1
00:00:02,440 --> 00:00:06,100
3, 2, 1, 0.
2
00:00:06,125 --> 00:00:17,200
Welcome aboard the System
3
00:00:17,225 --> 00:00:19,100
76 transmission Log.
4
00:00:19,125 --> 00:00:21,540
Our broadcast is about to begin.
5
00:00:21,565 --> 00:00:27,180
This is the latest on System 76
computers, manufacturing, and pop OS.
6
00:00:27,205 --> 00:00:30,260
Now for your in-orbit crew.
7
00:00:30,285 --> 00:00:34,340
Hey, Alex.
8
00:00:34,365 --> 00:00:36,940
Tell me about your experience
at All Things Open.
9
00:00:36,965 --> 00:00:38,340
It was my first time.
10
00:00:38,365 --> 00:00:43,800
I was helping work the booth with Stetson
and Aaron, who are from the sales team
11
00:00:43,825 --> 00:00:46,060
and support team, respectively.
12
00:00:46,085 --> 00:00:47,560
It was pretty fun.
13
00:00:47,585 --> 00:00:49,420
We had a lot of attention at the booth.
14
00:00:49,445 --> 00:00:54,780
We were the only consumer hardware company
that was there that I'm aware of.
15
00:00:54,805 --> 00:00:57,420
We had a lot of excitement.
16
00:00:57,445 --> 00:01:01,300
Ato was raffling off
three of our in laptops.
17
00:01:01,325 --> 00:01:03,400
We got a lot of questions about that.
18
00:01:03,425 --> 00:01:09,560
We were able to give the spiel
about System 76 and our open-source
19
00:01:09,585 --> 00:01:11,540
mission, our Denver manufacturing.
20
00:01:11,565 --> 00:01:17,095
It was great to educate
the enterprise masses about us.
21
00:01:17,120 --> 00:01:17,770
Cool.
22
00:01:17,800 --> 00:01:20,920
I feel like we were here
recording just yesterday.
23
00:01:20,960 --> 00:01:23,060
The month flew by pretty fast.
24
00:01:23,085 --> 00:01:25,290
Yeah, it's almost Thanksgiving.
It's crazy.
25
00:01:25,320 --> 00:01:29,780
Yeah, and when it's close to Thanksgiving,
we celebrate our birthday
26
00:01:29,805 --> 00:01:34,380
because System 76 was born
or founded on November seventh.
27
00:01:34,405 --> 00:01:40,060
We celebrated our 19th birthday
this month, and we had a feast with cake
28
00:01:40,085 --> 00:01:44,460
and presented Carl with a nice canvas
that had a company photo on it.
29
00:01:44,485 --> 00:01:48,490
Now he can see our smiling faces
on the wall every single day.
30
00:01:48,520 --> 00:01:52,570
Yeah, walking into the office
is like walking into grandma's house.
31
00:01:52,600 --> 00:01:56,040
You see your photo from
fifth grade up on the wall,
32
00:01:56,065 --> 00:01:59,210
and you're like, Oh, good times.
33
00:01:59,240 --> 00:02:00,780
Thanks, grandma.
34
00:02:00,805 --> 00:02:01,850
Grandma Carl.
35
00:02:01,880 --> 00:02:06,280
I like when we go on tours and
we stop by the photo wall on the way out,
36
00:02:06,305 --> 00:02:08,450
and I can tell stories of each photo.
37
00:02:08,480 --> 00:02:09,940
That's really fun.
38
00:02:09,965 --> 00:02:13,940
We have some meetups
coming up, too, as well.
39
00:02:13,965 --> 00:02:17,420
We go on factory tours during those.
40
00:02:17,445 --> 00:02:19,610
There's some catered food and drinks.
41
00:02:19,640 --> 00:02:21,170
We play some games.
42
00:02:21,200 --> 00:02:24,980
We get to chat with the community,
chat with locals.
43
00:02:25,005 --> 00:02:28,215
It's a good time.
The next tour, I believe, is in December.
44
00:02:28,240 --> 00:02:28,890
Is that right?
45
00:02:28,920 --> 00:02:33,735
We have our December December
and January meetups posted on meetup.
46
00:02:33,760 --> 00:02:34,500
Com.
47
00:02:34,525 --> 00:02:38,120
What would you say is
your favorite part on the tour?
48
00:02:38,145 --> 00:02:40,260
When we get to the Blue Laser.
49
00:02:40,285 --> 00:02:42,060
Why the Blue Laser?
50
00:02:42,085 --> 00:02:43,880
I can't tell you.
51
00:02:44,560 --> 00:02:45,800
Hang on.
Secrets.
52
00:02:45,825 --> 00:02:49,260
You have to go on the tour to find out.
53
00:02:49,285 --> 00:02:53,280
Okay, well, I'm going to
schedule a tour after this podcast.
54
00:02:53,520 --> 00:02:55,980
How do I schedule a tour
after this podcast?
55
00:02:56,005 --> 00:03:00,015
You have to email me,
emma, E-M-M-M-A@system70.
56
00:03:00,040 --> 00:03:00,220
Com.
6.
57
00:03:00,245 --> 00:03:05,260
Com directly, and you can schedule it, or
you can come on one of our group tours.
58
00:03:05,285 --> 00:03:09,080
I will email you to get those secrets.
59
00:03:09,240 --> 00:03:15,080
Okay.
All right, let's get into hardware news.
60
00:03:16,720 --> 00:03:22,200
We talked in-depth about Thelio ASTRA last
month, and there have been some
61
00:03:22,225 --> 00:03:24,420
benchmarking results released by Foronix.
62
00:03:24,445 --> 00:03:29,610
Those show some promising results,
and you can check it out on foronix.
63
00:03:29,640 --> 00:03:32,100
Com if you want to get some details there.
64
00:03:32,125 --> 00:03:36,780
Yeah, we featured Thelio ASTRA at ATO.
65
00:03:36,805 --> 00:03:42,460
People were excited to talk to us
about our new Arm 64 workstation.
66
00:03:42,485 --> 00:03:45,860
It's an Ampere processor up to 128
67
00:03:45,885 --> 00:03:50,610
cores, up to 6000 data graphics.
68
00:03:50,640 --> 00:03:53,460
So beast of a machine, beautiful looking.
69
00:03:53,485 --> 00:03:57,540
We were right next to the arm booth
at the conference, too.
70
00:03:57,565 --> 00:03:59,880
So that was funny.
71
00:03:59,905 --> 00:04:02,340
Sort of a funny coincidence.
72
00:04:02,365 --> 00:04:03,660
And yeah.
73
00:04:03,685 --> 00:04:06,900
Thaleoastra has really
been making the rounds.
74
00:04:06,925 --> 00:04:12,020
We're going to be at CoupaCon coming up
pretty soon here, and there'll be
75
00:04:12,045 --> 00:04:18,015
four Thaleo Astras at different booths
doing different demos at each booth.
76
00:04:18,040 --> 00:04:18,520
That's cool.
77
00:04:18,545 --> 00:04:23,420
We'll have more details
about that posted on our blog.
78
00:04:23,445 --> 00:04:28,220
Another newcomer
to the Thaleo desktop line is Thaleo
79
00:04:28,245 --> 00:04:32,140
Spark, our AMD-powered starter desktop.
80
00:04:32,165 --> 00:04:36,340
It is an energy-efficient work of art
that fits perfectly on any desk.
81
00:04:36,365 --> 00:04:41,860
You may remember Spark from
the first Thaleo Head music video.
82
00:04:41,885 --> 00:04:43,900
It was on our minds, and it still is.
83
00:04:43,925 --> 00:04:49,660
Now it's an AMD 8000 G-Series system and
84
00:04:49,685 --> 00:04:53,060
lower starting price at, I think, 799.
85
00:04:53,085 --> 00:04:55,400
Beautiful design, good starter desktop.
86
00:04:55,425 --> 00:04:58,820
If you're starting something new
or just want something fast
87
00:04:58,845 --> 00:05:00,980
that fits within budget.
88
00:05:01,005 --> 00:05:04,420
New desktops are great,
but what about desktop environments?
89
00:05:04,445 --> 00:05:05,740
What's new in cosmic?
90
00:05:05,765 --> 00:05:09,060
You got to talk about
the environment, Emma.
91
00:05:09,085 --> 00:05:14,280
So cosmic is a vast
environment of possibility.
92
00:05:14,305 --> 00:05:15,940
We are now in Alpha 3.
93
00:05:15,965 --> 00:05:17,260
Very exciting.
94
00:05:17,285 --> 00:05:21,000
We have some changes to cosmic
files, a lot of smoothing over
95
00:05:21,025 --> 00:05:24,340
with the preview mode and gallery mode.
96
00:05:24,365 --> 00:05:26,980
Preview mode is no longer
going to be on by default,
97
00:05:27,005 --> 00:05:29,860
and there's a new footer widget.
98
00:05:29,885 --> 00:05:34,180
So When you click on a file, the button
isn't going to change locations.
99
00:05:34,205 --> 00:05:35,960
It's just a lot neater.
100
00:05:36,080 --> 00:05:40,960
Then Cosmic Store now has
a made for cosmic section.
