Top all-time · #595
Andrej Karpathy
Personal blog · karpathy.aiaiprogramming
Andrej Karpathy's personal blog at karpathy.ai covers ai and programming. Since 2020, 103 different Hacker News readers have linked it 117 times, most often “Neural Networks: Zero to Hero”. No feed was found, so read it on the site itself.
Often recommended · 19 of 40 sampled comments recommend it?
Why HN links it
HN readers send people to Karpathy's zero-to-hero course, a hands-on series for learning neural networks and LLMs from scratch. Most links are plain citations.
- Recommended for · 19 comments
- Praised for hands-on, no-metaphor teaching of neural nets; Triplebyte analysis called fascinating.
Summarised by Claude from a sample of 40 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Most-linked posts?
- Neural Networks: Zero to Hero68 people · 81 mentions · first linked 2023-01-11
- US Job Market Visualizer – Andrej Karpathy4 people · 4 mentions · first linked 2026-03-14
- Licklider 19603 people · 3 mentions · first linked 2023-12-27
- /3 people · 3 mentions · first linked 2020-08-09
- I Love Calculator3 people · 3 mentions · first linked 2024-09-28
- Microgpt.py – train and inference a GPT in pure, dependency-free Python3 people · 3 mentions · first linked 2026-02-12
- Founding Fathers on Today's America3 people · 3 mentions · first linked 2024-12-17
- HN Time Capsule2 people · 2 mentions · first linked 2025-12-10
- /hncapsule/hall-of-fame.html2 people · 2 mentions · first linked 2025-12-10
- Lexicap: Lex Fridman Podcast Whisper Captions by Andrej Karpathy2 people · 2 mentions · first linked 2022-09-26
What Hacker News says?
This is a good one. https://karpathy.ai/zero-to-hero.html
Comment linking /zero-to-hero.html
on “Andrej Karpathy – It will take a decade to work through the issues with agents”Submitted US Job Market Visualizer – Andrej Karpathy as a story · 133 points
HN discussionI bet you we are all wrong and some random person is going to vibe code himself into something none of us expected. I half kid, if none of you have see it, highly suggest https://karpathy.ai/zero-to-hero.html
Comment linking /zero-to-hero.html
on “Andrej Karpathy – It will take a decade to work through the issues with agents”Submitted Lexicap: Lex Fridman Podcast Whisper Captions by Andrej Karpathy as a story · 42 points
HN discussionIf you are starting from scratch? https://karpathy.ai/zero-to-hero.html Do those, then come back to this and it will make more sense.
Comment linking /zero-to-hero.html
on “Llm.c – LLM training in simple, pure C/CUDA”Submitted Neural Networks: Zero to Hero as a story · 21 points
HN discussionNeural Networks: from zero to hero https://karpathy.ai/zero-to-hero.html
Comment linking /zero-to-hero.html
on “Hi everyone yes, I left OpenAI yesterday”Submitted HN Time Capsule as a story · 7 points
HN discussionThe most recent is this, which I believe was made after he left Tesla: https://github.com/karpathy/nanoGPT And it's accompanying video series: https://karpathy.ai/zero-to-hero.html Another example (although I honestly don't remember if he made this one between jobs) is: https://github.com/karpathy/micrograd
Comment linking /zero-to-hero.html
on “Hi everyone yes, I left OpenAI yesterday”Submitted I Love Calculator as a story · 6 points
HN discussionNobody writes code in terms of Nands but there is Nand to Tetris course ("The Elements of Computing Systems: Building a Modern Computer from First Principles" book) https://www.nand2tetris.org pytorch to LLMs has a lot to show even without Python to pytorch part. It reminds me of "Neural Networks: Zero to Hero" Andrej Karpathy https://m.youtube.com/playlist?list=PLAqhIrjkxbuWI23v9cThsA9... Prerequisites: solid progr…
Comment linking /zero-to-hero.html
on “Implementing a ChatGPT-like LLM from scratch, step by step”Submitted HN time capsule hn comments analyzed by AI with hindsights 10yr later as a story · 4 points
HN discussionNeural Networks: Zero to Hero[1] by Andrej Karpathy [1] https://karpathy.ai/zero-to-hero.html
Comment linking /zero-to-hero.html
on “Implementing a ChatGPT-like LLM from scratch, step by step”Submitted Neural Networks: Zero to Hero as a story · 4 points
HN discussionThis doesn't even seem to look at "predictions" if you dig into what it actually did. Looking at my own example (#210 on https://karpathy.ai/hncapsule/hall-of-fame.html with 4 comments), very little of what I said could be construed as "predictions" at all. I got an A for commenting on DF saying that I had not personally seen save corruption and listing weird bugs. It's true that weird bugs have long been a defining…
Comment linking /hncapsule/hall-of-fame.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted Licklider 1960 as a story · 4 points
HN discussion10 Years Ago, December 11, 2015 - Introducing Open AI -- very meta: https://karpathy.ai/hncapsule/2015-12-11/index.html#article-... The company has changed and it seems the mission has as well.
