episode
stringlengths
45
100
text
stringlengths
1
528
timestamp_link
stringlengths
56
56
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
and then we sent out all the emails to launch the,
https://karpathy.ai/lexicap/0014-large.html#00:20:45.120
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
launch the company and do the press thing. And then, you know,
https://karpathy.ai/lexicap/0014-large.html#00:20:47.920
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
we weren't quite actually there. And then
https://karpathy.ai/lexicap/0014-large.html#00:20:51.120
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
we thought, oh, well, you know, they're not going to announce it until
https://karpathy.ai/lexicap/0014-large.html#00:20:54.640
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
maybe 10 a.m. the next morning. And it's, I don't know, it's 5 p.m. now. So
https://karpathy.ai/lexicap/0014-large.html#00:20:58.640
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
how many hours do we have left? What is that? Like, you know, 17 hours to go.
https://karpathy.ai/lexicap/0014-large.html#00:21:02.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And, and that was, that was going to be fine.
https://karpathy.ai/lexicap/0014-large.html#00:21:06.080
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
Was the problem obvious? Did you understand
https://karpathy.ai/lexicap/0014-large.html#00:21:10.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
what could possibly, like, how complicated was the system at that point?
https://karpathy.ai/lexicap/0014-large.html#00:21:12.240
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
It was, it was pretty messy. So to get a live video feed that looked decent
https://karpathy.ai/lexicap/0014-large.html#00:21:16.400
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
working from anywhere in San Francisco, I put together this system where we had
https://karpathy.ai/lexicap/0014-large.html#00:21:22.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
like three or four cell phone data modems and
https://karpathy.ai/lexicap/0014-large.html#00:21:27.040
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
they were, like, we take the video stream and,
https://karpathy.ai/lexicap/0014-large.html#00:21:29.920
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
you know, sort of spray it across these three or four modems and then try to
https://karpathy.ai/lexicap/0014-large.html#00:21:32.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
catch all the packets on the other side, you know, with unreliable cell phone
https://karpathy.ai/lexicap/0014-large.html#00:21:36.000
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
networks. It's pretty low level networking.
https://karpathy.ai/lexicap/0014-large.html#00:21:39.040
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
Yeah, and putting these, like, you know, sort of
https://karpathy.ai/lexicap/0014-large.html#00:21:41.040
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
protocols on top of all that to, to reassemble and reorder the packets and
https://karpathy.ai/lexicap/0014-large.html#00:21:44.160
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
have time buffers and error correction and all that kind of stuff.
https://karpathy.ai/lexicap/0014-large.html#00:21:47.520
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And the night before it was just staticky. Every once in a while the image
https://karpathy.ai/lexicap/0014-large.html#00:21:50.880
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
would, would go to staticky and there would be this horrible,
https://karpathy.ai/lexicap/0014-large.html#00:21:55.600
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
like, screeching audio noise because the audio was also corrupted.
https://karpathy.ai/lexicap/0014-large.html#00:21:58.560
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And this would happen, like, every five to ten minutes or so and it was
https://karpathy.ai/lexicap/0014-large.html#00:22:01.920
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
a really, you know, off putting to the viewers.
https://karpathy.ai/lexicap/0014-large.html#00:22:05.280
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
How do you tackle that problem? What was the, uh, you're just freaking out behind a
https://karpathy.ai/lexicap/0014-large.html#00:22:08.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
computer. There's, are there other, other folks working
https://karpathy.ai/lexicap/0014-large.html#00:22:12.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
on this problem? Like, were you behind a whiteboard? Were you doing, uh,
https://karpathy.ai/lexicap/0014-large.html#00:22:16.160
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
Yeah, it was a little, it was a little, yeah, it's a little lonely because there's four of us
https://karpathy.ai/lexicap/0014-large.html#00:22:19.360
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
working on the company and only two people really wrote code.
https://karpathy.ai/lexicap/0014-large.html#00:22:23.680
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And Emmett wrote the website and the chat system and I wrote the
https://karpathy.ai/lexicap/0014-large.html#00:22:26.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
software for this video streaming device and video server.
https://karpathy.ai/lexicap/0014-large.html#00:22:30.080
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And so, you know, it's my sole responsibility to figure that out.