101
00:05:40,985 --> 00:05:46,100
You're going to see some nice cosmic apps
made by you guys, mostly.
102
00:05:46,125 --> 00:05:51,080
Things like a tasks manager
for your to-do list, a weather app.
103
00:05:51,105 --> 00:05:52,840
I think I saw a Pokédex in there.
104
00:05:52,865 --> 00:05:56,020
Someone is out there speaking my language.
105
00:05:56,045 --> 00:06:01,080
I'm just excited to see how long that list
is going to be by official release.
106
00:06:01,105 --> 00:06:03,060
It's pretty cool.
107
00:06:03,085 --> 00:06:09,140
Then lastly, we have screen reader
support in initial testable state.
108
00:06:09,165 --> 00:06:12,120
It's not going to be perfect,
but we're interested in hearing
109
00:06:12,145 --> 00:06:17,815
your feedback in the first testing
of accessibility in Cosmic.
110
00:06:17,840 --> 00:06:18,300
Yeah.
111
00:06:18,325 --> 00:06:23,120
If you want to give it a try,
go ahead and pop on over to system76.
112
00:06:23,145 --> 00:06:26,980
Com/cosmic,
and you can download the Alpha ISO.
113
00:06:27,005 --> 00:06:31,655
I can't forget to mention we
have a lot of bug fixes in Alpha 3.
114
00:06:31,680 --> 00:06:32,620
Thanks to you guys.
115
00:06:32,645 --> 00:06:36,060
You're doing great work
on these contributions here.
116
00:06:36,085 --> 00:06:39,700
It's really feeling like
a polished beta to me at this point,
117
00:06:39,725 --> 00:06:41,780
and we appreciate the support.
118
00:06:41,805 --> 00:06:44,680
I'm so excited for our guest this month.
119
00:06:44,705 --> 00:06:47,180
Who is our guest this month?
120
00:06:47,205 --> 00:06:52,600
We have Shantanu from Traslin Robotics
to talk about using System76 hardware
121
00:06:52,625 --> 00:06:56,480
and open source to create cutting edge
robotics solutions for research labs
122
00:06:56,505 --> 00:06:59,060
and educational institutions.
123
00:06:59,085 --> 00:07:02,460
That sounds awesome, and
I think we're going to learn a lot.
124
00:07:02,485 --> 00:07:06,220
So let's get into it.
125
00:07:06,245 --> 00:07:08,680
Thanks for joining us today, Shantanu.
126
00:07:08,705 --> 00:07:11,740
It's very exciting to have you
on our podcast today.
127
00:07:11,765 --> 00:07:15,340
This is Shantanu from Tracin Robotics.
128
00:07:15,365 --> 00:07:19,935
Can you introduce yourself and tell us
a little bit about what you do at Tracin?
129
00:07:19,960 --> 00:07:20,440
Hi.
130
00:07:20,465 --> 00:07:23,480
Thank you so much for having
me on this podcast.
131
00:07:23,680 --> 00:07:27,240
My name is Shantanu, and
I am a machine learning specialist
132
00:07:27,265 --> 00:07:28,740
here at Tracin Robotics.
133
00:07:28,765 --> 00:07:30,860
I work with the Aloha Kids.
134
00:07:30,885 --> 00:07:36,160
We develop Aloha Kids as a research
product for engineers and researchers so
135
00:07:36,185 --> 00:07:42,920
that they can easily set up their
experiment and get up and running with any
136
00:07:42,945 --> 00:07:47,900
task that they have in terms of machine
learning, especially imitation learning.
137
00:07:47,925 --> 00:07:54,280
So my goal here at Trouzen Robotics is
to make it as easy as possible for people
138
00:07:54,305 --> 00:07:57,280
and researchers to get started
with machine learning,
139
00:07:57,305 --> 00:08:02,900
may it be in terms of setting up their
hardware, software and making it
140
00:08:02,925 --> 00:08:08,080
as easy as possible to use so
that they don't have any difficulty
141
00:08:08,105 --> 00:08:12,500
setting up all these critical
functionalities in machine learning.
142
00:08:12,525 --> 00:08:15,260
So, yeah, that's my job here.
143
00:08:15,285 --> 00:08:19,135
Could you talk a little bit
about the Aloha kits and what those are?
144
00:08:19,160 --> 00:08:20,020
Yeah.
145
00:08:20,045 --> 00:08:24,400
So the Aloha kits started as
a collaboration between Stanford
146
00:08:24,425 --> 00:08:26,940
University and Truss and Robotics.
147
00:08:26,965 --> 00:08:33,040
Our researchers at Stanford University
were looking for a low-cost, affordable
148
00:08:33,065 --> 00:08:38,140
product that they can use to develop
machine learning for their research.
149
00:08:38,165 --> 00:08:42,220
So they collaborated with Truss
and Robotics to create these products
150
00:08:42,245 --> 00:08:45,020
that are today known as the Aloha Kids.
151
00:08:45,045 --> 00:08:50,040
Essentially, these are a leader-follower
arm pair in which you will be having
152
00:08:50,065 --> 00:08:54,260
a Widow X arm and a Viper X arm.
153
00:08:54,285 --> 00:08:58,460
It's essentially used for
teaching a robot how to do stuff
154
00:08:58,485 --> 00:09:00,620
using bimanual manipulation.
Education.
155
00:09:00,645 --> 00:09:05,260
As a human, it's easier for us to
look at something and learn it quickly.
156
00:09:05,285 --> 00:09:09,480
Maybe it might be opening
up the door, doing something
157
00:09:09,505 --> 00:09:12,020
as easy as tying a shoelace.
158
00:09:12,045 --> 00:09:15,180
But for robots,
this is a complicated task.
159
00:09:15,205 --> 00:09:19,240
In order to teach a robot how to do stuff,
there are multiple different ways,
160
00:09:19,265 --> 00:09:25,300
like reinforcement learning,
writing heavy code that tells a robot
161
00:09:25,325 --> 00:09:27,580
what traject it has to do.
162
00:09:27,605 --> 00:09:31,800
But as an easier solution, Imitation
163
00:09:31,825 --> 00:09:35,240
learning is something which is a go-to
164
00:09:35,265 --> 00:09:38,740
in terms of machine learning landscape.
165
00:09:38,765 --> 00:09:44,000
What people do is they use the leader arms
to tell the robot how something
166
00:09:44,025 --> 00:09:46,580
has to be done, like tying a shoelace.
167
00:09:46,605 --> 00:09:50,120
Once you collect enough episodes,
like maybe 50 or 100 episodes
168
00:09:50,145 --> 00:09:54,560
of this task, you can train it
using machine learning algorithm.
169
00:09:54,585 --> 00:09:58,660
Initially, it was act, which is
action-chunking transformers.
170
00:09:58,685 --> 00:10:02,640
Once the robot knows how to
solve this problem, it will be able
171
00:10:02,665 --> 00:10:04,980
to do it autonomously without any help.
172
00:10:05,005 --> 00:10:10,520
There are multiple papers
that are published recently in conference
173
00:10:10,545 --> 00:10:12,800
going on at Munich, Germany.
174
00:10:12,825 --> 00:10:17,680
There are papers which are also
deemed as the foundation models
175
00:10:17,705 --> 00:10:23,020
for the growth in the machine learning
landscape that we see right now.
176
00:10:23,045 --> 00:10:28,880
I can keep talking about this
all night, all day, but I guess
177
00:10:28,905 --> 00:10:34,220
that answers your question in terms
of what the Aloha kit can be used for.
178
00:10:34,245 --> 00:10:39,180
It's basically like an all-in-one
kit for machine learning?
179
00:10:39,205 --> 00:10:40,620
Yes.
180
00:10:40,645 --> 00:10:43,215
With the robot arm?
Yes.
181
00:10:43,240 --> 00:10:43,940
Cool.
182
00:10:43,965 --> 00:10:49,260
Some of our listeners might be roboticists
or enthusiasts about robotics.
183
00:10:49,285 --> 00:10:53,060
What in the industry
excites you most about robotics today?
184
00:10:53,085 --> 00:10:59,640
For anyone in the robotics industry
or even those people who are not in
185
00:10:59,665 --> 00:11:03,140
the robotics industry right now, people
know the hype about humanoid robots.
186
00:11:03,165 --> 00:11:08,840
There has been a rise in the companies
that are producing humanoid robots
187
00:11:08,865 --> 00:11:13,700
like Figur, Tesla, and Google.
188
00:11:13,725 --> 00:11:18,480
I know that these It seemed like
the future is near, but the amount of work
189
00:11:18,505 --> 00:11:23,300
that goes behind these robots to perform
simple tasks, the way I said it,
190
00:11:23,325 --> 00:11:27,740
tying a shoelace,
it's top of my mind because we just saw
191
00:11:27,765 --> 00:11:32,660
the release of a paper called as
Aloha Unleashed, which was
192
00:11:32,685 --> 00:11:34,900
published by Google DeepMind.
193
00:11:34,925 --> 00:11:39,740
If you see the core of the paper
is they used the Aloha kit
194
00:11:39,765 --> 00:11:47,100
that we provide, and they achieved
a very simple task of tying a shoelace.
