Home › Topics › AI, Machine Learning & Data
LLMs & generative AI: videos Hacker News recommends
Large language models and generative AI: GPT, ChatGPT, transformers, attention, diffusion, image generation, prompting, AI coding assistants, deepfakes, synthetic media, AI products and demos. Below: every video in this topic that at least three HN users linked with real commentary, in at least three different months (21 demos, 12 explainers, 11 interviews), best-scored first, with the comments on each video page.
- 13 HN recommenders 10 months 2024–2026HN treats it as the go-to visual introduction to transformers, and many say it finally made attention and embeddings click.
- 10 HN recommenders 8 months 2025–2026HN's most frequent recommendation for a clear, approachable foundation in how LLMs work without heavy math.
- 14 HN recommenders 8 months 2023–2025It is the go-to recommendation for learning how transformers and LLMs actually work, praised for its clarity and Karpathy's teaching.
- 8 HN recommenders 8 months 2023–2024Repeatedly named the best practical tutorial for using LLMs, with a well-liked fine-tuning section.
- 9 HN recommenders 6 months 2023–2026HN recommends it as the dense, approachable starting point for understanding LLMs and often cites it for specific claims about scaling and training.
- 7 HN recommenders 4 months 2023–2026S Sebastien Bubeck · lectureHN recommends it as a thought-provoking talk from a researcher with early GPT-4 access, especially for the unicorn degrading under safety tuning.
- 5 HN recommenders 5 months 2024–2025Recommended as the best deep dive on tokenizers, worth the two hours for anyone working with LLMs.
- 5 HN recommenders 4 months 2024–2026HN treats it as the go-to visual explanation of attention, recommended to people who find text-based transformer explainers hard to follow.
- 11 HN recommenders 4 months 2023–2024C Center for Humane Technology · talkFrequently recommended as the overview to get up to speed on AI risk, especially by people who valued The Social Dilemma.
- 4 HN recommenders 4 months 2020–2023S Seattle Applied Deep Learning · lectureRecommended as a clear explanation of why RNNs fall short and how transformers solve it.
- 4 HN recommenders 4 months 2020–2025HN recommends it as an accessible, clear walkthrough of the transformer paper for people with some AI background.
- 3 HN recommenders 4 months 2021–2025S sentdex · demoCited as an early, striking demo of neural networks acting as game engines and world models.
- 3 HN recommenders 3 months 2023–2026HN commenters point to it as an impressive example of what AI animation could do and of VFX artists adopting AI in their workflows.
- 3 HN recommenders 3 months 2025HN readers recommend it as an intuitive explanation of neural networks that shifts how people think about AI and language.
- 6 HN recommenders 3 months 2023–2025Recommended as the clearest overview of how LLM assistants are built and how to use them well.
- 3 HN recommenders 3 months 2024S Sasha Rush · lectureHN commenters recommend it for building intuition about state space models and the tradeoffs between transformer and SSM layers.
- 3 HN recommenders 3 months 2022–2026HN points to it as the definitive account of the GPT-4chan experiment, used in debates on LLM bias, truthfulness and bots.
- 3 HN recommenders 3 months 2023P Pascal Poupart · lectureHN recommends it to newcomers as a clear explanation of attention, including the query/key/value intuition.
- 2 HN recommenders 2 months 2024HN recommends it as a long but worthwhile way to demystify how language models work technically.
- 4 HN recommenders 3 months 2023–2025HN cites it for the argument that we cannot tell whether AI is early or late on its sigmoid curve until hindsight.
- 3 HN recommenders 2 months 2023–2024G Graphics in 5 Minutes · explainerHN recommends it as a quick, non-dumbed-down way to get up to speed on how LLMs are built.
- 4 HN recommenders 3 months 2023–2026HN points to it as the main place to hear Lattner's motivations and design ideas for Mojo, often in Mojo threads.
- 3 HN recommenders 3 months 2022–2025Pointed to as an accessible early explanation of transformers and attention for newcomers.
- 2 HN recommenders 2 months 2024HN cites it for the argument that LLMs capture language form but not meaning, and for how the brain's language network works.
- 2 HN recommenders 2 months 2023–2026A Acly · demoHN recommends it as a quick way to see how the Krita AI plugin works and how it matches Photoshop's generative features.
- 2 HN recommenders 2 months 2021–2025HN points to it as an early, impressive demonstration that LLMs could write working code.
- 2 HN recommenders 2 months 2025A Alex O'Connor · explainerHN treats it as the origin of the full-wine-glass meme and an illustration of how models lack concepts missing from their training data.
