're all ontologically equally existent.

s112
00:09:59.960 --> 00:10:05.840
They don't They're not the same, and they don't matter as much, but they all exist at the same degree.

s113
00:10:05.840 --> 00:10:11.080
And it really questions to what degree we as human beings are privileged in the world or the universe.

s114
00:10:11.080 --> 00:10:12.920
Like, are we really the most important thing?

s115
00:10:12.920 --> 00:10:14.120
Probably not.

s116
00:10:14.120 --> 00:10:20.440
And especially as AI comes in, that really should question and bring into question the ontology of of things existing.

s117
00:10:20.440 --> 00:10:22.520
So, that's very important in the work I do.

s118
00:10:22.520 --> 00:10:34.120
The second thing, as I'm kind of person who works with physical things, is obviously aesthetics are important, and taste and beauty is a huge discussion point now in Silicon Valley and SF and places like that.

s119
00:10:34.120 --> 00:10:41.000
But if you take a step back and look at the root of aesthetics, the original word is actually aesthesis, the word on the screen here.

s120
00:10:41.000 --> 00:10:43.040
That pertains to something much broader.

s121
00:10:43.040 --> 00:10:58.440
That pertains to the perceptual wisdom, sensory wisdom, and embodiment that is actually much more important than superficial external beauty, which aesthetics essentially condenses down to and has been condensed down to since the 19th century or so.

s122
00:10:58.440 --> 00:11:02.839
So, I want to reclaim the original sense when I'm designing for physical intelligence.

s123
00:11:02.839 --> 00:11:03.800
And the third thing is nature.

s124
00:11:03.800 --> 00:11:13.520
So, in the past, I worked a lot more directly with what you would consider traditionally nature, like a tree or a plant, because that is clearly nature to us.

s125
00:11:13.520 --> 00:11:21.400
But I see nature as something that is not external, that is something we are all part of, and everything that we engineer as people engineering AI or whatever else you're engineering,

s126
00:11:21.400 --> 00:11:23.080
that is also going to become part of nature.

s127
00:11:23.080 --> 00:11:28.240
So, you have to think in that manner, or I try to think in that manner as well.

s128
00:11:28.240 --> 00:11:37.280
So, with those three pillars in mind, I started thinking about AI embodiment, and I knew I didn't want to keep I didn't want to design something that was humanoid or even zoomorphic.

s129
00:11:37.280 --> 00:11:45.360
I wanted to think about something that existed outside the parameters or the traditional form factors that we might design with or design for.

s130
00:11:45.360 --> 00:11:51.320
And I also thought a lot about how I and other people are interacting with with artificial intelligence.

s131
00:11:51.320 --> 00:11:53.160
And mostly it exists without form.

s132
00:11:53.160 --> 00:11:57.520
It's similar to data, you know, it might live in a computer or a device.

s133
00:11:57.520 --> 00:11:58.920
Lives essentially in the cloud.

s134
00:11:58.920 --> 00:12:01.760
We have interfaces which are digital to interact with it.

s135
00:12:01.760 --> 00:12:04.840
But there's no like physical footprint of it normally around.

s136
00:12:04.840 --> 00:12:14.720
I mean, again, in this room, probably this doesn't apply quite as much as normally because there are literally humanoid robots walking by right now and things like that.

s137
00:12:14.720 --> 00:12:18.680
But typically AI is almost entirely without form.

s138
00:12:18.680 --> 00:12:25.040
So I wanted to think about what happens when you give it a form that it doesn't have clear affordances with.

s139
00:12:25.040 --> 00:12:31.440
It doesn't have like a a head you can clearly see or arms that can clearly be labeled or mistaken.

s140
00:12:31.440 --> 00:12:33.760
So not working with a lamp, for example.

s141
00:12:33.760 --> 00:12:40.560
And fortunately at the Media Lab we have some shape displays which are remnants of I think research in the 2010s.

s142
00:12:40.560 --> 00:12:42.520
These are not things that I built.

s143
00:12:42.520 --> 00:12:45.839
People who are far better at mechanical engineering built that.

s144
00:12:45.839 --> 00:12:53.040
And it's the perfect device or the perfect apparatus for what I was thinking because it has no clear affordances.

s145
00:12:53.040 --> 00:12:57.240
It's just a almost neutral surface which can move and do things.

s146
00:12:57.240 --> 00:13:00.680
It has no face, has no limbs, has no instruction manual.

s147
00:13:00.680 --> 00:13:04.680
So I began to work with this with the shape display.

s148
00:13:05.880 --> 00:13:09.000
And as you saw what it did initially, it breathed and so on and so forth.

s149
00:13:09.000 --> 00:13:16.720
But if I take a step back and think about what that meant, well, it tried to initially kind of act human in a way or do human pleasing things.

s150
00:13:16.720 --> 00:13:24.040
Tried to write to me in a language I understand, which is really not what I wanted a non-anthropomorphic surface to do.