195
00:11:47,125 --> 00:11:50,560
But for a normal person,
this is a very easy task,
196
00:11:50,585 --> 00:11:53,420
and no one knows how complex it is.
197
00:11:53,445 --> 00:11:57,520
But when you read into the paper,
you will see that to train the robot to
198
00:11:57,545 --> 00:12:03,740
do that task, it was so complicated that
it required immense amount of GPU power.
199
00:12:03,765 --> 00:12:07,340
It required immense amount
of training episodes.
200
00:12:07,365 --> 00:12:12,540
It took about 11 days for
it to train to do that simple task.
201
00:12:12,565 --> 00:12:17,960
The most exciting part for me is we
see these robots that are going
202
00:12:17,985 --> 00:12:24,560
to work in the industry, in factories
and other places, and they'll be soon
203
00:12:24,585 --> 00:12:26,220
collaborating with humans.
204
00:12:26,245 --> 00:12:30,160
But the fundamentals, training
these robots to do something, starts
205
00:12:30,185 --> 00:12:33,620
at very basic, that's the Aloha Kids.
206
00:12:33,645 --> 00:12:40,400
That's me being a part of it and me
being able to contribute to the community
207
00:12:40,425 --> 00:12:45,660
in a way where it becomes easier
for people to use these robots
208
00:12:45,685 --> 00:12:47,900
is something very exciting for me.
209
00:12:47,925 --> 00:12:49,020
That's awesome.
210
00:12:49,045 --> 00:12:54,015
Can you tell us a little bit about
the Interbotx open source repositories?
211
00:12:54,040 --> 00:12:54,720
Yeah.
212
00:12:54,745 --> 00:12:57,440
So Interbotx is why we
call it open source.
213
00:12:57,465 --> 00:13:02,240
We don't have any control roll
over the software in terms
214
00:13:02,265 --> 00:13:04,060
of we don't sell the software.
215
00:13:04,085 --> 00:13:06,500
We just are a hardware company.
216
00:13:06,525 --> 00:13:12,360
The reason why we create the software
is that people should not
217
00:13:12,385 --> 00:13:16,520
face any difficulties
while interfacing their robots
218
00:13:16,545 --> 00:13:19,140
with the computers they are going to use.
219
00:13:19,165 --> 00:13:25,040
They need something which is very sleek,
easy to use, and they don't
220
00:13:25,065 --> 00:13:27,940
spend a lot of time debugging stuff.
221
00:13:27,965 --> 00:13:33,520
What intervallex does is it gives you
a middleware framework, which is a layer
222
00:13:33,545 --> 00:13:39,080
of ROSS, which is an existing robotics
operating system framework, which
223
00:13:39,105 --> 00:13:41,220
is widely popular among roboticists.
224
00:13:41,245 --> 00:13:46,320
We use this communication protocol or
the middleware framework,
225
00:13:46,345 --> 00:13:52,080
build upon it using our Inovatex software
and make it as easy as possible for people
226
00:13:52,105 --> 00:13:57,080
to calibrate their robots,
run their robots, debug if anything goes
227
00:13:57,105 --> 00:14:02,640
wrong in terms of maybe servos,
not picking up the current correctly or
228
00:14:02,665 --> 00:14:07,560
the commands correctly or
servos dying in the middle of operation.
229
00:14:07,585 --> 00:14:14,360
We provide elaborate workflows
that will help them to debug anything at
230
00:14:14,385 --> 00:14:21,320
any point in time and get up and running
with their robots in about an hour after
231
00:14:21,345 --> 00:14:24,020
they have set up the robot correctly.
232
00:14:24,045 --> 00:14:26,280
Yeah, that's the software that we offer.
233
00:14:26,305 --> 00:14:27,920
It's pretty easy to use.
234
00:14:27,945 --> 00:14:32,900
It uses something which is called
ROS, which our robotists
235
00:14:32,925 --> 00:14:34,780
are highly familiar with.
236
00:14:34,805 --> 00:14:40,980
We offer C++ as well as Python binding,
so it makes it easier for you as a user
237
00:14:41,005 --> 00:14:44,740
to choose whatever you want.
238
00:14:44,765 --> 00:14:50,040
And because we support Python as
well, people who build upon it can easily
239
00:14:50,065 --> 00:14:53,040
integrate machine learning concepts
or deep learning concepts into
240
00:14:53,065 --> 00:14:57,180
their projects without worrying about
the software that run behind.
241
00:14:57,205 --> 00:15:01,415
It sounds like we're vastly reducing
the barrier entry for robotics here.
242
00:15:01,440 --> 00:15:02,380
That's awesome.
243
00:15:02,405 --> 00:15:06,460
Have you seen any standout projects yet
from the community?
244
00:15:06,485 --> 00:15:08,580
What are your highlights, would you say?
245
00:15:08,605 --> 00:15:13,920
There are multiple things that are going
out in the community right now
246
00:15:13,945 --> 00:15:18,080
because as you know, there is
a conference on reinforcement learning.
247
00:15:18,105 --> 00:15:23,400
That's the highly anticipated conference
in terms of reinforcement learning
248
00:15:23,425 --> 00:15:26,620
at Robotics that's going
on right now at Munich.
249
00:15:26,645 --> 00:15:31,880
When a conference is close by and people
will be releasing all their work
250
00:15:31,905 --> 00:15:33,780
that they have done throughout the years.
251
00:15:33,805 --> 00:15:40,480
We have noticed that at least five
of the top papers in this conference
252
00:15:40,505 --> 00:15:46,020
use the AloHA kits, use our hardware,
our software to some extent.
253
00:15:46,045 --> 00:15:51,360
There are big companies like
Google DeepMind just least the paper
254
00:15:51,385 --> 00:15:55,060
called as Aloha Unleashed,
which I was talking about before this.
255
00:15:55,085 --> 00:16:00,160
We also have seen a company called as Pai,
256
00:16:00,185 --> 00:16:03,000
which is physical intelligence, that just
257
00:16:03,025 --> 00:16:07,820
announced its own foundation model
that uses Aloha Kit again.
258
00:16:07,845 --> 00:16:14,600
We have seen, I guess, there are
three papers from a researcher from UC,
259
00:16:14,625 --> 00:16:20,980
Berkeley, along with researchers
that are leading Google DeepMind as well.
260
00:16:21,005 --> 00:16:26,000
The core of this paper is to
train something, to train a robot to do a
261
00:16:26,025 --> 00:16:31,620
very simple task, you need a lot of data,
and you to collect a lot of data sets.
262
00:16:31,645 --> 00:16:36,360
To reduce the amount of data
that you collect, people have come
263
00:16:36,385 --> 00:16:41,740
up with projects that let you
use any data on any other robot.
264
00:16:41,765 --> 00:16:46,340
It's something called
as a cross-training platforms.
265
00:16:46,365 --> 00:16:50,520
There are a lot of very interesting
papers, interesting researches
266
00:16:50,545 --> 00:16:53,660
that are going on in the industry.
267
00:16:53,685 --> 00:16:58,320
In terms of what I was saying,
the research papers that are coming out,
268
00:16:58,345 --> 00:17:02,520
and they are coming out so frequently,
and I guess to some extent,
269
00:17:02,545 --> 00:17:06,940
it's because of the way we are reducing
the barrier to entry with our products,
270
00:17:06,965 --> 00:17:10,920
making it as easy as possible
to set up your robots.
271
00:17:10,945 --> 00:17:15,440
We also provide with a community
that helps you publish your data sets,
272
00:17:15,465 --> 00:17:20,800
publish your trained models to a platform
where you can pick the data set
273
00:17:20,825 --> 00:17:23,810
from and train your own models.
274
00:17:23,840 --> 00:17:28,520
So what happens is researchers now
don't have to spend their time
275
00:17:28,545 --> 00:17:31,180
on setting up a hardware platform.
276
00:17:31,205 --> 00:17:36,740
They can invest more time
in coming up with new ideas, new models,
277
00:17:36,765 --> 00:17:41,560
and scale up their research
to an extent that they don't have to
278
00:17:41,585 --> 00:17:43,530
worry about anything in the background.
279
00:17:43,560 --> 00:17:48,420
I guess that's where cross
and robotics helps the researchers.
280
00:17:48,445 --> 00:17:53,900
That's our motto to help researchers
or help innovators innovate.
281
00:17:53,925 --> 00:17:57,880
That goes a lot with our model
of engineering the computer, too,
282
00:17:57,905 --> 00:18:02,480
is just trying to get all the fussing
around out of the way,
283
00:18:02,505 --> 00:18:07,050
all the installing drivers,
moving windows around, things like that,
284
00:18:07,080 --> 00:18:11,700
to just get people focused on what
they actually want to work on.
285
00:18:11,725 --> 00:18:18,040
Now, you're using the Serval workstation
for the Aloha stationery, too.
286
00:18:18,065 --> 00:18:20,810
What led to the decision for the Serval?
287
00:18:20,840 --> 00:18:28,220
As you said, your model
and our model is similar.
288
00:18:28,245 --> 00:18:32,250
The reason is when I I
started using this workstation.