https://karpathy.ai/lexicap/0014-large.html#00:22:34.080
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And I think, I think it's those, you know, setting,
https://karpathy.ai/lexicap/0014-large.html#00:22:37.200
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
setting deadlines, trying to move quickly and everything where you're in that
https://karpathy.ai/lexicap/0014-large.html#00:22:40.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
moment of intense pressure that sometimes people do their
https://karpathy.ai/lexicap/0014-large.html#00:22:42.960
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
best and most interesting work. And so even though that was a terrible moment,
https://karpathy.ai/lexicap/0014-large.html#00:22:45.520
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
I look back on it fondly because that's like, you know, that's one of those
https://karpathy.ai/lexicap/0014-large.html#00:22:48.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
character defining moments, I think. So in 2013, October, you founded
https://karpathy.ai/lexicap/0014-large.html#00:22:51.360
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
Cruise Automation. Yeah. So progressing forward, another
https://karpathy.ai/lexicap/0014-large.html#00:22:58.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
exceptionally successful company was acquired by GM in 16
https://karpathy.ai/lexicap/0014-large.html#00:23:02.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
for $1 billion. But in October 2013, what was on your mind?
https://karpathy.ai/lexicap/0014-large.html#00:23:06.800
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
What was the plan? How does one seriously start to tackle
https://karpathy.ai/lexicap/0014-large.html#00:23:14.000
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
one of the hardest robotics, most important impact for robotics
https://karpathy.ai/lexicap/0014-large.html#00:23:19.760
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
problems of our age? After going through Twitch, Twitch was,
https://karpathy.ai/lexicap/0014-large.html#00:23:23.200
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
was, and is today, pretty successful. But the, the work was,
https://karpathy.ai/lexicap/0014-large.html#00:23:27.760
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
the result was entertainment, mostly. Like, the better the product was,
https://karpathy.ai/lexicap/0014-large.html#00:23:35.120
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
the more we would entertain people and then, you know, make money on the ad
https://karpathy.ai/lexicap/0014-large.html#00:23:38.560
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
revenues and other things. And that was, that was a good thing. It
https://karpathy.ai/lexicap/0014-large.html#00:23:42.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
felt, felt good to entertain people. But I figured like, you know, what is really
https://karpathy.ai/lexicap/0014-large.html#00:23:44.960
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
the point of becoming a really good engineer and
https://karpathy.ai/lexicap/0014-large.html#00:23:48.400
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
developing these skills other than, you know, my own enjoyment? And I
https://karpathy.ai/lexicap/0014-large.html#00:23:51.120
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
realized I wanted something that scratched more of an existential
https://karpathy.ai/lexicap/0014-large.html#00:23:54.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
itch, like something that, that truly matters. And so I
https://karpathy.ai/lexicap/0014-large.html#00:23:56.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
basically made this list of requirements for a new, if I was going to
https://karpathy.ai/lexicap/0014-large.html#00:23:59.920
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
do another company, and the one thing I knew in the back of
https://karpathy.ai/lexicap/0014-large.html#00:24:05.120
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
my head that Twitch took like eight years to become successful.
https://karpathy.ai/lexicap/0014-large.html#00:24:07.600
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And so whatever I do, I better be willing to commit, you know, at least 10 years
https://karpathy.ai/lexicap/0014-large.html#00:24:12.160
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
to something. And when you think about things from that perspective,
https://karpathy.ai/lexicap/0014-large.html#00:24:16.160
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
you certainly, I think, raise the bar on what you choose to work on. So for me,
https://karpathy.ai/lexicap/0014-large.html#00:24:20.240
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
the three things were it had to be something where the technology
https://karpathy.ai/lexicap/0014-large.html#00:24:23.520
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
itself determines the success of the product,
https://karpathy.ai/lexicap/0014-large.html#00:24:26.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
like hard, really juicy technology problems, because that's what
https://karpathy.ai/lexicap/0014-large.html#00:24:28.880
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
motivates me. And then it had to have a direct and positive impact on society in
https://karpathy.ai/lexicap/0014-large.html#00:24:32.400
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
some way. So an example would be like, you know,
https://karpathy.ai/lexicap/0014-large.html#00:24:37.040
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
health care, self driving cars, because they save lives, other things where
https://karpathy.ai/lexicap/0014-large.html#00:24:39.120
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
there's a clear connection to somehow improving other people's lives.
https://karpathy.ai/lexicap/0014-large.html#00:24:42.000
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And the last one is it had to be a big business, because
https://karpathy.ai/lexicap/0014-large.html#00:24:45.040
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
for the positive impact to matter, it's got to be a large scale.
