The Best Videos to Understand AI, Picked by 20 Years of Hacker News Comments

A one-hour, general-audience introduction to large language models: what they are, how they're trained, how they compare to an operating system, and the security challenges that come with them, including jailbreaks.

Hacker News treats this as the dense but approachable starting point for understanding LLMs, and it is the single most mentioned video on this whole list, linked in 20 comments.

A three-hour-plus walkthrough of the full training stack behind a ChatGPT-style model, from pretraining through post-training, plus mental models for why these systems sometimes hallucinate and how to work around that.

Hacker News's most frequent recommendation for a clear foundation in how LLMs are actually built, with no heavy math required; it shows up in 18 comments.

The first chapter of a deep-learning series, explaining what neurons and layers actually are and the basic math underneath a neural network, illustrated with handwritten digit recognition.

Commenters call it the standard first stop for learning how neural networks work at all, praised for making the underlying math feel intuitive; it's referenced in 19 comments.

A visual explanation of how transformers and GPT-style language models work, following data through tokens, embeddings, and attention step by step.

Hacker News treats this as the go-to visual introduction to transformers; many comments say it's the video that finally made attention and embeddings click, with 16 mentions.

A closer, visual walkthrough of the attention mechanism itself: self-attention, multiple heads, and cross-attention, as chapter 6 of the same deep-learning series.

Recommended on Hacker News specifically for readers who found text-based explanations of attention hard to follow, with 8 mentions in the data.

A remastered, general-audience introduction to AI safety, adapted from a talk given in London, laying out why advanced AI could go wrong and answering the most common objections to that idea.

The default link Hacker News gives newcomers who want an approachable, even entertaining, introduction to AI risk and alignment; 7 comments point here first.

An explanation of instrumental convergence: why an AI system pursuing almost any goal would tend to also pursue sub-goals like self-preservation and gathering resources, whether or not anyone programmed it to.

Repeatedly recommended as the clearest explanation of why a smart AI system might behave badly by default, with no malice required; it has 8 mentions.

An explanation of the orthogonality thesis: how a system can be highly capable while pursuing goals that look strange or alien to humans, separating how smart something is from what it actually wants.

Hacker News uses it as the standard short answer to 'won't a smart AI just understand human values,' and it's one of the more-cited entries on this list at 12 mentions.

Tristan Harris and Aza Raskin lay out why they believe current AI capabilities carry real societal risk, arguing companies are racing to deploy faster than safety work can keep up, delivered just before GPT-4's public launch.

Frequently recommended as the single talk to watch to get up to speed on AI risk, especially by people who had already seen the same team's earlier work on social media; it has 20 mentions, tying it for the most-cited video on this list.

A fictional short film, backed by AI researcher Stuart Russell, depicting swarms of cheap autonomous drones using facial recognition to carry out targeted killings, made to argue for a ban on autonomous weapons.

Hacker News's standard reference whenever drone swarms, autonomous weapons, or military AI come up in the news, with 20 mentions over the years.