289
00:18:32,280 --> 00:18:36,880
To begin with any project, you need
to install a lot of dependencies, a lot
290
00:18:36,905 --> 00:18:42,010
of different things to interface
your robots with your workstation.
291
00:18:42,040 --> 00:18:43,880
I worked with different...
292
00:18:43,905 --> 00:18:48,480
When I was a student, I worked
with different laptops, and I have seen
293
00:18:48,505 --> 00:18:52,680
other people work with different laptops
and the difficulties that they face while
294
00:18:52,705 --> 00:18:57,460
installing simple things like Linux,
Cuda drivers, or any other thing.
295
00:18:57,485 --> 00:19:03,200
But with this workstation, when I started,
it was within an hour, I was able
296
00:19:03,225 --> 00:19:05,740
to set up all the necessary dependencies.
297
00:19:05,765 --> 00:19:09,840
It took an hour because downloading
NVIDIA code takes an hour.
298
00:19:09,865 --> 00:19:12,460
It wasn't the system
doing something weird.
299
00:19:12,485 --> 00:19:16,600
But yeah, it was a very smooth and error
It was a free experience because most
300
00:19:16,625 --> 00:19:22,380
of the hardware is compatible
with the NVIDIA drivers that you need.
301
00:19:22,405 --> 00:19:27,520
It was as simple as running a command line
argument, like seeing a pseudo
302
00:19:27,545 --> 00:19:29,570
install NVIDIA server.
303
00:19:29,600 --> 00:19:31,330
So It was that simple.
304
00:19:31,360 --> 00:19:35,620
The Linux operation and
the Linux setup was very easy.
305
00:19:35,645 --> 00:19:42,380
I guess that's the reason why we are
partners with System76 to get our laptops
306
00:19:42,405 --> 00:19:47,200
so that the users that we give these
laptops to, the researchers,
307
00:19:47,225 --> 00:19:52,280
don't have to face any other difficulties
that you might face when you get a laptop
308
00:19:52,305 --> 00:19:57,380
which doesn't support the necessary
drivers or the hardware interfaces.
309
00:19:57,405 --> 00:20:01,940
I guess that's the core reason
why we go with system 76.
310
00:20:01,965 --> 00:20:04,620
Awesome.
We're happy to hear that it's going well.
311
00:20:04,645 --> 00:20:08,520
I heard the MeerCAT was
going to be in use soon.
312
00:20:08,545 --> 00:20:12,090
Do you know about that project coming up?
Yes.
313
00:20:12,120 --> 00:20:14,700
The mini PC?
Yeah.
314
00:20:14,725 --> 00:20:18,960
As you see, the Aloha stationery
is a bimanual robotic system.
315
00:20:18,985 --> 00:20:24,400
We had noticed that in the research
industry, people are also
316
00:20:24,425 --> 00:20:26,760
working with single-arm devices.
317
00:20:26,785 --> 00:20:30,810
They don't need two arms
to achieve or complete a task.
318
00:20:30,840 --> 00:20:34,720
So what we noticed was people
required one R interface.
319
00:20:34,745 --> 00:20:40,440
And if we are providing a cheaper
alternative to the Aloha, we also need
320
00:20:40,465 --> 00:20:45,480
something that is as compact as possible
so that people can move around the robots
321
00:20:45,505 --> 00:20:47,040
easily, record their episodes.
322
00:20:47,065 --> 00:20:51,740
And we were looking for a system
that can be useful for this.
323
00:20:51,765 --> 00:20:58,480
And then we saw the System76 MeerKAT
Mini, and it had super processing powers
324
00:20:58,505 --> 00:20:59,700
and interfaces.
325
00:20:59,725 --> 00:21:04,440
And Again, because it is System76,
Linux drivers, Cuda drivers,
326
00:21:04,465 --> 00:21:07,330
and everything was super easy to install.
327
00:21:07,360 --> 00:21:12,220
Given that and how portable
it is, we are soon going to launch
328
00:21:12,245 --> 00:21:14,620
a product called as Aloha Solo.
329
00:21:14,645 --> 00:21:17,460
That that will be using the Meer cat mini.
330
00:21:17,485 --> 00:21:20,380
That's really awesome.
Do you have a timeline for that yet?
331
00:21:20,405 --> 00:21:23,695
It should be out by December first,
officially.
332
00:21:23,720 --> 00:21:24,570
Okay, cool.
333
00:21:24,600 --> 00:21:28,620
Are there any other things that
you'd like to add about trust in robotics
334
00:21:28,645 --> 00:21:31,770
or about the robotics Linux industry
and Linux in general?
335
00:21:31,800 --> 00:21:34,460
Yeah, I'm happy about this partnership.
336
00:21:34,485 --> 00:21:38,000
We are so satisfied that we
have partners like you.
337
00:21:38,025 --> 00:21:42,320
The great parts being the customer
support, the customization that you guys
338
00:21:42,345 --> 00:21:48,280
provide, and how easy it is to get
something resolved in terms of anything
339
00:21:48,305 --> 00:21:53,400
that we need apart from the product,
the support in terms of getting a product
340
00:21:53,425 --> 00:22:00,080
in time, sending it to the customer,
having this beautiful a stream line
341
00:22:00,105 --> 00:22:03,880
process where the customer orders
the product, you get it in time,
342
00:22:03,905 --> 00:22:07,140
and we are able to get it
to the customer in time.
343
00:22:07,165 --> 00:22:11,480
There is absolutely no error in between
because we are able to get the software
344
00:22:11,505 --> 00:22:17,120
and the hardware integrated with System76
with so much ease that it feels
345
00:22:17,145 --> 00:22:19,775
like it's one product itself.
346
00:22:19,800 --> 00:22:22,320
It doesn't look like there are
two different companies working on it.
347
00:22:22,345 --> 00:22:24,780
It just feels like it's
a one full company.
348
00:22:24,805 --> 00:22:28,020
I guess that's the best part
about working with you guys.
349
00:22:28,045 --> 00:22:31,615
It's definitely what we go for here,
the the whole experience.
350
00:22:31,640 --> 00:22:33,120
Well, thank you so much, Shantanu.
351
00:22:33,145 --> 00:22:35,680
Thank you so much.
352
00:22:36,440 --> 00:22:39,020
All right, let's play a game.
353
00:22:39,045 --> 00:22:41,120
A game?
354
00:22:41,440 --> 00:22:42,290
What's the game?
355
00:22:42,320 --> 00:22:47,380
I'm going to play a sound, and it's a
sound that you would hear in our factory,
356
00:22:47,405 --> 00:22:49,655
and you have to guess what the sound is.
357
00:22:49,680 --> 00:22:50,050
Okay.
358
00:22:50,080 --> 00:22:54,855
I haven't won one of these yet,
so let's see how this goes.
359
00:22:54,880 --> 00:22:55,320
Okay.
360
00:22:55,345 --> 00:22:57,855
It can be anything-It's not
another USB stick, is it?
361
00:22:57,880 --> 00:22:58,440
No.
Okay.
362
00:22:58,465 --> 00:23:02,290
It could be anything that you
encounter at the System 76 factory.
363
00:23:02,320 --> 00:23:03,860
All right.
364
00:23:03,885 --> 00:23:07,480
Here we go.
Anything I can encounter.
365
00:23:23,880 --> 00:23:25,140
Any guesses?
366
00:23:25,165 --> 00:23:32,160
Mostly, I just heard Nathaniel and Aaron,
so it feels like when we used to
367
00:23:32,720 --> 00:23:38,500
put the switches and keys in by hand.
368
00:23:38,525 --> 00:23:41,780
I didn't really hear any
machinery in there.
369
00:23:41,805 --> 00:23:44,090
There were some fans, but...
370
00:23:44,120 --> 00:23:45,740
You're totally right.
371
00:23:45,765 --> 00:23:47,290
You win.
Oh, really?
372
00:23:47,320 --> 00:23:48,980
It win.
373
00:23:49,005 --> 00:23:53,180
When we used to install keycaps
by hand on the launch keyboard.
374
00:23:53,205 --> 00:23:54,500
Oh, my God.
I got one.
375
00:23:54,525 --> 00:23:56,020
You win.
376
00:23:56,045 --> 00:23:57,520
Yay.
377
00:23:57,960 --> 00:23:59,330
All right.
Well, that's it.
378
00:23:59,360 --> 00:24:02,980
That's all we We have for you
this month on the transmission log.
379
00:24:03,005 --> 00:24:06,380
So thanks for joining us,
and we'll see you next month.
380
00:24:06,405 --> 00:24:11,220
We will see you next month,
and thanks for listening.
381
00:24:11,245 --> 00:24:15,000
Three, two, one.
382
00:24:15,025 --> 00:24:26,280
This has been the System
383
00:24:26,305 --> 00:24:28,220
76 transmission log.
384
00:24:28,245 --> 00:24:32,460
For more inspiration, check out the
website and follow us on social media.
385
00:24:32,485 --> 00:24:35,560
On your descent back to Earth,
please keep your hands and feet
386
00:24:35,585 --> 00:24:38,500
inside the transport beam at all times.
387
00:24:38,525 --> 00:24:40,800
Captain, sign off in transmission.
