TechnicalAI glossary
Transformer
The architecture invented in 2017 that made modern LLMs possible. Its strength: "attention," which links each word to every other to grasp context.
In other words
The common engine of almost all current AIs, their "internal combustion engine."
Alyriaz articles that mention it
- Deutsche Telekom aims to save 2.5 billion with AI
Deutsche Telekom aims to save 2.5 billion euros through AI by 2030. Hotline, network, call translation, and two figures missing.
- Accenture: AI was supposed to ruin it, it's making it soar
Accenture was supposed to be a victim of AI. Its results sent its shares up as much as 22%: companies need help adopting it.
- Revising with AI without getting caught out
An AI summary gives the impression of having understood, without retaining anything. Five ways to use an assistant to actually learn.
- SubQ: the startup that may have killed the transformer
Since 2017 and the publication of the paper "Attention is All You Need", every major AI model has been based on the transformer architecture. SubQ, launched on 5 May 2026, is the first commercial model to break with this paradigm. Here is what changes technically, and why the AI community is deeply divided over the implications.
- AI's biggest failures: the billion-dollar products that crashed and burned
For three years, AI was presented as the revolution that would transform everything. In 2026, the first major products are starting to die publicly: Sora abandoned by OpenAI, Apple hit with a $250 million judgment, Microsoft Designer buried, Humane Pin remotely deactivated. Here is what these five failures really teach us.
- How AI works: The Fortune Teller's Paradox
People often think an AI thinks or understands what it is being told. In reality, it is a gigantic statistical guessing tool. Discover the secret.