Comment linking /hncapsule/2015-12-11/index.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted Karpathy – Neural Networks: Zero to Hero as a story · 4 points
HN discussionThe analysis of the 2015 article about Triplebyte is fascinating [1]. Particularly the Awards section. 1. https://karpathy.ai/hncapsule/2015-12-08/index.html#article-...
Comment linking /hncapsule/2015-12-08/index.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted Lexicap: Lex Fridman podcast transcriptions generated by OpenAI Whisper as a story · 4 points
HN discussion> https://karpathy.ai/hncapsule/2015-12-24/index.html#article-... I wonder why ChatGPT refused to analyze it? The HN article was "Brazil declares emergency after 2,400 babies are born with brain damage" but the page says "No analysis available".
Comment linking /hncapsule/2015-12-24/index.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted State of GPT in 2023 (A Refresher, for Context, as We Start 2024) [pdf] as a story · 3 points
HN discussionIt somehow feels right to see what GPT-5 thinks of the article titled "Machine learning works spectacularly well, but mathematicians aren’t sure why" and its discussion: https://karpathy.ai/hncapsule/2015-12-04/index.html#article-...
Comment linking /hncapsule/2015-12-04/index.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted Andrej Karpathy: Books as a story · 3 points
HN discussionExamples: tptacek gets an 'A' for his comment on DF which the LLM claiming that the user "captured DF's unforgiving nature, where 'can't do x or it crashes is just another feature to learn' which remained true until it was fixed on ..." Link to LLM review: https://karpathy.ai/hncapsule/2015-12-02/index.html#article-... . So the LLM is praising a comment as describing DF as unforgiving (a characterization of the pres…
Comment linking /hncapsule/2015-12-02/index.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted Microgpt.py – train and inference a GPT in pure, dependency-free Python as a story · 2 points
HN discussionAre you sure? The third section of each review lists the “Most prescient” and “Most wrong” comments. That sounds exactly like what you're looking for. For example, on the "Kickstarter is Debt" article, here is the LLM's analysis of the most prescient comment. The analysis seems accurate and helpful to me. https://karpathy.ai/hncapsule/2015-12-03/index.html#article-... phire > “Oculus might end up being the most succ…
Comment linking /hncapsule/2015-12-03/index.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted microgpt as a story · 2 points
HN discussionThis is a perfect example of the power and problems with LLMs. I took the narcissistic approach of searching for myself. Here's a grade of one of my comments[1]: >slg: B- (accurate characterization of PH’s “networking & facade” feel, but implicitly underestimates how long that model can persist) And here's the actual comment I made[2]: >And maybe it is the cynical contrarian in me, but I think the "real world" aspec…
Comment linking /hncapsule/2015-12-18/index.html
on “Auto-grading decade-old Hacker News discussions with hindsight”Submitted Founding Fathers on Today's America as a story · 2 points
HN discussion