00:00:02,440 --> 00:00:06,100
3, 2, 1, 0.
2
00:00:06,125 --> 00:00:17,200
Welcome aboard the System
3
00:00:17,225 --> 00:00:19,100
76 transmission Log.
4
00:00:19,125 --> 00:00:21,540
Our broadcast is about to begin.
5
00:00:21,565 --> 00:00:27,180
This is the latest on System 76
computers, manufacturing, and pop OS.
6
00:00:27,205 --> 00:00:30,260
Now for your in-orbit crew.
7
00:00:30,285 --> 00:00:34,340
Hey, Alex.
8
00:00:34,365 --> 00:00:36,940
Tell me about your experience
at All Things Open.
9
00:00:36,965 --> 00:00:38,340
It was my first time.
10
00:00:38,365 --> 00:00:43,800
I was helping work the booth with Stetson
and Aaron, who are from the sales team
11
00:00:43,825 --> 00:00:46,060
and support team, respectively.
12
00:00:46,085 --> 00:00:47,560
It was pretty fun.
13
00:00:47,585 --> 00:00:49,420
We had a lot of attention at the booth.
14
00:00:49,445 --> 00:00:54,780
We were the only consumer hardware company
that was there that I'm aware of.
15
00:00:54,805 --> 00:00:57,420
We had a lot of excitement.
16
00:00:57,445 --> 00:01:01,300
Ato was raffling off
three of our in laptops.
17
00:01:01,325 --> 00:01:03,400
We got a lot of questions about that.
18
00:01:03,425 --> 00:01:09,560
We were able to give the spiel
about System 76 and our open-source
19
00:01:09,585 --> 00:01:11,540
mission, our Denver manufacturing.
20
00:01:11,565 --> 00:01:17,095
It was great to educate
the enterprise masses about us.
21
00:01:17,120 --> 00:01:17,770
Cool.
22
00:01:17,800 --> 00:01:20,920
I feel like we were here
recording just yesterday.
23
00:01:20,960 --> 00:01:23,060
The month flew by pretty fast.
24
00:01:23,085 --> 00:01:25,290
Yeah, it's almost Thanksgiving.
It's crazy.
25
00:01:25,320 --> 00:01:29,780
Yeah, and when it's close to Thanksgiving,
we celebrate our birthday
26
00:01:29,805 --> 00:01:34,380
because System 76 was born
or founded on November seventh.
27
00:01:34,405 --> 00:01:40,060
We celebrated our 19th birthday
this month, and we had a feast with cake
28
00:01:40,085 --> 00:01:44,460
and presented Carl with a nice canvas
that had a company photo on it.
29
00:01:44,485 --> 00:01:48,490
Now he can see our smiling faces
on the wall every single day.
30
00:01:48,520 --> 00:01:52,570
Yeah, walking into the office
is like walking into grandma's house.
31
00:01:52,600 --> 00:01:56,040
You see your photo from
fifth grade up on the wall,
32
00:01:56,065 --> 00:01:59,210
and you're like, Oh, good times.
33
00:01:59,240 --> 00:02:00,780
Thanks, grandma.
34
00:02:00,805 --> 00:02:01,850
Grandma Carl.
35
00:02:01,880 --> 00:02:06,280
I like when we go on tours and
we stop by the photo wall on the way out,
36
00:02:06,305 --> 00:02:08,450
and I can tell stories of each photo.
37
00:02:08,480 --> 00:02:09,940
That's really fun.
38
00:02:09,965 --> 00:02:13,940
We have some meetups
coming up, too, as well.
39
00:02:13,965 --> 00:02:17,420
We go on factory tours during those.
40
00:02:17,445 --> 00:02:19,610
There's some catered food and drinks.
41
00:02:19,640 --> 00:02:21,170
We play some games.
42
00:02:21,200 --> 00:02:24,980
We get to chat with the community,
chat with locals.
43
00:02:25,005 --> 00:02:28,215
It's a good time.
The next tour, I believe, is in December.
44
00:02:28,240 --> 00:02:28,890
Is that right?
45
00:02:28,920 --> 00:02:33,735
We have our December December
and January meetups posted on meetup.
46
00:02:33,760 --> 00:02:34,500
Com.
47
00:02:34,525 --> 00:02:38,120
What would you say is
your favorite part on the tour?
48
00:02:38,145 --> 00:02:40,260
When we get to the Blue Laser.
49
00:02:40,285 --> 00:02:42,060
Why the Blue Laser?
50
00:02:42,085 --> 00:02:43,880
I can't tell you.
51
00:02:44,560 --> 00:02:45,800
Hang on.
Secrets.
52
00:02:45,825 --> 00:02:49,260
You have to go on the tour to find out.
53
00:02:49,285 --> 00:02:53,280
Okay, well, I'm going to
schedule a tour after this podcast.
54
00:02:53,520 --> 00:02:55,980
How do I schedule a tour
after this podcast?
55
00:02:56,005 --> 00:03:00,015
You have to email me,
emma, E-M-M-M-A@system70.
56
00:03:00,040 --> 00:03:00,220
Com.
6.
57
00:03:00,245 --> 00:03:05,260
Com directly, and you can schedule it, or
you can come on one of our group tours.
58
00:03:05,285 --> 00:03:09,080
I will email you to get those secrets.
59
00:03:09,240 --> 00:03:15,080
Okay.
All right, let's get into hardware news.
60
00:03:16,720 --> 00:03:22,200
We talked in-depth about Thelio ASTRA last
month, and there have been some
61
00:03:22,225 --> 00:03:24,420
benchmarking results released by Foronix.
62
00:03:24,445 --> 00:03:29,610
Those show some promising results,
and you can check it out on foronix.
63
00:03:29,640 --> 00:03:32,100
Com if you want to get some details there.
64
00:03:32,125 --> 00:03:36,780
Yeah, we featured Thelio ASTRA at ATO.
65
00:03:36,805 --> 00:03:42,460
People were excited to talk to us
about our new Arm 64 workstation.
66
00:03:42,485 --> 00:03:45,860
It's an Ampere processor up to 128
67
00:03:45,885 --> 00:03:50,610
cores, up to 6000 data graphics.
68
00:03:50,640 --> 00:03:53,460
So beast of a machine, beautiful looking.
69
00:03:53,485 --> 00:03:57,540
We were right next to the arm booth
at the conference, too.
70
00:03:57,565 --> 00:03:59,880
So that was funny.
71
00:03:59,905 --> 00:04:02,340
Sort of a funny coincidence.
72
00:04:02,365 --> 00:04:03,660
And yeah.
73
00:04:03,685 --> 00:04:06,900
Thaleoastra has really
been making the rounds.
74
00:04:06,925 --> 00:04:12,020
We're going to be at CoupaCon coming up
pretty soon here, and there'll be
75
00:04:12,045 --> 00:04:18,015
four Thaleo Astras at different booths
doing different demos at each booth.
76
00:04:18,040 --> 00:04:18,520
That's cool.
77
00:04:18,545 --> 00:04:23,420
We'll have more details
about that posted on our blog.
78
00:04:23,445 --> 00:04:28,220
Another newcomer
to the Thaleo desktop line is Thaleo
79
00:04:28,245 --> 00:04:32,140
Spark, our AMD-powered starter desktop.
80
00:04:32,165 --> 00:04:36,340
It is an energy-efficient work of art
that fits perfectly on any desk.
81
00:04:36,365 --> 00:04:41,860
You may remember Spark from
the first Thaleo Head music video.
82
00:04:41,885 --> 00:04:43,900
It was on our minds, and it still is.
83
00:04:43,925 --> 00:04:49,660
Now it's an AMD 8000 G-Series system and
84
00:04:49,685 --> 00:04:53,060
lower starting price at, I think, 799.
85
00:04:53,085 --> 00:04:55,400
Beautiful design, good starter desktop.
86
00:04:55,425 --> 00:04:58,820
If you're starting something new
or just want something fast
87
00:04:58,845 --> 00:05:00,980
that fits within budget.
88
00:05:01,005 --> 00:05:04,420
New desktops are great,
but what about desktop environments?
89
00:05:04,445 --> 00:05:05,740
What's new in cosmic?
90
00:05:05,765 --> 00:05:09,060
You got to talk about
the environment, Emma.
91
00:05:09,085 --> 00:05:14,280
So cosmic is a vast
environment of possibility.
92
00:05:14,305 --> 00:05:15,940
We are now in Alpha 3.
93
00:05:15,965 --> 00:05:17,260
Very exciting.
94
00:05:17,285 --> 00:05:21,000
We have some changes to cosmic
files, a lot of smoothing over
95
00:05:21,025 --> 00:05:24,340
with the preview mode and gallery mode.
96
00:05:24,365 --> 00:05:26,980
Preview mode is no longer
going to be on by default,
97
00:05:27,005 --> 00:05:29,860
and there's a new footer widget.
98
00:05:29,885 --> 00:05:34,180
So When you click on a file, the button
isn't going to change locations.
99
00:05:34,205 --> 00:05:35,960
It's just a lot neater.
100
00:05:36,080 --> 00:05:40,960
Then Cosmic Store now has
a made for cosmic section.
