The Step Everyone Skips
Geoff Woods' CRIT framework asks the AI to interview you before it does anything. Most people skip that step, and the reason they skip it is exactly why it matters.
Strategy that survives contact with reality.
I write for the CTOs, CIOs and executives who have to live with the result: how AI and modernisation decisions hold up once they meet a real organisation.
What I am useful for
Most teams hand AI to their senior engineers and keep the juniors away from it, on the theory that only experienced judgement can catch a confident wrong answer. I think that gets it backwards: enable the seniors, yes, but empower the juniors with a harness. Tests that must pass, review gates, constrained scope, an environment where a wrong answer is cheap to produce and quick to catch. Restricting the tool to people who already have judgement is the resigned answer; building the system that supplies judgement is the engineering one.
Strategyintent and bets
Executionteams and systems
Realityconstraints, incentives, legacy, people
what reality teaches revises the strategy
What I think about
- AI that works in reality
- Architecture, operating models and engineering practice, not the demo.
- Enterprise transformation
- A system of constraints, incentives, legacy and people, never a technology swap.
- Systems thinking
- Models that explain why something works, where it breaks, and what moves when one variable does.
- Execution over theatre
- Ideas earn credibility through implementation: the decisions, the constraints, the lessons.
- Mentoring through complexity
- Explaining difficult things clearly, without dumbing them down, so judgment compounds.
Work you can inspect
Open-source repositories and systems, pulled live from GitHub at build time rather than described.