Archives

11 posts published

  1. T

    Transformer son organisation tech pour la rendre AI-driven

    Ce qui change quand construire coûte moins cher que décider. Comment repenser l'organisation d'une scale-up de 80 personnes ou plus quand l'IA accélère la construction.

    -team-performanceai
  2. L

    Le craft est mort, vive le craft

    C'est ce que m'a dit un candidat en entretien après qu'on lui ait dit que sa culture craft était insuffisante. Sa réponse est franche : avec l'IA, ce qui compte, c'est d'aller vite. On pourrait balayer ça d'un revers. Sauf que dans les équipes que j'accompagne, j'entends la même chose, formulé différemment.

    -engineering-practicesai
  3. O

    On a doublé l'équipe et rien n'a accéléré - Une histoire de performance

    Pourquoi doubler une équipe ne double pas sa vélocité. Le throughput d'une organisation dépend de ses porteurs, les quelques personnes capables de transformer un sujet flou en résultat livré, pas de son effectif total. Cet article déconstruit le réflexe "plus de gens = plus de vitesse" et explore ce qui se passe quand on ajoute de l'IA à une équipe qui manque de porteurs.

    -team-performanceleadership
  4. "

    "C’est historique."

    "C'est historique." Pendant vingt ans, c'était un coût humain acceptable. L'IA le transforme en coût de production direct. Chaque décision non documentée devient une source d'erreur quotidienne pour les agents. La connaissance orale ne suffit plus.

    -aileadership
  5. E

    Everyone is busy, nothing ships

    Everyone is busy and nothing ships. The problem isn't that people aren't productive enough. It's that they're too productive, in the wrong place.

    -productteam-performance
  6. F

    From Copilot to harness engineering

    Individual AI initiatives are everywhere. Going from there to a team that builds and maintains its own AI harness is a different problem. The gap is smaller than it looks.

    -engineering-practicesai
  7. A

    AI doesn't replace junior developers. But it changes how we train them.

    AI speeds up production, not learning. Cutting junior hiring is rational today and a disaster in 5 years. The real question is whether we still know how to train them.

    -team-performanceai
  8. L

    Let it burn - Why the best organizations choose what they don't fix

    Three projects max. The rest burns. Why triage and iteration beat reorgs for improving team efficiency

    -team-performanceleadership
  9. D

    Devs produce twice as many PRs. Delivery hasn't moved.

    Devs produce twice as many PRs with AI. Delivery hasn't kept pace. The real ROI isn't in the tokens burned, it's in what teams do with the time AI frees up.

    -metricsai
  10. F

    Feature Overdose

    A team ships 11 features in a quarter. The board is happy. Delivery is green. When someone checks the usage data, 3 are actually adopted. The other 8 exist in the product, in the documentation, in the support scope. Not in user habits.

    -productai
  11. C

    Cheap to build, costly to keep

    AI has broken the cost of writing code. It hasn't touched the cost of owning it. How to measure what your velocity dashboards won't show you.

    -engineering-practicesmetrics
Copyright © 2026Shape and ShipPowered by Writizzy