s151
00:13:24.040 --> 00:13:28.080
It was also very slow technically, so you know, I'm I'm prompting it.

s152
00:13:28.080 --> 00:13:39.640
I'm trying to have a conversation with this other intelligence which has this body and it would take, you know, 45 seconds, a minute, 2 minutes, whatever, to respond to me.

s153
00:13:39.640 --> 00:13:48.000
The latency was very uncomfortable because if I speak to you and then you take 2 minutes to reply with a nod, that's not very good.

s154
00:13:48.000 --> 00:13:49.400
So, we needed to work on that.

s155
00:13:49.400 --> 00:13:57.880
And the third thing was of course it doesn't remember anything because it was February and no one had thought about memory and recollection at that point.

s156
00:13:57.880 --> 00:14:01.240
So, I started developing this system which I call Numelab.

s157
00:14:01.240 --> 00:14:02.840
I won't explain the the name.

s158
00:14:02.840 --> 00:14:06.120
There's a blog post you can read about why it's called Numelab.

s159
00:14:06.120 --> 00:14:13.760
And essentially what Numelab is is a closed-loop system for generating a body language for this system.

s160
00:14:13.760 --> 00:14:23.400
So, the reason behind that was because basically the latency, if I if we could design a body language and give the intelligence a repertoire of gestures like a shrug, a nod,

s161
00:14:23.400 --> 00:14:35.200
a shake of the head, a way to express a smile and things like that, it could probably respond much more quickly and in time for a conversation with me, which is what I'm aiming or was aiming to do.

s162
00:14:35.200 --> 00:14:46.920
So, it works in this kind of this loop where it looks into a database of different gestures, emotions, expressions it should try to emote or provide, go through some validation gates

s163
00:14:46.920 --> 00:14:52.960
to try to make sure that the expressions are legible or readable for humans for example.

s164
00:14:52.960 --> 00:14:56.160
An agent then scores those as they come out.

s165
00:14:56.160 --> 00:15:04.920
There are many, many, many, many of these produced and then at the end there's a human in the loop who kind of like validates, verifies, make sure things are appropriate

s166
00:15:04.920 --> 00:15:08.280
and somehow readable and then we store that and move on to the next thing.

s167
00:15:08.280 --> 00:15:15.040
And that's been running for several weeks at the lab and looks something like this.

s168
00:15:15.040 --> 00:15:17.680
The basically a lot of cameras pointed at the shape display.

s169
00:15:17.680 --> 00:15:27.240
In the back end the agent is just going through loops and loops and loops of gestures trying to create different kinds of expression, scoring things, moving on.

s170
00:15:27.240 --> 00:15:30.160
And after several weeks, it created a language.

s171
00:15:30.160 --> 00:15:38.360
So, this hasn't been published yet, neither in videos or in any other form, but it has now achieved something like 32 gestures.

s172
00:15:38.360 --> 00:15:41.120
There are many, many more, but there are 32 pretty good gestures.

s173
00:15:41.120 --> 00:15:44.040
These are not the gestures, this is just some cool visual art.

s174
00:15:44.040 --> 00:15:47.320
These are the gestures here, or some of the gestures here.

s175
00:15:47.320 --> 00:15:49.120
And you can see some of them moving around.

s176
00:15:49.120 --> 00:15:50.600
Some of them are duplicates as well.

s177
00:15:50.600 --> 00:15:55.800
But, essentially, what's happening is the language model is using this shape display as its body, has a body language now.

s178
00:15:55.800 --> 00:16:00.920
I can talk to it via any kind of I can talk to it, I can write to it, and so on and so forth.

s179
00:16:00.920 --> 00:16:02.160
I I can even wave at it.

s180
00:16:02.160 --> 00:16:04.920
I can I can body language to body language.

s181
00:16:04.920 --> 00:16:05.760
And it responds.

s182
00:16:05.760 --> 00:16:11.880
And what's interesting about it is that the the body part responds faster than the language part at this point.

s183
00:16:11.880 --> 00:16:14.400
The latency is actually really, really quick.

s184
00:16:14.400 --> 00:16:18.840
If I ask it a yes or no question, the nod happens almost instantly.

s185
00:16:18.840 --> 00:16:22.760
So, that's where I'm at with with this thing right now.

s186
00:16:22.760 --> 00:16:28.240
And it's going beyond this, and it's currently in kind of experiment testing mode.

s187
00:16:28.240 --> 00:16:38.920
People are coming into the lab, having sessions with the agent, leaving, feeling really worried about the future, and or unsettled about the future, not worried quite as much.

s188
00:16:38.920 --> 00:16:42.560
But, where I think this is going is kind of summed up on this page.

s189
00:16:42.560 --> 00:16:52.800
So, as I already touched on, I think we're focusing too much right now, especially with when we're thinking about physical things, there's a too much talk about taste and aesthetics.

s190
00:16:52.800 --> 00:16:59.000
And I want to do some other stuff with that word and put AI in front of it, apparently, and make it aestethics.

