In my mind, education is the very difficult technical process of building ramps to knowledge.
Words found on the source page on Nov 6, 2026.
More from Andrej Karpathy
Blog2025 LLM Year in ReviewDec 19, 2025
And I've vibe coded entire ephemeral apps just to find a single bug because why not - code is suddenly free, ephemeral, malleable, discardable after single use. Vibe coding will terraform software and alter job descriptions.Blog2025 LLM Year in ReviewDec 19, 2025
Anthropic got this order of precedence correct and packaged CC into a delightful, minimal CLI form factor that changed what AI looks like - it's not just a website you go to like Google, it's a little spirit/ghost that "lives" on your computer.Blog2025 LLM Year in ReviewDec 19, 2025
LLMs are emerging as a new kind of intelligence, simultaneously a lot smarter than I expected and a lot dumber than I expected.Blog2025 LLM Year in ReviewDec 19, 2025
Training on the test set is a new art form. What does it look like to crush all the benchmarks but still not get AGI?Blog2025 LLM Year in ReviewDec 19, 2025
We're not "evolving/growing animals", we are "summoning ghosts".BlogVerifiabilityNov 17, 2025
Software 1.0 easily automates what you can specify. Software 2.0 easily automates what you can verify.BlogVerifiabilityNov 17, 2025
The more a task/job is verifiable, the more amenable it is to automation in the new programming paradigm.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
I do, but it’s business as usual because we’re in an intelligence explosion already and have been for decades.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
If you talk to a billion parameter model, I think in 20 years, you can have a very productive conversation.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
In my mind, this is more accurately described as the decade of agents.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
Reinforcement learning is a lot worse than I think the average person thinks.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
The reason you don’t do it today is because they just don’t work.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
The way I like to put it is you’re sucking supervision through a straw.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
We’re building ghosts or spirits or whatever people want to call it, because we’re not doing training by evolution. We’re doing training by imitation of humans and the data that they’ve put on the Internet.PodcastDwarkesh Podcast: AGI is still a decade awayOct 17, 2025
What takes the long amount of time and the way to think about it is that it’s a march of nines.BlogPower to the people: How LLMs flip the script on technology diffusionApr 7, 2025
As far as I can tell, the average person has never experienced a technological unlock this dramatic, this fast.BlogPower to the people: How LLMs flip the script on technology diffusionApr 7, 2025
Mary, Jim and Joes are experiencing the majority of the benefit, not Google or the government of the United States.BlogPower to the people: How LLMs flip the script on technology diffusionApr 7, 2025
So it strikes me as quite unique and remarkable that LLMs display a dramatic reversal of this pattern - they generate disproportionate benefit for regular people, while their impact is a lot more muted and lagging in corporations and governments.
On the same topics
BlogI do not want your brains to rotSep 17, 2026
Cognitive offloading is a strategic delegation that costs nothing. Cognitive surrender is something different; an uncritical abdication of reasoning itself. And there is something about AI, about its allure and potency, that could make surrender far more widespread.BlogI do not want your brains to rotSep 17, 2026
Now advanced AI encourages us to delegate more and more, simpler and simpler tasks, weakening the practices that maintain our cognitive capacities.EssayHow Universities Should Prepare FoundersAugust 2026
And even if you have AIs writing all your code for you, you're still in the position of an engineering manager, and good engineering managers should be able to do the work of those working for them.EssayBanning AI in Law School: We've Seen This BeforeJul 11, 2026
In the case of AI, the connection of AI to freedom of expression is unshakable and irrefutable.EssayWhat if everyone could afford an expert?Jun 18, 2026
More specifically, I believe it will become a norm, in the way that smartphones have become a norm, for billions of people worldwide to have a designated AI health advisor, an AI legal advisor, and a general-purpose AI tutor that they regularly consult with as they navigate their lives.BlogAbundant IntelligenceSep 23, 2025
Maybe with 10 gigawatts of compute, AI can figure out how to cure cancer. Or with 10 gigawatts of compute, AI can figure out how to provide customized tutoring to every student on earth.
About this quote
Quotesby Glyph