2 HN users recommended this demo3 months2014–2016score 3.5
Years with an HN recommendation, 2007–2026
HN commenters use it as the canonical demo of deep reinforcement learning, noting that watching it teaches more than an article does.
A short smartphone-recorded clip from a 2014 event in Paris where DeepMind CEO Demis Hassabis presents an AI learning to play old Atari games. The system learns from only the video signal and the score. Audio quality is poor.
“you learn more about the tech from a youtube video than from an article” (HN, 2014)
Thanks for the correction. The "arbitrary" qualifier is not in TFA, but (as, indeed, you said) that's the point of the demo, e.g.: [this video] Note that they're using just the video signal from the game as input. It's really a sad comment on the state of reporting at MIT Tech Review that you learn more about the tech from a youtube video than from an article. (My complaint is not with the DeepMind people, it's with the article, which should put the work in context.)
DeepMind demo learning games from only the video signal
In psychology, this is called "Reinforcement"[1] and that goes back to Skinner (1938) and Thorndike (1911). In AI research its called "Reinforcement learning"[2]. One of the earliest papers relating reinforcement learning to AI was “Theory of Neural-Analog Reinforcement Systems”[3] by Minsky, 1954 A cool reinforcement learning demo is the DeepMind Atari video from last year[4] [1] [link] [2] [link] [3] [link] [4] [this video] (this is awesome)
Creating a GAI which can learn how to act human, on its own, in virtual worlds and in robots with realistic bodies isn't serious research? [this video]
I was also under the misapprehension that deep learning is just about classification, but that isn't true. Yes, reinforcement learning is the path to general intelligence, and the deep learning community is showing impressive progress on that front as well. The Deepmind demo [1] and the recent robotics work at Berkeley[2] are good examples. Thanks for the link to Stan Franklin's work. I'm glad to hear there is work to integrate the two approaches. [1] [this video] [2] [link]
Example of impressive deep reinforcement learning progress
AGI is the idea of creating general intelligence. Which almost certainly requires some reinforcement learning (the ways dogs and humans learn, by having a goal and an internal reward system). Deepmind's Atari player is based on deep reinforcement learning, where increasing the score represents a reward: [this video] I used to believe the moving goalpost idea, that AI is anything that isn't yet possible. I now disagree. I saved myself a lot of energy by avoiding entirely the question "what counts as AI?" by switching to the question, "what counts as AGI?" which is a term with a clearer threshold: [link]
Video list and Hacker News comment labels from HackerTube (prompted by @karpathy), released under ODC-By 1.0, built on the open-index/hacker-news archive. Hacker News data is provided by Y Combinator; comments remain the work of their authors. Video titles, channel names and thumbnails belong to their creators and are shown via YouTube. HackerTube Watch is an unofficial edition. Data export of 2026-10-31.
Plays via YouTube's privacy-enhanced embed
6 Hacker News comments
From the HN archive; up to 20 kept per video, favoring different commenters. Recommendations first.
Citations, criticism and jokes
Read more