episode
stringlengths
45
100
text
stringlengths
1
528
timestamp_link
stringlengths
56
56
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
I mean, I'm not an expert in anthropology,
https://karpathy.ai/lexicap/0010-large.html#00:39:44.660
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
but it seems like we're very kind of good
https://karpathy.ai/lexicap/0010-large.html#00:39:46.660
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
within our own tribe, but need to be taught
https://karpathy.ai/lexicap/0010-large.html#00:39:49.260
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
to be nice to other tribes.
https://karpathy.ai/lexicap/0010-large.html#00:39:52.860
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Well, if you look at Steven Pinker,
https://karpathy.ai/lexicap/0010-large.html#00:39:54.660
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
he highlights this pretty nicely in
https://karpathy.ai/lexicap/0010-large.html#00:39:56.300
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Better Angels of Our Nature,
https://karpathy.ai/lexicap/0010-large.html#00:40:00.740
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
where he talks about violence decreasing over time
https://karpathy.ai/lexicap/0010-large.html#00:40:02.340
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
consistently.
https://karpathy.ai/lexicap/0010-large.html#00:40:04.980
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
So whatever tension, whatever teams we pick,
https://karpathy.ai/lexicap/0010-large.html#00:40:05.800
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
it seems that the long arc of history
https://karpathy.ai/lexicap/0010-large.html#00:40:08.340
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
goes towards us getting along more and more.
https://karpathy.ai/lexicap/0010-large.html#00:40:11.100
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
So. I hope so.
https://karpathy.ai/lexicap/0010-large.html#00:40:14.220
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
So do you think that, do you think it's possible
https://karpathy.ai/lexicap/0010-large.html#00:40:15.420
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
to teach RL based robots this kind of kindness,
https://karpathy.ai/lexicap/0010-large.html#00:40:20.620
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
this kind of ability to interact with humans,
https://karpathy.ai/lexicap/0010-large.html#00:40:26.180
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
this kind of policy, even to, let me ask a fun one.
https://karpathy.ai/lexicap/0010-large.html#00:40:28.380
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Do you think it's possible to teach RL based robot
https://karpathy.ai/lexicap/0010-large.html#00:40:32.860
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
to love a human being and to inspire that human
https://karpathy.ai/lexicap/0010-large.html#00:40:35.140
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
to love the robot back?
https://karpathy.ai/lexicap/0010-large.html#00:40:38.580
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
So to like RL based algorithm that leads to a happy marriage.
https://karpathy.ai/lexicap/0010-large.html#00:40:40.020
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
That's an interesting question.
https://karpathy.ai/lexicap/0010-large.html#00:40:47.540
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Maybe I'll answer it with another question, right?
https://karpathy.ai/lexicap/0010-large.html#00:40:48.860
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Because I mean, but I'll come back to it.
https://karpathy.ai/lexicap/0010-large.html#00:40:52.820
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
So another question you can have is okay.
https://karpathy.ai/lexicap/0010-large.html#00:40:56.700
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
I mean, how close does some people's happiness get
https://karpathy.ai/lexicap/0010-large.html#00:40:58.940
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
from interacting with just a really nice dog?
https://karpathy.ai/lexicap/0010-large.html#00:41:03.560
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Like, I mean, dogs, you come home,
https://karpathy.ai/lexicap/0010-large.html#00:41:07.620
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
that's what dogs do.
https://karpathy.ai/lexicap/0010-large.html#00:41:09.900
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
They greet you, they're excited,
https://karpathy.ai/lexicap/0010-large.html#00:41:10.740
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
makes you happy when you come home to your dog.
https://karpathy.ai/lexicap/0010-large.html#00:41:12.660
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
You're just like, okay, this is exciting.
https://karpathy.ai/lexicap/0010-large.html#00:41:14.700
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
They're always happy when I'm here.
