TechnicalAI glossary
Attention
The mechanism at the heart of Transformers: for each word, the model "looks at" the other words and weighs their importance. This is what gives it a sense of context.
In other words
In "the mouse eats the cheese", knowing that "eats" links the mouse to the cheese.
Alyriaz articles that mention it
- Can you confide in a machine that sells ads?
People entrust ChatGPT with their health, their money, their doubts. It now displays adverts. Even without influencing its answers, that changes the relationship.
- Which AI assistant knows the most about you?
A study compares the data collected by AI assistants. Meta AI and Muse declare almost twice as much as the others. How to verify.
- Overwhelmed by AI, arXiv rations access for researchers
Overwhelmed by AI-written papers, arXiv limits each researcher to two submissions per month. A first, and a telling one for science.
- Should we be worried about good things?
A critique published this week accuses the dominant narrative of diverting attention from concrete harms towards spectacular risks. We examine the argument, what is right about it, and how to distribute one's concern.
- The day we stop talking about it
The technologies that transform the world all, one day, stop being a topic. We examine what happens in that moment of silence, why it is more decisive than the period of effervescence, and what we should take from it.
- Does AI rob us of boredom?
AI instantly fills every moment of waiting or uncertainty. We explore what boredom actually produced, why its disappearance is not trivial, and what remains to be done with it.