diffusion
A generative modeling method that produces outputs by iteratively denoising, used for continuous action generation.
As defined in Fully autonomous robots are much closer than you think – Sergey Levine
Other definitions
The spread of a technology through an economy as people and firms adopt it, which determines who captures its benefits.
Scaling with Soul
Where it comes up · 9
- Sarah Guo - Funding the Frontier - [Invest Like the Best, EP.489]3 claims
- “People Don’t Realize This” about Climate Change, Elon Musk Says1 claim
- Applied Intuition: A Billion Intelligent Machines1 claim
- Scaling with Soul1 claim
- An audio version of my blog post, Thoughts on AI progress (Dec 2025)1 claim
- Fully autonomous robots are much closer than you think – Sergey Levine1 claim
- Tyler Cowen — The #1 bottleneck to AI progress is humans1 claim
- Why Dictators Always Seem Unstoppable - Until They Re Not Marcel Dirsus1 claim
and 1 more
Most important claims · 10 of 11Search all claims for “diffusion” →
- The human body extracts oxygen from air to survive, generating CO₂ which is exhaled; exhalation depends on diffusion of CO₂ from the blood back into the atmosphere at the lungs, and higher atmospheric CO₂ makes this diffusion more difficult, causing CO₂ to accumulate in the blood and lower blood pH“People Don’t Realize This” about Climate Change, Elon Musk Says
- Historically, the countries that benefited most from major technological breakthroughs were not those that created the technology but those that adopted it most deeply, a process called diffusion, and Latin America can benefit by diffusing AI quickly even without model-building clusters.Scaling with Soul
- Uprisings diffuse across borders, so a dictator should watch neighboring countries closely: if people in a similar nearby system rise up, especially one sharing language or regime type, the dictator is likely to be in trouble and may even use preemptive repression.Why Dictators Always Seem Unstoppable - Until They Re Not Marcel Dirsus
- Unlike text which is discrete tokens, the robot's actions are represented using flow matching and diffusion because they are continuous and require very precise control for dexterous manipulation.Fully autonomous robots are much closer than you think – Sergey Levine
- The claim that AI isn't widely deployed because technology takes a long time to diffuse is cope used to gloss over the fact that current models simply lack the capabilities necessary for broad economic value outside coding.An audio version of my blog post, Thoughts on AI progress (Dec 2025)
- Many use cases are too expensive, too sensitive, or too slow to run on frontier-provider models today, so open-source models unlock diffusion of capability into the economy, and this gap is expected to grow as companies learn to do more with AI.Sarah Guo - Funding the Frontier - [Invest Like the Best, EP.489]
- Guo argues investors should learn from founders who are building surprising products or believe something the investor does not, rather than reasoning about grand strategic frameworks for which AI layer will win, because such frameworks are not useful for decision-making.Sarah Guo - Funding the Frontier - [Invest Like the Best, EP.489]
- The consistent epistemic move is to trust the relevant experts in each domain: AI experts (who see the promise of strong AI, against asleep economists) and technology-diffusion experts (who universally find diffusion slow with no rapid takeoff); combining both expert views yields the conclusion that AI is amazing in the long run but will be a slow, bottleneck-ridden slog.Tyler Cowen — The #1 bottleneck to AI progress is humans
- Intelligence diffuses slowly into physical machines compared with phones and laptops because manufacturers and farm or mine operators face impedances, and because buyers keep expensive vehicles for 10 to 15 years and over half of Americans have small savings, so a purchased machine is not replaced by a new autonomous option.Applied Intuition: A Billion Intelligent Machines
- Trying to predict which AI layer will win based on what frontier labs are doing is not useful for investment decisions, and investors should instead judge the competitiveness of each effort and focus on the next 99 percent of diffusion.Sarah Guo - Funding the Frontier - [Invest Like the Best, EP.489]