Dexter Awoyemi explores what increasingly capable AI means for individuals and organisations, and helps teams decide what AI is allowed to change.
One way to get the outcomes that we actually want is to be intentional with our asks.
Surprise is information: it’s the signal that our model of reality needs an update.
The most valuable thing is often the hardest: knowing which decisions to delegate to AI and which ones stay with you.
AI doesn’t resolve the ambiguity in your business.
I believe AI presents existential risk AND I believe in accelerating the deployment of AI.
An insurance company is building two teams. One will own actions, and the other will control them.
Why human attention gets more valuable as AI grows
By the time we have a good explanation, the AI has already moved on.
What are we willing to delegate?
Software turns prediction into action
Are you automating irreversible decisions?
True acts of creation involve selection
An encoded model of living with personal agents
The goal of communication at every scale
How natural language, uncertainty and structure define modern AI
AI amplifies what we notice
The upstream constraint that determines what AI amplifies
Reducing ambiguity so intent can survive execution
AI exposes structural bottlenecks to the bandwidth of information flow
Attention as the new discipline for the age of cheap cognition
The new power dynamic of AI: individuals move faster than organisations
Why human-in-the-loop isn’t optional for agentic AI
Expert systems 2.0: still human-in-the-loop and targeted at real business value
6 (+1) ways to engineer reasoning above LLMs to improve their capabilities
Open reasoning with the combined strengths of text and symbols
Giving people control by moving from big data to self-sovereign data
Comparing features of open finance APIs. Part two of several.
Connecting apps to your bank account
Comparing features of open finance APIs. Part one of several.
How and why institutions are building open APIs.
Helped fix the matching problem after major firms had struggled with it for years.
Found what was blocking delivery and shipped the workflow in one month.
Clarified who needed to approve it, how it would work, and what had to be checked before launch.
Showed the real issue in the acquisition pipeline before a premature platform build.