Home Speakers Dwarkesh Patel Quote We see these AIs as a galaxy glittering with capabilities, but at their center, invisible to the naked eye, holding all the constellations together, is an unimaginably massive black hole of data.
Words found on the source page on Nov 6, 2026.
More from Dwarkesh Patel Podcast What Robert Caro learned about power after studying it for 50 years Oct 30, 2026
If it allows it, Airlock takes action on behalf of the agent, which means your agents never see your API keys or other credentials. Podcast What will computer chips look like in 2045? Oct 22, 2026
There’s never been as big a shock to the semiconductor industry as the current surge in demand from AI. Podcast Si Sheppard – How did a few hundred Spanish soldiers topple two empires? Oct 1, 2026
That way, your software gets tested against scenarios you’d never imagine but that could still arise in the real world. Essay 8 Predictions for the Era of Continual Learning Aug 7, 2026
Continual learning will create a clear moat for leading AI labs that they currently lack. Essay 8 Predictions for the Era of Continual Learning Aug 7, 2026
I don’t think you can have AIs that perform whole jobs as competently as humans if they’re forced to just write Markdown files from session to session. Essay 8 Predictions for the Era of Continual Learning Aug 7, 2026
If you want to change what AI you are using, you basically have to fire an employee that has months of context on your organization and replace them with a fresh one that you have to retrain from scratch. Essay 8 Predictions for the Era of Continual Learning Aug 7, 2026
This is one of many reasons why I think it’s unwise to lock in some kind of regulatory safety regime right now. We simply don’t know what kind of technology we’re going to be looking at even in a year, let alone in five years or ten years, and we’d be entrenching an archaic and potentially counterproductive approach to dealing with the threats from AI. Essay 8 Predictions for the Era of Continual Learning Aug 7, 2026
When deployment becomes part of training, the returns to being ahead will accelerate. Essay Why compute might get 10x+ more expensive in coming years Jul 29, 2026
I want to emphasize the key conclusion here: as AI models become smarter, they’ll better monetize the same amount of compute. If a true human-level software engineer that could run on an H100 equivalent, at current market rates for software engineers, that H100 should rent for over $250k a year. Essay Why compute might get 10x+ more expensive in coming years Jul 29, 2026
I wish we didn’t live in a world with such strong economies of scale of intelligence (because I’m worried about power concentration). Essay Why compute might get 10x+ more expensive in coming years Jul 29, 2026
If by 2028 we’ve automated software engineering and the price of compute is 15x higher than it is right now, then it’s going to be much more difficult for you, with no revenue, to compete for compute against the frontier labs. Essay Why compute might get 10x+ more expensive in coming years Jul 29, 2026
If you’re spending most of your compute on inference, you’re kind of declaring that AI progress has stalled, because it’s not worth investing more in training, and your business is basically that of a cloud provider. Essay The data black hole at the center of AI Jun 19, 2026
I think the reason it is relatively easy for open source and previous laggards to catch up to within months of the frontier is that data is the real driver of progress. Essay The mistake of conflating intelligence and power May 16, 2026
But in the real world, power is more the product of having the authority and trust to get lots of people to collaborate with you, rather than some galaxy brain scheming capability. Essay The mistake of conflating intelligence and power May 16, 2026
I’m simply pointing out that the way AI systems are currently becoming smarter (by getting trained to be to be really good at specific economically valuable tasks like coding) is not that strongly correlated with power.
On the same topics Blog Arguments in Favor of AI Fair Use Sep 7, 2026
As LLMs continue to scale up, the value of any particular source continues to diminish. Blog Arguments in Favor of AI Fair Use Sep 7, 2026
Much, if not most, of the value of LLMs derives from their ability to generate new benefits beyond what the source material alone can yield. Essay Making Startups Powerful September 2026
The obvious AI variant is to let your users opt in to training your model on their interactions with it. Many will resist that, but if some don't, the model they get to use will outperform the vanilla one used by the others. Filing Palantir Q2 2026 earnings release (SEC Form 8-K, Exhibit 99.1) Aug 3, 2026
Our customers trust us to provide them with maximal control over their operations, data, and decisions. Their competitive advantage should never become the training data for future models. Blog Who's Afraid of Chinese Models? Jul 20, 2026
Second, intelligence isn’t in fact a perfect commodity, in part because applied intelligence makes itself smarter. Blog The Reverse Information Paradox Jul 12, 2026
If learning flows in only one direction, economic value converges toward the owners of the learning infrastructure rather than the creators of the knowledge itself. About this quote 
Quotesby Glyph