101
00:05:40,985 --> 00:05:46,100
You're going to see some nice cosmic apps
made by you guys, mostly.
102
00:05:46,125 --> 00:05:51,080
Things like a tasks manager
for your to-do list, a weather app.
103
00:05:51,105 --> 00:05:52,840
I think I saw a Pokédex in there.
104
00:05:52,865 --> 00:05:56,020
Someone is out there speaking my language.
105
00:05:56,045 --> 00:06:01,080
I'm just excited to see how long that list
is going to be by official release.
106
00:06:01,105 --> 00:06:03,060
It's pretty cool.
107
00:06:03,085 --> 00:06:09,140
Then lastly, we have screen reader
support in initial testable state.
108
00:06:09,165 --> 00:06:12,120
It's not going to be perfect,
but we're interested in hearing
109
00:06:12,145 --> 00:06:17,815
your feedback in the first testing
of accessibility in Cosmic.
110
00:06:17,840 --> 00:06:18,300
Yeah.
111
00:06:18,325 --> 00:06:23,120
If you want to give it a try,
go ahead and pop on over to system76.
112
00:06:23,145 --> 00:06:26,980
Com/cosmic,
and you can download the Alpha ISO.
113
00:06:27,005 --> 00:06:31,655
I can't forget to mention we
have a lot of bug fixes in Alpha 3.
114
00:06:31,680 --> 00:06:32,620
Thanks to you guys.
115
00:06:32,645 --> 00:06:36,060
You're doing great work
on these contributions here.
116
00:06:36,085 --> 00:06:39,700
It's really feeling like
a polished beta to me at this point,
117
00:06:39,725 --> 00:06:41,780
and we appreciate the support.
118
00:06:41,805 --> 00:06:44,680
I'm so excited for our guest this month.
119
00:06:44,705 --> 00:06:47,180
Who is our guest this month?
120
00:06:47,205 --> 00:06:52,600
We have Shantanu from Traslin Robotics
to talk about using System76 hardware
121
00:06:52,625 --> 00:06:56,480
and open source to create cutting edge
robotics solutions for research labs
122
00:06:56,505 --> 00:06:59,060
and educational institutions.
123
00:06:59,085 --> 00:07:02,460
That sounds awesome, and
I think we're going to learn a lot.
124
00:07:02,485 --> 00:07:06,220
So let's get into it.
125
00:07:06,245 --> 00:07:08,680
Thanks for joining us today, Shantanu.
126
00:07:08,705 --> 00:07:11,740
It's very exciting to have you
on our podcast today.
127
00:07:11,765 --> 00:07:15,340
This is Shantanu from Tracin Robotics.
128
00:07:15,365 --> 00:07:19,935
Can you introduce yourself and tell us
a little bit about what you do at Tracin?
129
00:07:19,960 --> 00:07:20,440
Hi.
130
00:07:20,465 --> 00:07:23,480
Thank you so much for having
me on this podcast.
131
00:07:23,680 --> 00:07:27,240
My name is Shantanu, and
I am a machine learning specialist
132
00:07:27,265 --> 00:07:28,740
here at Tracin Robotics.
133
00:07:28,765 --> 00:07:30,860
I work with the Aloha Kids.
134
00:07:30,885 --> 00:07:36,160
We develop Aloha Kids as a research
product for engineers and researchers so
135
00:07:36,185 --> 00:07:42,920
that they can easily set up their
experiment and get up and running with any
136
00:07:42,945 --> 00:07:47,900
task that they have in terms of machine
learning, especially imitation learning.
137
00:07:47,925 --> 00:07:54,280
So my goal here at Trouzen Robotics is
to make it as easy as possible for people
138
00:07:54,305 --> 00:07:57,280
and researchers to get started
with machine learning,
139
00:07:57,305 --> 00:08:02,900
may it be in terms of setting up their
hardware, software and making it
140
00:08:02,925 --> 00:08:08,080
as easy as possible to use so
that they don't have any difficulty
141
00:08:08,105 --> 00:08:12,500
setting up all these critical
functionalities in machine learning.
142
00:08:12,525 --> 00:08:15,260
So, yeah, that's my job here.
143
00:08:15,285 --> 00:08:19,135
Could you talk a little bit
about the Aloha kits and what those are?
144
00:08:19,160 --> 00:08:20,020
Yeah.
145
00:08:20,045 --> 00:08:24,400
So the Aloha kits started as
a collaboration between Stanford
146
00:08:24,425 --> 00:08:26,940
University and Truss and Robotics.
147
00:08:26,965 --> 00:08:33,040
Our researchers at Stanford University
were looking for a low-cost, affordable
148
00:08:33,065 --> 00:08:38,140
product that they can use to develop
machine learning for their research.
149
00:08:38,165 --> 00:08:42,220
So they collaborated with Truss
and Robotics to create these products
150
00:08:42,245 --> 00:08:45,020
that are today known as the Aloha Kids.
151
00:08:45,045 --> 00:08:50,040
Essentially, these are a leader-follower
arm pair in which you will be having
152
00:08:50,065 --> 00:08:54,260
a Widow X arm and a Viper X arm.
153
00:08:54,285 --> 00:08:58,460
It's essentially used for
teaching a robot how to do stuff
154
00:08:58,485 --> 00:09:00,620
using bimanual manipulation.
Education.
155
00:09:00,645 --> 00:09:05,260
As a human, it's easier for us to
look at something and learn it quickly.
156
00:09:05,285 --> 00:09:09,480
Maybe it might be opening
up the door, doing something
157
00:09:09,505 --> 00:09:12,020
as easy as tying a shoelace.
158
00:09:12,045 --> 00:09:15,180
But for robots,
this is a complicated task.
159
00:09:15,205 --> 00:09:19,240
In order to teach a robot how to do stuff,
there are multiple different ways,
160
00:09:19,265 --> 00:09:25,300
like reinforcement learning,
writing heavy code that tells a robot
161
00:09:25,325 --> 00:09:27,580
what traject it has to do.
162
00:09:27,605 --> 00:09:31,800
But as an easier solution, Imitation
163
00:09:31,825 --> 00:09:35,240
learning is something which is a go-to
164
00:09:35,265 --> 00:09:38,740
in terms of machine learning landscape.
165
00:09:38,765 --> 00:09:44,000
What people do is they use the leader arms
to tell the robot how something
166
00:09:44,025 --> 00:09:46,580
has to be done, like tying a shoelace.
167
00:09:46,605 --> 00:09:50,120
Once you collect enough episodes,
like maybe 50 or 100 episodes
168
00:09:50,145 --> 00:09:54,560
of this task, you can train it
using machine learning algorithm.
169
00:09:54,585 --> 00:09:58,660
Initially, it was act, which is
action-chunking transformers.
170
00:09:58,685 --> 00:10:02,640
Once the robot knows how to
solve this problem, it will be able
171
00:10:02,665 --> 00:10:04,980
to do it autonomously without any help.
172
00:10:05,005 --> 00:10:10,520
There are multiple papers
that are published recently in conference
173
00:10:10,545 --> 00:10:12,800
going on at Munich, Germany.
174
00:10:12,825 --> 00:10:17,680
There are papers which are also
deemed as the foundation models
175
00:10:17,705 --> 00:10:23,020
for the growth in the machine learning
landscape that we see right now.
176
00:10:23,045 --> 00:10:28,880
I can keep talking about this
all night, all day, but I guess
177
00:10:28,905 --> 00:10:34,220
that answers your question in terms
of what the Aloha kit can be used for.
178
00:10:34,245 --> 00:10:39,180
It's basically like an all-in-one
kit for machine learning?
179
00:10:39,205 --> 00:10:40,620
Yes.
180
00:10:40,645 --> 00:10:43,215
With the robot arm?
Yes.
181
00:10:43,240 --> 00:10:43,940
Cool.
182
00:10:43,965 --> 00:10:49,260
Some of our listeners might be roboticists
or enthusiasts about robotics.
183
00:10:49,285 --> 00:10:53,060
What in the industry
excites you most about robotics today?
184
00:10:53,085 --> 00:10:59,640
For anyone in the robotics industry
or even those people who are not in
185
00:10:59,665 --> 00:11:03,140
the robotics industry right now, people
know the hype about humanoid robots.
186
00:11:03,165 --> 00:11:08,840
There has been a rise in the companies
that are producing humanoid robots
187
00:11:08,865 --> 00:11:13,700
like Figur, Tesla, and Google.
188
00:11:13,725 --> 00:11:18,480
I know that these It seemed like
the future is near, but the amount of work
189
00:11:18,505 --> 00:11:23,300
that goes behind these robots to perform
simple tasks, the way I said it,
190
00:11:23,325 --> 00:11:27,740
tying a shoelace,
it's top of my mind because we just saw
191
00:11:27,765 --> 00:11:32,660
the release of a paper called as
Aloha Unleashed, which was
192
00:11:32,685 --> 00:11:34,900
published by Google DeepMind.
193
00:11:34,925 --> 00:11:39,740
If you see the core of the paper
is they used the Aloha kit
194
00:11:39,765 --> 00:11:47,100
that we provide, and they achieved
a very simple task of tying a shoelace.
