HackerTube. Ideas worth watching.
Discover talks, tutorials and unexpected rabbit holes recommended on Hacker News. Browse by topic, follow the discussion, and save something worth your time.
Neural networks,
built from the ground up.
Andrej Karpathy's first Neural Networks: Zero to Hero lecture, which builds micrograd, a tiny autograd engine, from scratch in Python. It explains backpropagation and neural network training step by step, assuming only basic Python and some high-school calculus.
Read the discussion ↗Recommended again and again
Lasting recommendations from the Hacker News archive
Ruby Conf 12 - Y Not- Adventures in Functional Programming by Jim Weirich
Indistinguishable From Magic: Manufacturing Modern Computer Chips
How to Speak
The spelled-out intro to neural networks and backpropagation: building micrograd
8 Bit & '8 Bitish' Graphics-Outside the Box
How Python was Shaped by leaky Internals, Armin Ronacher, Flask framework
How The Economic Machine Works by Ray Dalio
In Search Of A Flat Earth
1. Introduction to Human Behavioral Biology
AT&T Archives: Similiarities of Wave Behavior (Bonus Edition)
Around The Corner - How Differential Steering Works (1937)
Find your next rabbit hole on HackerTube
A good recommendation deserves more than a passing link. This library brings together conference talks, visual explanations, engineering deep dives and documentaries that people have shared on Hacker News. Start with a subject you know, or follow a connection into something entirely new.
Recommendations with a trail
Every video has a page with a summary, its topic and the original discussion. Follow the Hacker News links to see the context for yourself. The recommender count describes people in the source dataset, rather than likes or views on this website.
Our discovery order uses the source dataset’s recommendation score, which reflects recurring recommendations over time. It is a useful starting point for exploration, not an objective measure of a video’s quality or accuracy.
A video library for curious people
Explore functional programming, visual mathematics, machine learning, science and the craft of making things. Browse all topics for focused collections, or search by title, creator and subject across the complete library.
Make room for a longer idea
Save a talk to Watch later and come back when you have time. Your saved list and recently opened videos stay in this browser, with no account required. Click Play to load the YouTube player; there is also a direct YouTube link when an embed is unavailable.
This is an independent edition built from the open HackerTube dataset. Summaries and labels were generated by AI and may contain mistakes. Original videos remain with their creators; attribution and data sources are available for every reader.
HackerTube, explained
What is HackerTube?
HackerTube is a browsable collection of YouTube videos linked and recommended in Hacker News discussions. This independent .run edition uses the original project’s open dataset, with its own interface and static pages.
Can I watch videos here for free?
Browsing and saving videos do not require an account or payment. Videos play through YouTube’s privacy-enhanced embed after you press Play. Availability, regional restrictions and advertising are controlled by YouTube and the video owner.
Where do the recommendations come from?
The October 31, 2026 public export contains 6,454 video records and 39,834 labelled Hacker News comments. It is an archive snapshot, not a live feed. Visit each video page for original discussion links, and the About page for licensing details.
Can I upload a video or sync my saved list?
This edition is a discovery library. It does not accept uploads or provide accounts. Watch later and history are stored on this device in browser storage and do not sync between devices.











