AI Video Learning Paths
Tell us what you want to learn. Get an ordered path through videos recommended by the Hacker News community.
Example learning path
From linear algebra to your first neural network, then on to the ideas behind large language models. Basic Python is useful for the coding stage.
Understand vectors, matrices and the geometry behind a learning algorithm.
3Blue1Brown
Why this video: Start with a visual introduction to the linear algebra used throughout machine learning.
HN repeatedly recommends the series as the best intuition-first, visual introduction to linear algebra and a benchmark for clear math explanation.
11 distinct positive recommenders in the source snapshotOriginal HN discussion
MIT OpenCourseWare
Why this video: Follow the intuition with a university lecture that develops the mathematical foundations.
It is HN's default recommendation for learning linear algebra, praised for Strang's teaching and for building real understanding useful in machine learning.
11 distinct positive recommenders in the source snapshotOriginal HN discussionDraw a matrix transformation and explain what it does to a vector.
Connect neurons, gradient descent and backpropagation before using a framework.
3Blue1Brown
Why this video: A visual explanation of what a neural network computes, suited to a first encounter.
HN treats it as the standard first stop for learning neural networks, praised for making the math intuitive.
16 distinct positive recommenders in the source snapshotOriginal HN discussion
3Blue1Brown
Why this video: Build on the first video by seeing how gradient descent turns a network into a learning system.
HN recommends the series as a very accessible way to understand the math behind machine learning and LLMs.
4 distinct positive recommenders in the source snapshotOriginal HN discussionExplain how changing a weight changes a network’s prediction and loss.
Move from diagrams to code by implementing automatic differentiation and a small network.
Andrej Karpathy
Why this video: Karpathy builds micrograd in code, connecting the earlier backpropagation concepts to an implementation.
HN treats it as the best starting point for really understanding backprop and neural nets, because Karpathy explains the why with intuition and builds everything in code.
17 distinct positive recommenders in the source snapshotOriginal HN discussion
Devoxx
Why this video: Extend that foundation with a practical introduction to TensorFlow and deep learning.
Repeatedly called an excellent, practical introduction that covers a lot in two hours and explains issues academic textbooks skip.
6 distinct positive recommenders in the source snapshotOriginal HN discussionRebuild a tiny network, then change its training data and compare the loss.
Understand transformers and how large language models are trained and used.
3Blue1Brown
Why this video: Use a visual introduction to transformers to bridge neural-network basics and language models.
HN treats it as the go-to visual introduction to transformers, and many say it finally made attention and embeddings click.
13 distinct positive recommenders in the source snapshotOriginal HN discussion
Andrej Karpathy
Why this video: Finish with Karpathy’s overview of LLM training, capabilities and limitations.
HN recommends it as the dense, approachable starting point for understanding LLMs and often cites it for specific claims about scaling and training.
9 distinct positive recommenders in the source snapshotOriginal HN discussionSketch how tokens, attention and next-token prediction fit together.
Videos and HN context come from the public HackerTube catalog. The learning sequence and practice prompts are added by this site. HN discussion links let you review the original recommendations.
Your completion progress is saved in this browser.
Searching for where to learn machine learning with videos can leave you with dozens of disconnected tabs. AI Video Learning Paths connects the HackerTube archive into a sequence: understand the prerequisites, build the core concepts, work through an example, then explore further. Each stage includes a learning objective, relevant videos, an explanation of the match and a small practice prompt. Review the original Hacker News discussions, open each video on its watch page, and track your progress at your own pace.
How to use the video archive