What should AI learn to optimise?
What psychology, slime mould, evolutionary algorithms, Pareto frontiers and governance are teaching me while I build the Agape Duck drone simulator.
← Home
Insights
Practical notes from the work itself.
Pieces about the things we build, test and sometimes abandon. Written from first-hand work, with sources and uncertainty made visible where they matter.
Articles
How to spot a five minute admin task worth automating
Small repeated tasks compound. A simple way to work out which ones are worth changing, and which are better left alone.
Build the prototype before making the claim
Why the Sky Pilot simulator and the 3D print shelf are labelled as prototypes and experiments rather than finished products.
What human review should actually look like
Practical review points for routine communications and approved-information workflows, so a person stays responsible for what goes out.
Newsletter
Agape Duck Field Notes
Occasional updates covering practical AI, automation, websites, prototypes and lessons from real builds.
There is no sign-up form yet, so this is done by hand: email Jezz and you will be added to the list.
Email me to join the Field Notes listHave a practical problem worth solving?
Send a short, non-confidential outline of a recurring task or useful idea and we can consider whether there is a useful next step.
Start a conversation