- 2 HN recommenders 2 months 2024–2025Recommended as a well-made, digestible introduction to scaling laws and why models cluster along a similar efficiency frontier.
- 1 HN recommenders 2 months 2020–2026S Sassy Justice · comedy videoBrought up in every deepfake thread as a funny and technically convincing early example of synthetic media used for satire.
- 2 HN recommenders 2 months 2025–2026Pointed to as a non-promotional, realistic long session of an experienced engineer working with AI on non-trivial code.
- 2 HN recommenders 2 months 2026Shared as a funny example of jailbreaking AI voice bots, and as the source of the bolognese prompt-injection phrase.
- 2 HN recommenders 2 months 2016–2024A Adobe Creative Cloud · demoHN cites it as the early example of voice cloning and audio deepfakes, with some recommending Peele's reaction and others calling the demo fake or never shipped.
- 2 HN recommenders 2 months 2021–2025I ISL and Collaborators · demoUsed as a go-to example of neural networks enhancing game rendering, referenced in later AI-graphics discussions.
- 2 HN recommenders 2 months 2023HN points to it as the accessible explanation of why glitch tokens exist whenever tokenizers or odd GPT behavior come up.
- 3 HN recommenders 2 months 2022–2023Commenters cite it during Stable Diffusion release debates as a source for Mostaque's vision and Stability's principles.
- 2 HN recommenders 2 months 2019–2022C Ctrl Shift Face · comedy videoPointed to as one of the most impressive early deepfakes, where a good impersonation makes the effect convincing and funny.
- 2 HN recommenders 2 months 2023–2026Cited in AI-and-education threads as the optimistic case and demo for AI tutoring, and as a source on Khan Academy's LLM approach.
- 2 HN recommenders 2 months 2023U UC Berkeley EECS · talkRecommended in LLM threads as an authoritative explanation from a ChatGPT architect of why models hallucinate and why imitation is not enough.
- 2 HN recommenders 2 months 2024–2025HN commenters point to it as the best first-hand account of Cursor's technical internals, straight from its founders.
- 3 HN recommenders 2 months 2023C Cognitive Revolution "How AI Changes Everything" · interviewCited as a firsthand account of the raw, unfiltered GPT-4 before safety tuning, especially the part after 45 minutes.
- 3 HN recommenders 2 months 2023C CBS Mornings · interviewRecommended as a more nuanced look at Hinton's views than headlines, especially his argument that LLMs are more than auto-complete.
- 2 HN recommenders 2 months 2025–2026A Alex Kantrowitz · podcast episodePointed to as LeCun's argument for the limits of LLMs and the need for world models.
- 4 HN recommenders 2 months 2023HN commenters send it to people confused about ChatGPT's quirks, explaining how human-feedback training makes it a pleasing yes-man that bullshits.
- 2 HN recommenders 2 months 2025–2026HN points to it in AI music and data-poisoning threads as the source for this technique.
- 1 HN recommenders 1 months 2024Cited for LeCun's views on LLM limits and alternative approaches to intelligence.
- 1 HN recommenders 1 months 2023–2025A Acly · demoHN points to it as a demonstration of local, controllable AI painting and a way to learn the workflow.
- 1 HN recommenders 1 months 2023–2024Mostly cited as the canonical statement of AI doomer views, with some recommending it and many criticizing it as irresponsible or wrong.
- 1 HN recommenders 1 months 2023B Bankless · interviewHN links it as a balanced, full-length interview with a leading alignment researcher on AI risk.
- 1 HN recommenders 1 months 2025HN mostly cites it as a source for Karpathy's framing of programming in English and his generate-then-verify workflow, with a few critics finding it hype.
- 1 HN recommenders 1 months 2016–2017HN cites it as an early, quirky example of machine-generated screenplays and what LSTMs could do.
- 1 HN recommenders 1 months 2023–2026Linked as the approachable explanation whenever HN discusses detecting or watermarking AI-generated text.
- 1 HN recommenders 1 months 2025–2026V Vibe Coding With Steve and Gene · demoHN shares it as a fun demo of an LLM driving a live Emacs process and orchestrating its long-standing ecosystem.
- 1 HN recommenders 1 months 2023–2024HN cites it as an example of temporally stable AI video stylization built from ordinary production gear.
- 1 HN recommenders 1 months 2025–2026HN uses it as a clear explainer for diffusion mechanics, such as negative prompts and faster sampling, and for joint text-image encoders.
- 1 HN recommenders 1 months 2023–2025HN refers back to it as the primary source for what GPT-4 could do at launch, especially the sketch-to-website and multimodal demos.
- 1 HN recommenders 1 months 2023–2024HN cites it as an example of what a very small crew can make with generative AI tools, while noting how much manual work it still takes.
