Context
On April 7, 2026, the French National Assembly's fact-finding mission on artificial intelligence (Mission d'information sur l'intelligence artificielle) held a hearing with Laurent Alexandre and Olivier Babeau. A clip from this hearing was shared on X by Fabien Mikol.
Laurent Alexandre's thesis
Laurent Alexandre puts forward a deliberately provocative argument: we are massively underestimating the current power of AI, and we need to "stop the denial."
Key claims
- AI is no longer a simple assistance tool. It already surpasses experts on several cognitive tasks — coding being the most visible example.
- The "calming" discourse is outdated. He directly challenges Luc Julia, whom he presents as overly reassuring about AI's actual capabilities.
- Models like Claude already outperform many humans on specific tasks, particularly in programming.
- Continuing to minimize this is a strategic mistake — both for individuals planning their careers and for institutions shaping policy.
The central idea
The real disruption is already here, especially for code and intellectual professions. The gap between what AI can do and what the public discourse acknowledges is growing wider, not narrower.
What to make of it
Laurent Alexandre is a well-known figure of the French AI debate, often positioned as deliberately alarmist. Some of his claims rest on solid observations (LLM performance on coding benchmarks, rapid capability gains), while others are rhetorical provocations designed to shake complacency.
The interesting tension is between two legitimate positions:
- The "tool" framing (Julia, Villani): AI is advanced software, powerful but ultimately controllable and bounded.
- The "rupture" framing (Alexandre): AI capabilities are crossing thresholds that make the "just a tool" narrative dangerously misleading.
The truth likely involves both — but the policy implications differ sharply depending on which frame dominates.
Source: Fabien Mikol on X — clip from the National Assembly hearing, April 7, 2026.