The Louvre was a spectacular setting for two very hot days at RAISE Summit in Paris!
Getting AI to do something useful is becoming easier. Getting it to work reliably inside a company is not.
In the opening keynote with Mark Cuban and Anton Osika they shared that more than one million new projects are created on Lovable every week!
Chris Clark, co-founder and COO of OpenRouter, shared that inference could become a dominant operating expense for knowledge businesses. List price isn't the same as real-world cost, benchmarks are no substitute for your own evals, and reliable inference depends on routing, observability and failover.
Another theme at Raise, was that AI adoption is increasingly an organisational design challenge.
Marco Argenti, from Goldman Sachs, described non-technical colleagues prototyping solutions, with forward-deployed and engineering teams turning the strongest ideas into production-grade systems.
Duncan Clark CEO Pro Design Canva shared how they paused normal work for an AI learning week, giving people space to explore, build and learn without demanding an immediate outcome.
Guillaume Princen from Anthropic offered an important reminder: AI should help us think further, not replace our thinking. The difficult question is deciding which decisions should remain human.
Sachin Katti, Head of Industrial Compute at OpenAI, described the progression from supervised "AI interns" carrying out workflows towards more autonomous research loops. Hearing that Codex has become the default work interface inside OpenAI made that future feel much less abstract.
The All-In Podcast Live session was another highlight and very entertaining!
It was even better to spend both days exploring the conference with my brother, Eliot Payne. The swag was excellent. IYKYK.
It was also great to represent AI Tinkerers Hong Kong & Greater Bay Area, great to meet Jake Laes, Joe Heitzeberg and as well as to join the AI Tinkerers Paris dinner with fellow founders, leaders and builders Abdoulaye D., Shivnarayan Rajappa, Matthias Lübken, Clara Gainon de Forsan de Gabriac, Ph.D., Simon Zilinskas - Inta, Oleg Sokolov, Vatsal Bhardwaj.
Paris was beautiful. The heat (and limited aircon) was relentless but the question I'm taking home: how do we give more people the freedom to experiment while building the infrastructure, evals and human decision boundaries needed to trust what reaches production?