195
00:11:47,125 --> 00:11:50,560
But for a normal person,
this is a very easy task,
196
00:11:50,585 --> 00:11:53,420
and no one knows how complex it is.
197
00:11:53,445 --> 00:11:57,520
But when you read into the paper,
you will see that to train the robot to
198
00:11:57,545 --> 00:12:03,740
do that task, it was so complicated that
it required immense amount of GPU power.
199
00:12:03,765 --> 00:12:07,340
It required immense amount
of training episodes.
200
00:12:07,365 --> 00:12:12,540
It took about 11 days for
it to train to do that simple task.
201
00:12:12,565 --> 00:12:17,960
The most exciting part for me is we
see these robots that are going
202
00:12:17,985 --> 00:12:24,560
to work in the industry, in factories
and other places, and they'll be soon
203
00:12:24,585 --> 00:12:26,220
collaborating with humans.
204
00:12:26,245 --> 00:12:30,160
But the fundamentals, training
these robots to do something, starts
205
00:12:30,185 --> 00:12:33,620
at very basic, that's the Aloha Kids.
206
00:12:33,645 --> 00:12:40,400
That's me being a part of it and me
being able to contribute to the community
207
00:12:40,425 --> 00:12:45,660
in a way where it becomes easier
for people to use these robots
208
00:12:45,685 --> 00:12:47,900
is something very exciting for me.
209
00:12:47,925 --> 00:12:49,020
That's awesome.
210
00:12:49,045 --> 00:12:54,015
Can you tell us a little bit about
the Interbotx open source repositories?
211
00:12:54,040 --> 00:12:54,720
Yeah.
212
00:12:54,745 --> 00:12:57,440
So Interbotx is why we
call it open source.
213
00:12:57,465 --> 00:13:02,240
We don't have any control roll
over the software in terms
214
00:13:02,265 --> 00:13:04,060
of we don't sell the software.
215
00:13:04,085 --> 00:13:06,500
We just are a hardware company.
216
00:13:06,525 --> 00:13:12,360
The reason why we create the software
is that people should not
217
00:13:12,385 --> 00:13:16,520
face any difficulties
while interfacing their robots
218
00:13:16,545 --> 00:13:19,140
with the computers they are going to use.
219
00:13:19,165 --> 00:13:25,040
They need something which is very sleek,
easy to use, and they don't
220
00:13:25,065 --> 00:13:27,940
spend a lot of time debugging stuff.
221
00:13:27,965 --> 00:13:33,520
What intervallex does is it gives you
a middleware framework, which is a layer
222
00:13:33,545 --> 00:13:39,080
of ROSS, which is an existing robotics
operating system framework, which
223
00:13:39,105 --> 00:13:41,220
is widely popular among roboticists.
224
00:13:41,245 --> 00:13:46,320
We use this communication protocol or
the middleware framework,
225
00:13:46,345 --> 00:13:52,080
build upon it using our Inovatex software
and make it as easy as possible for people
226
00:13:52,105 --> 00:13:57,080
to calibrate their robots,
run their robots, debug if anything goes
227
00:13:57,105 --> 00:14:02,640
wrong in terms of maybe servos,
not picking up the current correctly or
228
00:14:02,665 --> 00:14:07,560
the commands correctly or
servos dying in the middle of operation.
229
00:14:07,585 --> 00:14:14,360
We provide elaborate workflows
that will help them to debug anything at
230
00:14:14,385 --> 00:14:21,320
any point in time and get up and running
with their robots in about an hour after
231
00:14:21,345 --> 00:14:24,020
they have set up the robot correctly.
232
00:14:24,045 --> 00:14:26,280
Yeah, that's the software that we offer.
233
00:14:26,305 --> 00:14:27,920
It's pretty easy to use.
234
00:14:27,945 --> 00:14:32,900
It uses something which is called
ROS, which our robotists
235
00:14:32,925 --> 00:14:34,780
are highly familiar with.
236
00:14:34,805 --> 00:14:40,980
We offer C++ as well as Python binding,
so it makes it easier for you as a user
237
00:14:41,005 --> 00:14:44,740
to choose whatever you want.
238
00:14:44,765 --> 00:14:50,040
And because we support Python as
well, people who build upon it can easily
239
00:14:50,065 --> 00:14:53,040
integrate machine learning concepts
or deep learning concepts into
240
00:14:53,065 --> 00:14:57,180
their projects without worrying about
the software that run behind.
241
00:14:57,205 --> 00:15:01,415
It sounds like we're vastly reducing
the barrier entry for robotics here.
242
00:15:01,440 --> 00:15:02,380
That's awesome.
243
00:15:02,405 --> 00:15:06,460
Have you seen any standout projects yet
from the community?
244
00:15:06,485 --> 00:15:08,580
What are your highlights, would you say?
245
00:15:08,605 --> 00:15:13,920
There are multiple things that are going
out in the community right now
246
00:15:13,945 --> 00:15:18,080
because as you know, there is
a conference on reinforcement learning.
247
00:15:18,105 --> 00:15:23,400
That's the highly anticipated conference
in terms of reinforcement learning
248
00:15:23,425 --> 00:15:26,620
at Robotics that's going
on right now at Munich.
249
00:15:26,645 --> 00:15:31,880
When a conference is close by and people
will be releasing all their work
250
00:15:31,905 --> 00:15:33,780
that they have done throughout the years.
251
00:15:33,805 --> 00:15:40,480
We have noticed that at least five
of the top papers in this conference
252
00:15:40,505 --> 00:15:46,020
use the AloHA kits, use our hardware,
our software to some extent.
253
00:15:46,045 --> 00:15:51,360
There are big companies like
Google DeepMind just least the paper
254
00:15:51,385 --> 00:15:55,060
called as Aloha Unleashed,
which I was talking about before this.
255
00:15:55,085 --> 00:16:00,160
We also have seen a company called as Pai,
256
00:16:00,185 --> 00:16:03,000
which is physical intelligence, that just
257
00:16:03,025 --> 00:16:07,820
announced its own foundation model
that uses Aloha Kit again.
258
00:16:07,845 --> 00:16:14,600
We have seen, I guess, there are
three papers from a researcher from UC,
259
00:16:14,625 --> 00:16:20,980
Berkeley, along with researchers
that are leading Google DeepMind as well.
260
00:16:21,005 --> 00:16:26,000
The core of this paper is to
train something, to train a robot to do a
261
00:16:26,025 --> 00:16:31,620
very simple task, you need a lot of data,
and you to collect a lot of data sets.
262
00:16:31,645 --> 00:16:36,360
To reduce the amount of data
that you collect, people have come
263
00:16:36,385 --> 00:16:41,740
up with projects that let you
use any data on any other robot.
264
00:16:41,765 --> 00:16:46,340
It's something called
as a cross-training platforms.
265
00:16:46,365 --> 00:16:50,520
There are a lot of very interesting
papers, interesting researches
266
00:16:50,545 --> 00:16:53,660
that are going on in the industry.
267
00:16:53,685 --> 00:16:58,320
In terms of what I was saying,
the research papers that are coming out,
268
00:16:58,345 --> 00:17:02,520
and they are coming out so frequently,
and I guess to some extent,
269
00:17:02,545 --> 00:17:06,940
it's because of the way we are reducing
the barrier to entry with our products,
270
00:17:06,965 --> 00:17:10,920
making it as easy as possible
to set up your robots.
271
00:17:10,945 --> 00:17:15,440
We also provide with a community
that helps you publish your data sets,
272
00:17:15,465 --> 00:17:20,800
publish your trained models to a platform
where you can pick the data set
273
00:17:20,825 --> 00:17:23,810
from and train your own models.
274
00:17:23,840 --> 00:17:28,520
So what happens is researchers now
don't have to spend their time
275
00:17:28,545 --> 00:17:31,180
on setting up a hardware platform.
276
00:17:31,205 --> 00:17:36,740
They can invest more time
in coming up with new ideas, new models,
277
00:17:36,765 --> 00:17:41,560
and scale up their research
to an extent that they don't have to
278
00:17:41,585 --> 00:17:43,530
worry about anything in the background.
279
00:17:43,560 --> 00:17:48,420
I guess that's where cross
and robotics helps the researchers.
280
00:17:48,445 --> 00:17:53,900
That's our motto to help researchers
or help innovators innovate.
281
00:17:53,925 --> 00:17:57,880
That goes a lot with our model
of engineering the computer, too,
282
00:17:57,905 --> 00:18:02,480
is just trying to get all the fussing
around out of the way,
283
00:18:02,505 --> 00:18:07,050
all the installing drivers,
moving windows around, things like that,
284
00:18:07,080 --> 00:18:11,700
to just get people focused on what
they actually want to work on.
285
00:18:11,725 --> 00:18:18,040
Now, you're using the Serval workstation
for the Aloha stationery, too.
286
00:18:18,065 --> 00:18:20,810
What led to the decision for the Serval?
287
00:18:20,840 --> 00:18:28,220
As you said, your model
and our model is similar.
288
00:18:28,245 --> 00:18:32,250
The reason is when I I
started using this workstation.
