TechnicalAI glossary
Diffusion model
The dominant image/video generation technique: the model starts from random noise and "denoises" it step by step until the final image, guided by your prompt.
In other words
A sculptor who starts from a shapeless block and removes material until the statue emerges.
Alyriaz articles that mention it
- They ask ChatGPT who to vote for
American voters are preparing their vote with ChatGPT. Six months out from the French presidential election, what that teaches us, and the habits to adopt.
- Claude has broken a theoretical physics record
One sentence, one week, 96 processors: Claude pushed a particle physics record one step further. And the record holder himself checked it.
- Opus 5.5: what actually changes compared with the previous one
Beyond benchmarks, we detail what Opus 5.5 does differently from Opus 5 in practice: speed, token consumption, endurance on long tasks, and the shape of its answers.
- Is the race a good metaphor?
The race metaphor dominates political discourse on AI. We examine what it presupposes, what it makes unthinkable, and what might replace it.
- A model that writes a text in one go instead of word by word
Diffusion models applied to text generate an entire response in parallel rather than word by word. We explain the principle, its speed advantages and its limitations.
- How does an AI create an image? The answer is stranger than you think.
AI image generators don't draw in the way we understand it. They start from a chaos of random pixels and progressively denoise it until an image emerges. We explain how diffusion models work, without jargon, and what this process implies about the nature of these creations.