https://karpathy.ai/lexicap/0010-large.html#00:41:16.460
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
And if they don't greet you, cause maybe whatever,
https://karpathy.ai/lexicap/0010-large.html#00:41:18.340
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
your partner took them on a trip or something,
https://karpathy.ai/lexicap/0010-large.html#00:41:21.300
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
you might not be nearly as happy when you get home, right?
https://karpathy.ai/lexicap/0010-large.html#00:41:23.540
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
And so the kind of, it seems like the level of reasoning
https://karpathy.ai/lexicap/0010-large.html#00:41:26.100
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
a dog has is pretty sophisticated,
https://karpathy.ai/lexicap/0010-large.html#00:41:30.260
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
but then it's still not yet at the level of human reasoning.
https://karpathy.ai/lexicap/0010-large.html#00:41:32.200
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
And so it seems like we don't even need to achieve
https://karpathy.ai/lexicap/0010-large.html#00:41:35.660
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
human level reasoning to get like very strong affection
https://karpathy.ai/lexicap/0010-large.html#00:41:37.840
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
with humans.
https://karpathy.ai/lexicap/0010-large.html#00:41:40.460
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
And so my thinking is why not, right?
https://karpathy.ai/lexicap/0010-large.html#00:41:41.700
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Why couldn't, with an AI, couldn't we achieve
https://karpathy.ai/lexicap/0010-large.html#00:41:44.220
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
the kind of level of affection that humans feel
https://karpathy.ai/lexicap/0010-large.html#00:41:47.140
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
among each other or with friendly animals and so forth?
https://karpathy.ai/lexicap/0010-large.html#00:41:51.460
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
So question, is it a good thing for us or not?
https://karpathy.ai/lexicap/0010-large.html#00:41:57.480
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
That's another thing, right?
https://karpathy.ai/lexicap/0010-large.html#00:41:59.740
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Because I mean, but I don't see why not.
https://karpathy.ai/lexicap/0010-large.html#00:42:01.380
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Why not, yeah, so Elon Musk says love is the answer.
https://karpathy.ai/lexicap/0010-large.html#00:42:05.980
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Maybe he should say love is the objective function
https://karpathy.ai/lexicap/0010-large.html#00:42:09.020
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
and then RL is the answer, right?
https://karpathy.ai/lexicap/0010-large.html#00:42:12.660
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Well, maybe.
https://karpathy.ai/lexicap/0010-large.html#00:42:14.700
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Oh, Peter, thank you so much.
https://karpathy.ai/lexicap/0010-large.html#00:42:17.660
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
I don't want to take up more of your time.
https://karpathy.ai/lexicap/0010-large.html#00:42:18.880
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Thank you so much for talking today.
https://karpathy.ai/lexicap/0010-large.html#00:42:20.260
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Well, thanks for coming by.
https://karpathy.ai/lexicap/0010-large.html#00:42:21.900
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Great to have you visit.
https://karpathy.ai/lexicap/0010-large.html#00:42:23.500
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
The following is a conversation with Jürgen Schmidhuber.
https://karpathy.ai/lexicap/0011-large.html#00:00:00.000
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
He's the co director of the CS Swiss AI Lab
https://karpathy.ai/lexicap/0011-large.html#00:00:03.520
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
and a co creator of long short term memory networks.
https://karpathy.ai/lexicap/0011-large.html#00:00:06.320
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
LSDMs are used in billions of devices today
https://karpathy.ai/lexicap/0011-large.html#00:00:10.360
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
for speech recognition, translation, and much more.
https://karpathy.ai/lexicap/0011-large.html#00:00:13.720
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
Over 30 years, he has proposed a lot of interesting
https://karpathy.ai/lexicap/0011-large.html#00:00:17.400
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
out of the box ideas on meta learning, adversarial networks,
https://karpathy.ai/lexicap/0011-large.html#00:00:20.800
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
computer vision, and even a formal theory of quote,
https://karpathy.ai/lexicap/0011-large.html#00:00:24.800
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
creativity, curiosity, and fun.