- 1 HN recommenders 1 months 2025–2026HN cites it for insight into Google researchers' views on AI progress and publishing, though one commenter finds the interview weak.
- 1 HN recommenders 1 months 2021–2024S Shamook · demoHN cites it as an example of deepfakes producing a more convincing CGI human than the studio's own effort.
- cited 6× 0 months 2025–2026HN commenters cite it as the serious skeptic's case against LLMs reaching general intelligence, usually in debates about whether scaling is enough.
- cited 4× 0 months 2023–2025T Tom Gally · talkCited as an expert's first-hand evidence that GPT translation is near human quality for distant language pairs.
- cited 4× 0 months 2023–2025T Tom Gally · commentaryHN links it as an informed view from a working translator on using GPT for translation.
- cited 3× 0 months 2023–2026S Simons Institute for the Theory of Computing · talkCited in discussions of learning as compression and of why neural networks generalize.
- cited 3× 0 months 2018–2020Cited as a concrete early example that AI-faked video was already convincing, in discussions of misinformation and trust in media.
- cited 3× 0 months 2023A Anyscale · interviewCited as a source for the fact that OpenAI uses Ray for training, and for Brockman's views on developer speed.
- cited 3× 0 months 2021–2022D DevOps Directive · demoCited as evidence that AI can solve interview-style algorithm puzzles, in threads questioning the value of Leetcode.
- cited 7× 0 months 2016–2025S Sony CSL (Paris) · music videoCited as an early example of AI-composed music, usually with the caveat that humans did much of the production.
- cited 8× 0 months 2023HN mostly cites it for specific Altman quotes, such as models as reasoning engines rather than databases, while many commenters found the interview shallow or insincere.
- cited 3× 0 months 2023–2025P PromptJungle · tutorialHN cites it as a process example of how such AI-generated videos are made, including on a phone.
- Ilya Sutskever (OpenAI Chief Scientist) — Why next-token prediction could surpass human intelligencecited 8× 0 months 2023–2025HN commenters cite it mainly for Sutskever's statements that training data is far from exhausted.
- cited 5× 0 months 2020–2023HN commenters cite it as a concise explanation of GPT-3's scaling jump and surprising abilities at its release.
- cited 6× 0 months 2018–2023B BuzzFeedVideo · demoUsed as an early, widely seen example of deepfakes in discussions of fake video and its risks.
- cited 3× 0 months 2023–2025B Bloc · demoPointed to as an early example of LLM-powered NPC dialog in games.
- cited 4× 0 months 2023–2025D demonflyingfox · comedy videoHN cites it as the original of the Balenciaga AI meme family and as an early viral AI mashup moment.
- cited 3× 0 months 2021–2022R River's Educational Channel · demoCited as an early, amusing example of LLM-driven command-line toys, with commenters sharing similar experiments.
- Sam Altman: OpenAI, GPT-5, Sora, Board Saga, Elon Musk, Ilya, Power & AGI | Lex Fridman Podcast #419cited 6× 0 months 2024–2026HN uses it as a primary source for what Altman said on topics like programming changing, GPT-5's expected leap and startup structure, and often posts timestamps or paraphrases.
- cited 4× 0 months 2018–2019T Tero Karras FI · demoCited as an example of how realistic GAN-generated faces had become, and of what a well-structured latent space can do.
- cited 19× 0 months 2016–2021M Matthias Niessner · demoHN uses it as a reference point in deepfake and fake-news discussions, showing that convincing real-time video manipulation was already possible in 2016.
- cited 12× 0 months 2025–2026A AI Engineer · talkCited in AI-coding debates as one of the largest data-driven studies, notably for finding no correlation between self-assessed and measured productivity.
- cited 5× 0 months 2024–2025Mostly cited as a source for statements by Anthropic leaders, such as views on AI progress and on sycophancy.
Questions
What is the most recommended llms & generative ai video on Hacker News?
Transformers, the tech behind LLMs | Deep Learning Chapter 5 by 3Blue1Brown: HN treats it as the go-to visual introduction to transformers, and many say it finally made attention and embeddings click.
How is this llms & generative ai list ranked?
By HackerTube's score: the number of distinct months in which an HN comment recommended the video, plus half a point for every month with a strong recommendation. One viral thread counts less than steady recommendations over years.
Where do the videos come from?
From every Hacker News comment between 2007 and 2026 that linked a YouTube video, labeled by an AI model as a recommendation, a citation, a joke or a criticism. The list is published by HackerTube under ODC-By 1.0; this edition adds indexable pages and the comment context.
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.