https://karpathy.ai/lexicap/0014-large.html#00:24:48.240
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And I was thinking about that for a while, and I made like, I tried
https://karpathy.ai/lexicap/0014-large.html#00:24:51.200
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
writing a Gmail clone and looked at some other ideas.
https://karpathy.ai/lexicap/0014-large.html#00:24:54.640
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And then it just sort of light bulb went off, like self driving cars, like that
https://karpathy.ai/lexicap/0014-large.html#00:24:57.520
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
was the most fun I had ever had in college working on that.
https://karpathy.ai/lexicap/0014-large.html#00:25:00.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And like, well, what's the state of the technology? It's been 10 years.
https://karpathy.ai/lexicap/0014-large.html#00:25:03.920
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
Maybe times have changed, and maybe now is the time to make this work.
https://karpathy.ai/lexicap/0014-large.html#00:25:07.280
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And I poked around and looked at, the only other thing out there really at the
https://karpathy.ai/lexicap/0014-large.html#00:25:10.640
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
time was the Google self driving car project.
https://karpathy.ai/lexicap/0014-large.html#00:25:14.000
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And I thought, surely there's a way to, you know, have an entrepreneur mindset
https://karpathy.ai/lexicap/0014-large.html#00:25:16.560
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
and sort of solve the minimum viable product here.
https://karpathy.ai/lexicap/0014-large.html#00:25:20.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
And so I just took the plunge right then and there and said, this is something I
https://karpathy.ai/lexicap/0014-large.html#00:25:23.440
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
know I can commit 10 years to. It's the probably the greatest
https://karpathy.ai/lexicap/0014-large.html#00:25:26.400
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
applied AI problem of our generation. And if it works, it's going to be both a
https://karpathy.ai/lexicap/0014-large.html#00:25:29.600
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
huge business and therefore like, probably the most positive impact I can
https://karpathy.ai/lexicap/0014-large.html#00:25:33.440
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
possibly have on the world. So after that light bulb went off, I went
https://karpathy.ai/lexicap/0014-large.html#00:25:37.120
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
all in on cruise immediately and got to work.
https://karpathy.ai/lexicap/0014-large.html#00:25:41.680
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
Did you have an idea how to solve this problem? Which aspect of the problem to
https://karpathy.ai/lexicap/0014-large.html#00:25:45.520
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
solve? You know, slow, like we just had Oliver
https://karpathy.ai/lexicap/0014-large.html#00:25:48.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
from Voyage here, slow moving retirement communities,
https://karpathy.ai/lexicap/0014-large.html#00:25:52.560
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
urban driving, highway driving. Did you have, like, did you have a vision of the
https://karpathy.ai/lexicap/0014-large.html#00:25:56.400
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
city of the future where, you know, the transportation is
https://karpathy.ai/lexicap/0014-large.html#00:26:00.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
largely automated, that kind of thing? Or was it sort of
https://karpathy.ai/lexicap/0014-large.html#00:26:05.200
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
more fuzzy and gray area than that? My analysis of the situation is that
https://karpathy.ai/lexicap/0014-large.html#00:26:09.760
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
Google is putting a lot, had been putting a lot of money into that project. They
https://karpathy.ai/lexicap/0014-large.html#00:26:15.680
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
had a lot more resources. And so, and they still hadn't cracked
https://karpathy.ai/lexicap/0014-large.html#00:26:19.200
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
the fully driverless car. You know, this is 2013, I guess.
https://karpathy.ai/lexicap/0014-large.html#00:26:23.680
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
So I thought, what can I do to sort of go from zero to,
https://karpathy.ai/lexicap/0014-large.html#00:26:29.520
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
you know, significant scale so I can actually solve the real problem, which is
https://karpathy.ai/lexicap/0014-large.html#00:26:34.320
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
the driverless cars. And I thought, here's the strategy. We'll
https://karpathy.ai/lexicap/0014-large.html#00:26:37.520
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
start by doing a really simple problem or solving a
https://karpathy.ai/lexicap/0014-large.html#00:26:40.720
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
really simple problem that creates value for people. So,
https://karpathy.ai/lexicap/0014-large.html#00:26:44.480
Kyle Vogt: Cruise Automation | Lex Fridman Podcast #14
eventually ended up deciding on automating highway driving,
https://karpathy.ai/lexicap/0014-large.html#00:26:49.040