s191
00:16:59.000 --> 00:17:02.280
And that reclaims again this essence of aesthesis.

s192
00:17:02.280 --> 00:17:04.319
And that does four things.

s193
00:17:04.319 --> 00:17:16.280
I think when I've been working with this system or entity or agent or being or whatever we want to call it, it's definitely been very different to any kind of machine or experiment

s194
00:17:16.280 --> 00:17:21.439
or anything I've really done before apart from really encountering people or other beings.

s195
00:17:21.439 --> 00:17:25.000
And so, it feels like this machine can sense me.

s196
00:17:25.000 --> 00:17:30.720
It definitely can sense me, technically, but it also feels like it can sense me in a very strange way.

s197
00:17:30.720 --> 00:17:39.080
And I can sense it, and that's a completely different interaction than anything else I've ever explored technically.

s198
00:17:39.080 --> 00:17:40.920
And other people are also sharing this, by the way.

s199
00:17:40.920 --> 00:17:42.760
This is not just my delusion.

s200
00:17:42.760 --> 00:17:51.960
And the second point is that by developing this body language, this gesture vocabulary, whatever you want to call it, it goes beyond what I thought.

s201
00:17:51.960 --> 00:17:58.240
I thought initially people might read this as an emoji or something, and I was really quite tentative about the testing.

s202
00:17:58.240 --> 00:18:09.040
I thought this would definitely break down with other people, but actually everyone seems to feel like this this expression is really important and adds a whole other layer of value

s203
00:18:09.040 --> 00:18:12.000
to up to interacting with what is essentially just a chatbot.

s204
00:18:12.000 --> 00:18:15.200
It's still the same chatbots that you use every day.

s205
00:18:15.200 --> 00:18:17.920
And this isn't it's not like a visualizer.

s206
00:18:17.920 --> 00:18:22.440
It adds this richness, texture, feeling, sensation, whatever.

s207
00:18:22.440 --> 00:18:23.800
All of these words added.

s208
00:18:23.800 --> 00:18:26.120
It's hard to put words really to it.

s209
00:18:26.120 --> 00:18:28.200
It's really a feeling.

s210
00:18:28.200 --> 00:18:40.200
And then the third thing is that clearly we're beginning to through this work we can show that it's actually very easy and quite exciting to diverge from humanoid or zoomorphic forms of physical intelligence.

s211
00:18:40.200 --> 00:18:46.040
And I know lots of people are already doing much more interesting work than this, but for me this was very very new to see.

s212
00:18:46.040 --> 00:18:50.920
And also the fact that we don't have to just operationalize AI to be our helper.

s213
00:18:50.920 --> 00:18:53.120
It can also be other things.

s214
00:18:53.120 --> 00:18:57.280
And I'm not saying what this is right now, but there are other things it definitely can be.

s215
00:18:57.280 --> 00:19:01.840
And then finally, like this idea of using I'm not I'm I don't know, maybe you can tell at this point.

s216
00:19:01.840 --> 00:19:04.840
I'm not really a big problem-solving person.

s217
00:19:04.840 --> 00:19:11.440
I don't I like using things and I like tools and so on because they help me to achieve things, but this is far more interesting to me.

s218
00:19:11.440 --> 00:19:15.680
Creating things like this, which again create the sensation, this feeling.

s219
00:19:15.680 --> 00:19:24.880
And I think that this can be combined into things which are productive and useful and create new associations with things that just add more value in our world, make us feel

s220
00:19:24.880 --> 00:19:32.520
like a bit more magic, a bit It's a bit Pixar, really, but in the real world, not just on a on a 2D screen that we're watching.

s221
00:19:32.520 --> 00:19:39.040
So, all of this contributes to what I'm building at MIT and what I'll be building after MIT.

s222
00:19:39.040 --> 00:19:53.640
I'm literally writing Well, my thesis is Aesthetic Machines and I'm writing a thesis which is called Aesthetic Machines, which basically encompasses all of this thinking to think about how AI could leave the screen and enter the real world and be accepted by people

s223
00:19:53.640 --> 00:19:56.360
and not be quite so terrifying or scary.

s224
00:19:56.360 --> 00:20:01.000
And I think my big hunch on this is that this word aesthetics is important.

s225
00:20:01.000 --> 00:20:13.240
We need to think about physical intelligence that is sensory, is perceptual, is embodied in ways that we can understand that isn't feeling crazy, alien, scary to us, but feels relevant,

s226
00:20:13.240 --> 00:20:19.880
welcoming, affectionate, expressive in ways that we can engage with it and understand.

s227
00:20:20.120 --> 00:20:22.920
So, that's that.

s228
00:20:22.920 --> 00:20:25.200
And thank you very much for listening.

s229
00:20:25.200 --> 00:20:28.794
You can find me on the internet everywhere.

s230
00:20:28.794 --> 00:20:30.794
[applause]

s231
00:20:45.824 --> 00:20:47.824
[music]