289
00:18:32,280 --> 00:18:36,880
To begin with any project, you need
to install a lot of dependencies, a lot
290
00:18:36,905 --> 00:18:42,010
of different things to interface
your robots with your workstation.
291
00:18:42,040 --> 00:18:43,880
I worked with different...
292
00:18:43,905 --> 00:18:48,480
When I was a student, I worked
with different laptops, and I have seen
293
00:18:48,505 --> 00:18:52,680
other people work with different laptops
and the difficulties that they face while
294
00:18:52,705 --> 00:18:57,460
installing simple things like Linux,
Cuda drivers, or any other thing.
295
00:18:57,485 --> 00:19:03,200
But with this workstation, when I started,
it was within an hour, I was able
296
00:19:03,225 --> 00:19:05,740
to set up all the necessary dependencies.
297
00:19:05,765 --> 00:19:09,840
It took an hour because downloading
NVIDIA code takes an hour.
298
00:19:09,865 --> 00:19:12,460
It wasn't the system
doing something weird.
299
00:19:12,485 --> 00:19:16,600
But yeah, it was a very smooth and error
It was a free experience because most
300
00:19:16,625 --> 00:19:22,380
of the hardware is compatible
with the NVIDIA drivers that you need.
301
00:19:22,405 --> 00:19:27,520
It was as simple as running a command line
argument, like seeing a pseudo
302
00:19:27,545 --> 00:19:29,570
install NVIDIA server.
303
00:19:29,600 --> 00:19:31,330
So It was that simple.
304
00:19:31,360 --> 00:19:35,620
The Linux operation and
the Linux setup was very easy.
305
00:19:35,645 --> 00:19:42,380
I guess that's the reason why we are
partners with System76 to get our laptops
306
00:19:42,405 --> 00:19:47,200
so that the users that we give these
laptops to, the researchers,
307
00:19:47,225 --> 00:19:52,280
don't have to face any other difficulties
that you might face when you get a laptop
308
00:19:52,305 --> 00:19:57,380
which doesn't support the necessary
drivers or the hardware interfaces.
309
00:19:57,405 --> 00:20:01,940
I guess that's the core reason
why we go with system 76.
310
00:20:01,965 --> 00:20:04,620
Awesome.
We're happy to hear that it's going well.
311
00:20:04,645 --> 00:20:08,520
I heard the MeerCAT was
going to be in use soon.
312
00:20:08,545 --> 00:20:12,090
Do you know about that project coming up?
Yes.
313
00:20:12,120 --> 00:20:14,700
The mini PC?
Yeah.
314
00:20:14,725 --> 00:20:18,960
As you see, the Aloha stationery
is a bimanual robotic system.
315
00:20:18,985 --> 00:20:24,400
We had noticed that in the research
industry, people are also
316
00:20:24,425 --> 00:20:26,760
working with single-arm devices.
317
00:20:26,785 --> 00:20:30,810
They don't need two arms
to achieve or complete a task.
318
00:20:30,840 --> 00:20:34,720
So what we noticed was people
required one R interface.
319
00:20:34,745 --> 00:20:40,440
And if we are providing a cheaper
alternative to the Aloha, we also need
320
00:20:40,465 --> 00:20:45,480
something that is as compact as possible
so that people can move around the robots
321
00:20:45,505 --> 00:20:47,040
easily, record their episodes.
322
00:20:47,065 --> 00:20:51,740
And we were looking for a system
that can be useful for this.
323
00:20:51,765 --> 00:20:58,480
And then we saw the System76 MeerKAT
Mini, and it had super processing powers
324
00:20:58,505 --> 00:20:59,700
and interfaces.
325
00:20:59,725 --> 00:21:04,440
And Again, because it is System76,
Linux drivers, Cuda drivers,
326
00:21:04,465 --> 00:21:07,330
and everything was super easy to install.
327
00:21:07,360 --> 00:21:12,220
Given that and how portable
it is, we are soon going to launch
328
00:21:12,245 --> 00:21:14,620
a product called as Aloha Solo.
329
00:21:14,645 --> 00:21:17,460
That that will be using the Meer cat mini.
330
00:21:17,485 --> 00:21:20,380
That's really awesome.
Do you have a timeline for that yet?
331
00:21:20,405 --> 00:21:23,695
It should be out by December first,
officially.
332
00:21:23,720 --> 00:21:24,570
Okay, cool.
333
00:21:24,600 --> 00:21:28,620
Are there any other things that
you'd like to add about trust in robotics
334
00:21:28,645 --> 00:21:31,770
or about the robotics Linux industry
and Linux in general?
335
00:21:31,800 --> 00:21:34,460
Yeah, I'm happy about this partnership.
336
00:21:34,485 --> 00:21:38,000
We are so satisfied that we
have partners like you.
337
00:21:38,025 --> 00:21:42,320
The great parts being the customer
support, the customization that you guys
338
00:21:42,345 --> 00:21:48,280
provide, and how easy it is to get
something resolved in terms of anything
339
00:21:48,305 --> 00:21:53,400
that we need apart from the product,
the support in terms of getting a product
340
00:21:53,425 --> 00:22:00,080
in time, sending it to the customer,
having this beautiful a stream line
341
00:22:00,105 --> 00:22:03,880
process where the customer orders
the product, you get it in time,
342
00:22:03,905 --> 00:22:07,140
and we are able to get it
to the customer in time.
343
00:22:07,165 --> 00:22:11,480
There is absolutely no error in between
because we are able to get the software
344
00:22:11,505 --> 00:22:17,120
and the hardware integrated with System76
with so much ease that it feels
345
00:22:17,145 --> 00:22:19,775
like it's one product itself.
346
00:22:19,800 --> 00:22:22,320
It doesn't look like there are
two different companies working on it.
347
00:22:22,345 --> 00:22:24,780
It just feels like it's
a one full company.
348
00:22:24,805 --> 00:22:28,020
I guess that's the best part
about working with you guys.
349
00:22:28,045 --> 00:22:31,615
It's definitely what we go for here,
the the whole experience.
350
00:22:31,640 --> 00:22:33,120
Well, thank you so much, Shantanu.
351
00:22:33,145 --> 00:22:35,680
Thank you so much.
352
00:22:36,440 --> 00:22:39,020
All right, let's play a game.
353
00:22:39,045 --> 00:22:41,120
A game?
354
00:22:41,440 --> 00:22:42,290
What's the game?
355
00:22:42,320 --> 00:22:47,380
I'm going to play a sound, and it's a
sound that you would hear in our factory,
356
00:22:47,405 --> 00:22:49,655
and you have to guess what the sound is.
357
00:22:49,680 --> 00:22:50,050
Okay.
358
00:22:50,080 --> 00:22:54,855
I haven't won one of these yet,
so let's see how this goes.
359
00:22:54,880 --> 00:22:55,320
Okay.
360
00:22:55,345 --> 00:22:57,855
It can be anything-It's not
another USB stick, is it?
361
00:22:57,880 --> 00:22:58,440
No.
Okay.
362
00:22:58,465 --> 00:23:02,290
It could be anything that you
encounter at the System 76 factory.
363
00:23:02,320 --> 00:23:03,860
All right.
364
00:23:03,885 --> 00:23:07,480
Here we go.
Anything I can encounter.
365
00:23:23,880 --> 00:23:25,140
Any guesses?
366
00:23:25,165 --> 00:23:32,160
Mostly, I just heard Nathaniel and Aaron,
so it feels like when we used to
367
00:23:32,720 --> 00:23:38,500
put the switches and keys in by hand.
368
00:23:38,525 --> 00:23:41,780
I didn't really hear any
machinery in there.
369
00:23:41,805 --> 00:23:44,090
There were some fans, but...
370
00:23:44,120 --> 00:23:45,740
You're totally right.
371
00:23:45,765 --> 00:23:47,290
You win.
Oh, really?
372
00:23:47,320 --> 00:23:48,980
It win.
373
00:23:49,005 --> 00:23:53,180
When we used to install keycaps
by hand on the launch keyboard.
374
00:23:53,205 --> 00:23:54,500
Oh, my God.
I got one.
375
00:23:54,525 --> 00:23:56,020
You win.
376
00:23:56,045 --> 00:23:57,520
Yay.
377
00:23:57,960 --> 00:23:59,330
All right.
Well, that's it.
378
00:23:59,360 --> 00:24:02,980
That's all we We have for you
this month on the transmission log.
379
00:24:03,005 --> 00:24:06,380
So thanks for joining us,
and we'll see you next month.
380
00:24:06,405 --> 00:24:11,220
We will see you next month,
and thanks for listening.
381
00:24:11,245 --> 00:24:15,000
Three, two, one.
382
00:24:15,025 --> 00:24:26,280
This has been the System
383
00:24:26,305 --> 00:24:28,220
76 transmission log.
384
00:24:28,245 --> 00:24:32,460
For more inspiration, check out the
website and follow us on social media.
385
00:24:32,485 --> 00:24:35,560
On your descent back to Earth,
please keep your hands and feet
386
00:24:35,585 --> 00:24:38,500
inside the transport beam at all times.
387
00:24:38,525 --> 00:24:40,800
Captain, sign off in transmission.