https://karpathy.ai/lexicap/0011-large.html#00:00:28.720
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
This conversation is part of the MIT course
https://karpathy.ai/lexicap/0011-large.html#00:00:32.360
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
on artificial general intelligence
https://karpathy.ai/lexicap/0011-large.html#00:00:34.920
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
and the artificial intelligence podcast.
https://karpathy.ai/lexicap/0011-large.html#00:00:36.560
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
If you enjoy it, subscribe on YouTube, iTunes,
https://karpathy.ai/lexicap/0011-large.html#00:00:38.840
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
or simply connect with me on Twitter
https://karpathy.ai/lexicap/0011-large.html#00:00:41.960
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
at Lex Friedman spelled F R I D.
https://karpathy.ai/lexicap/0011-large.html#00:00:43.960
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
And now here's my conversation with Jürgen Schmidhuber.
https://karpathy.ai/lexicap/0011-large.html#00:00:47.280
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
Early on you dreamed of AI systems
https://karpathy.ai/lexicap/0011-large.html#00:00:53.080
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
that self improve recursively.
https://karpathy.ai/lexicap/0011-large.html#00:00:55.640
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
When was that dream born?
https://karpathy.ai/lexicap/0011-large.html#00:00:58.680
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
When I was a baby.
https://karpathy.ai/lexicap/0011-large.html#00:01:01.800
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
No, that's not true.
https://karpathy.ai/lexicap/0011-large.html#00:01:03.160
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
When I was a teenager.
https://karpathy.ai/lexicap/0011-large.html#00:01:04.240
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
And what was the catalyst for that birth?
https://karpathy.ai/lexicap/0011-large.html#00:01:06.520
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
What was the thing that first inspired you?
https://karpathy.ai/lexicap/0011-large.html#00:01:09.680
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
When I was a boy, I was thinking about what to do in my life
https://karpathy.ai/lexicap/0011-large.html#00:01:12.920
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
and then I thought the most exciting thing
https://karpathy.ai/lexicap/0011-large.html#00:01:20.200
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
is to solve the riddles of the universe.
https://karpathy.ai/lexicap/0011-large.html#00:01:23.880
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
And that means you have to become a physicist.
https://karpathy.ai/lexicap/0011-large.html#00:01:28.200
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
However, then I realized that there's something even grander.
https://karpathy.ai/lexicap/0011-large.html#00:01:30.920
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
You can try to build a machine
https://karpathy.ai/lexicap/0011-large.html#00:01:35.840
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
that isn't really a machine any longer
https://karpathy.ai/lexicap/0011-large.html#00:01:39.880
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
that learns to become a much better physicist
https://karpathy.ai/lexicap/0011-large.html#00:01:42.120
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
than I could ever hope to be.
https://karpathy.ai/lexicap/0011-large.html#00:01:44.520
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
And that's how I thought maybe I can multiply
https://karpathy.ai/lexicap/0011-large.html#00:01:47.080
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
my tiny little bit of creativity into infinity.
https://karpathy.ai/lexicap/0011-large.html#00:01:50.320
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
But ultimately that creativity will be multiplied
https://karpathy.ai/lexicap/0011-large.html#00:01:54.520
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
to understand the universe around us.
https://karpathy.ai/lexicap/0011-large.html#00:01:57.280
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
That's the curiosity for that mystery that drove you.
https://karpathy.ai/lexicap/0011-large.html#00:01:59.280
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
Yes, so if you can build a machine
https://karpathy.ai/lexicap/0011-large.html#00:02:05.800
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
that learns to solve more and more complex problems
https://karpathy.ai/lexicap/0011-large.html#00:02:08.440
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
and more and more general problem solver
https://karpathy.ai/lexicap/0011-large.html#00:02:13.880
Juergen Schmidhuber: Godel Machines, Meta-Learning, and LSTMs | Lex Fridman Podcast #11
then you basically have solved all the problems,
https://karpathy.ai/lexicap/0011-large.html#00:02:16.880