Deep Dive into LLMs like ChatGPT
What it is
Andrej Karpathy's three-hour-plus general-audience deep dive into how ChatGPT-style LLMs are built, covering the full training stack from pretraining to post-training. It also gives mental models for their quirks, like hallucination, and how to use them well.
Why Hacker News recommends it
HN's most frequent recommendation for a clear, approachable foundation in how LLMs work without heavy math.
- 10 people recommended it
- 8 different months
- 6 strong recommendations
- first 2025, last 2026
What HN commenters said
“He goes over a ton of different stuff related to model training, but makes it very easy to understand.”
Long but easy-to-follow overview of LLM training.
InvidFlower, 2025-03 · Strong recommendation · in “Ladder: Self-improving LLMs through recursive problem decomposition”
“the first thing I would recommend is to watch Karpathy's 'deep dive into LLMs' video”
Gives instinctive understanding of LLMs and saves time.
GodelNumbering, 2025-04 · Strong recommendation · in “Show HN: I made a free tool that analyzes SEC filings and posts detailed reports”
“Anyway, I really like's Karpathy's video because he's very good at explaining LLMs at every level.”
Approachable deep dive into how LLMs work.
runjake, 2025-10 · Strong recommendation · in “Ask HN: Build Your Own LLM?”
“Karpathy's excellent "deep dive into LLMs like chatgpt" video[0] explains and demonstrates this aspect pretty well”
Explains and demonstrates LLM stochastic nature.
pengaru, 2025-11 · Strong recommendation · in “LLMs are bullshitters. But that doesn't mean they're not useful”
“I highly recommend Andrej Karpathy's Deep Dive into LLMs like ChatGPT”
General intro making sense of the tech without heavy math.
romanhn, 2026-02 · Strong recommendation · in “Ask HN: What would you recommend a vibe coder learn about how all this works?”
“I highly recommend Karpathy’s intro to LLMs videos on YouTube.”
Worthwhile even with basic LLM understanding.
etothet, 2026-03 · Strong recommendation · in “Microgpt”
“Karpathy has a good intro video on this.”
Good intro on base training versus post-training.
wzdd, 2025-03 · Recommendation · in “Tracing the thoughts of a large language model”
“There is a very good intro on how LLM work by Kharpaty.”
Very good intro to how LLMs work.
highhedgehog, 2025-03 · Recommendation · in “Ask HN: If you had 100 hours to learn something new, what would you learn?”
Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.
Topics: llm training, chatgpt, pretraining, hallucination, tokenization, andrej karpathy
Also recommended by the same people
explainerTransformers, the tech behind LLMs | Deep Learning Chapter 53Blue1Brown · 13 HN recommenders · 2024–2026
demoThe effects of DBS on the motor symptoms of Parkinson's DiseaseAndrew Johnson · 5 HN recommenders · 2013–2016
talkDavid Deutsch: A new way to explain explanationTED · 2 HN recommenders · 2022–2026
talkClassic Game Postmortem - Another WorldGDC Festival of Gaming · 4 HN recommenders · 2020–2023
talkClassic Game Postmortem: Alone in the DarkGDC Festival of Gaming · 3 HN recommenders · 2019–2023
comedyGeorge Carlin - Germs, Immune SystemGeen Naam · 2 HN recommenders · 2018–2022
documentaryThis Is What Winning Looks Like (Full Length)VICE · 7 HN recommenders · 2014–2023
talkDr. Matthew Phillips - 'Metabolic Strategies as Therapies in Cancer and Neurodegeneration'Low Carb Down Under · 2 HN recommenders · 2022–2023