What is HackerTube?
HackerTube is a video discovery project built around YouTube links shared in Hacker News discussions. Andrej Karpathy introduced it on October 31, 2026, linking to hackertube.org. Talkshelf uses its exported catalog to help you find videos, understand the discussion around them, and study selected videos with notes and practice materials.
Why Karpathy built HackerTube
In his launch post, Karpathy describes wanting a better way to find videos than his YouTube recommendations. HackerTube turns links from Hacker News comments into a browsable catalog. He also made a data export available for people building their own versions.
The useful question for a viewer is what brought a video into the conversation. Was someone recommending an explanation, answering a technical question, or pointing to a particular example? Reading the linked discussion adds context to the title and thumbnail.
How videos are selected
The catalog comes from discussions on Hacker News. Records distinguish recommendations from other references to a video. Recommendation scores reflect those discussion records, rather than YouTube view counts.
Use the score as a starting point, then check the video's topic, level and discussion. A widely recommended lecture may assume background you have not studied yet; a less frequently mentioned tutorial may answer exactly the question you have today.
On Talkshelf, each video includes an overview and its available HN comments. You can browse categories, search for a topic or creator, and filter for videos with study materials.
HackerTube and Talkshelf: which site does what?
| Site | What you will find |
|---|---|
| hackertube.org | The HackerTube project linked in Karpathy's launch post. |
| hackertube.video | A site using the HackerTube name, with video rankings, topics, channels and HN discussion context. |
| Talkshelf | A video library using the exported catalog, with learning guides and selected packs of notes, quizzes, flashcards and PPT outlines. |
Talkshelf's material tags show which resources are available for a video. Open a video to watch it, read its overview and comments, and find the relevant study tools below the player.
Where to start watching
For neural networks and language models, the Andrej Karpathy guide offers starting points for both conceptual learning and coding. Start with an overview if the terminology is new, or follow the coding path if you are ready to work through the examples.
To find something you can study with supporting resources, browse videos with materials. Each video page lists its available resources, so you can choose a video and a study format together.