Values
What matters to you? What should the agent protect, prioritise or refuse?
AI tips · 001
Practical use
One of the most important shifts in using AI is to stop thinking of an agent as a disposable chatbot.
The shift
That means giving the agent enough of your preferences, values, working methods, standards and boundaries that it can begin to operate as a digital version of your way of thinking.
Peter Steinberger’s SOUL.md project offers one useful model: writing down the qualities and values that give an agent continuity. Your own version can be much simpler.
The boundary: the aim is not to hand over your judgment. It is to make the agent a more faithful working partner.
Write down the pattern
What matters to you? What should the agent protect, prioritise or refuse?
What does good work look like? What counts as finished, accurate or trustworthy?
What must the agent never assume? When must it stop and ask you?
Start today
Write a short document explaining how you make decisions, what you care about and how you want to be spoken to. Then let the agent use it while you work through one small, repeatable task.
Over time, the document becomes more useful because it is tested against real work. Your digital extension is not discovered in one grand prompt. It is built through attention, correction and continuity.
Disclosure
Disclaimer: Human content polished by AI.
This is the first entry in a continuing collection of practical AI tips.