korrents
David Heinemeier Hansson

What David Heinemeier Hansson thinks about LLMs

@dhh · 747 positions · 10 changes of mind · On ppll ↗

Creator of Ruby on Rails, CTO of 37signals, and creator of Omarchy.

David Heinemeier Hansson did not write this page.

We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.

Filter LLMs · Newest first
Topic
All 747LLMs 15Apple 46management 36Europe 30ambition 27coding agents 27agency 26America 26design 26
104 more topics
open source 26politics 25behavioural science 24happiness 23optimism 22the cloud 22hiring 21Linux 20startups 20motivation 19SaaS 19AI and jobs 17social media 17government 16AI agents 15creativity 15humility 15simplicity 14capitalism 13censorship 13immigration 13family 12Ruby 11product management 9team size 9Anthropic 8cybersecurity 8infrastructure 8software quality 8vibe coding 8Google 7remote work 7Ruby on Rails 7taste 7benchmarks 6developer tools 6health and fitness 6JavaScript 6Microsoft 6patience 6Psychotherapy 6war 6energy 5innovation 5Omarchy 5religion 5attitude change 4code generation 4courage 4data centers 4education 4empathy 4Hyprland 4ideas vs execution 4investing 4market efficiency 4Russia 4self-driving cars 4software performance 4AI alignment 3Arch Linux 3climate change 3democracy 3libertarianism 3privacy 3semiconductors 3venture capital 3AGI 2AI and human skill 2AI slop 2AI writing 2aviation 2compilers 2crypto 2demographics 2discipline 2electric vehicles 2friendship 2Japan 2Silicon Valley 2solar 2video games 2Addiction 1batteries 1China 1CSS 1curiosity 1gender equality 1journalism 1mathematics 1MCP 1metascience 1Nix 1OpenAI 1permitting 1personal websites 1physics 1Python 1racism 1scaling laws 1stock market 1sustainability 1TypeScript 1writing 1
Sort
NewestOldest

Only David Heinemeier Hansson on LLMs. See everyone on LLMs

15 dated positions, 2023 to 2026, in their own words. Our reading of what David Heinemeier Hansson has said — not written or endorsed by them.

  1. Our readingRegarding AI as just another tool is a category error.

    Their wordsIf you're stuck at "AI is just another tool", you're making a category error.

    ↗@dhh on Xx.com

  2. 7 days earlier

  3. Our readingRationing AI tokens reflects a weird scarcity mindset when more tokens are easily affordable than one could ever use.

    Their wordsBut it's all grounded in a weird scarcity mindset. I don't have a token limit! I'll never be able to use all the tokens I could easily afford.

    ↗@dhh on Xx.com

  4. 1 day earlier

  5. Our readingThe number of programmers worldwide capable of outperforming collaborating frontier AI models is negligible.

    Their wordsThe number of programmers worldwide who can beat a collaborating pair of frontier models rounds to zero.

    ↗@dhh on Xx.com

  7. Our readingGetting stuck on the 'stochastic parrot' meme about AI leads people to miss the creativity and novel thinking in front of them.

    creativity

    Their wordsDon't get stuck on the stochastic parrot meme. You'll miss all the sparkling creativity and novel thinking in front of your eyes.

    ↗@dhh on Xx.com

  8. 4 weeks earlier

  9. Our readingWork produced by one frontier model should be reviewed by a different one, for the same reason a good programmer's work improves under a good peer's review.

    Their wordsWhy are we surprised? Even if you're a good programmer, if you finish a job and you ask your also very good peer to review it, you're gonna end up with better code. Of course you're gonna end up with better code. So build that into your process.

    ↗DHH: Future of Programming, AI, Agentic Engineering, Vibe Coding & Linux | Lex Fridman Podcast #501youtube.com 25th of 41 in this recording

  11. Our reading · no longer heldAn open-weight model is already all anyone needs; it is as good as the frontier models.

    Recalls holdingWhen I was instructing the agent, I was like, "Oh my God, Kimi's just as good as these other models."

    SaidBut I've been wrong about this before.

    ↗DHH: Future of Programming, AI, Agentic Engineering, Vibe Coding & Linux | Lex Fridman Podcast #501youtube.com 29th of 41 in this recording

  13. Our readingThe two standard criticisms of AI — that it is not deterministic and that it is not creative — cannot both be true.

    Their wordsThere's both the charge that it's not deterministic and therefore bad, and also that it is not creative. It's one or the other, bro.

    ↗DHH: Future of Programming, AI, Agentic Engineering, Vibe Coding & Linux | Lex Fridman Podcast #501youtube.com 34th of 41 in this recording

  15. Our readingLarge language models could still plateau, and that possibility should be held open even though no evidence of it has appeared.

    humilityscaling laws

    Their wordsSo we should also have the humility. As amazing as the LLMs are now, it could be that they eventually plateau. We haven't seen any evidence of it yet, and I think this is also why we're seeing this absolute gobsmacking levels of investment, because so far the scaling laws are true, and the more billions are poured in, the more intelligence comes out.

    ↗DHH: Future of Programming, AI, Agentic Engineering, Vibe Coding & Linux | Lex Fridman Podcast #501youtube.com 35th of 41 in this recording

  16. 4 weeks earlier

  17. Our readingOpen-weight models are necessary to protect against ideological tyranny

    Their wordswe desperately need strong open-weight models to protect ourselves against this kind of soft ideological tyranny

    ↗I'm sorry, Daveworld.hey.com 2nd of 2 in this piece

  18. 14 months earlier

  19. Our readingCode quality still matters even as LLMs make writing code much faster

    software quality

    Their wordsAI is a superb pair programmer, but I'd retire before permanently handing it the keyboard to drive the code

    ↗Coding should be a vibe!world.hey.com

  20. 6 weeks earlier

  21. Our readingAI becoming original is a likely next step

    Their wordsYou're a fool if you don't believe originality is a likely next step

    ↗Great AI Stealsworld.hey.com

  22. 3 months earlier

  23. Our readingAI models are not yet capable of producing great code

    Their wordsas of right now, I've yet to see any of the AI models I've been using for the past year produce great code within domains that I'm very familiar with

    ↗The premise trapworld.hey.com 1st of 2 in this piece

  25. Our readingThe next big leap for AI models probably isn't likely until they are exposed to proprietary corporate code

    Their wordsI think the next big leap for these models under the current paradigm probably isn't likely to happen until they're exposed to a vast corpus on proprietary, corporate code

    ↗The premise trapworld.hey.com 2nd of 2 in this piece

  26. 10 months earlier

  27. Our readingFallible humans have an inherent advantage over AI by being forgivable

    Their wordsMaybe fallible humans have an inherent advantage over AI by being forgivable?

    ↗It’s easier to forgive a human than a robotworld.hey.com

  28. 10 months earlier

  29. Our readingThe recent advance of AI may be the beginning of the end for software development as we formerly knew it

    Their wordsMaybe we have finally reached the beginning of the end for software development as we formerly knew it

    ↗How to continue making kerosene lamps on the eve of electricityworld.hey.com