Human-Led AI Letters.
One letter a week on using AI without surrendering the judgement, context and responsibility that consequential decisions require.
Each one leaves you with something usable: a clearer distinction, a better question, or a firmer basis for a decision you are currently holding.
Free. One email a week. Unsubscribe at any time.
What Arrives.
One letter, most weeks on the same day. Five minutes to read.
Some work through a single distinction — the difference between verifying an answer and merely re-reading it, or between a decision that deserves reopening and one that is only making you uncomfortable. Others take a practical decision problem, or a development in AI that changes something about how we should be deciding.
No tool announcements. No prompt collections. Tools come up when they answer a larger question; the centre of this work is the person using them, not the machinery.
Some of The Questions These Letters Work Through:
- When should AI be trusted, and when should its output be checked?
- How much verification is proportionate?
- How do I know whether I have enough information, or am simply avoiding the decision?
- What happens when prediction, values and responsibility point in different directions?
- How does a decision stay stable enough to act on and open enough to learn from?
- What should never be delegated without thought?
- Which human capabilities become more valuable as machines become more capable?
Who This is for.
You may recognise yourself here:
- You use AI daily and find it genuinely useful, and something about that still bothers you.
- You have asked it the same question more than once and received different answers.
- You have caught yourself repeating a conclusion you could not fully explain in your own words.
- You are carrying a decision you have already researched enough.
- You want the benefits without becoming dependent on the tool to think.
- You are answerable to someone — a business, a team, a family — for choices that are not straightforward.
You do not need technical knowledge. You do not need to work in AI. You only need to care about how people continue to think, decide and remain responsible as these systems get better.
What I will not promise you.
Perfect decisions. Permanent certainty. Freedom from mistakes. A framework that guarantees the outcome.
Human judgement is imperfect. AI is imperfect. The work is in combining them responsibly, and the goal is not certainty — it is judgement you can defend and live with.
I will not treat AI as a saviour or an enemy. It is a powerful and increasingly influential participant in human decisions, and its role deserves careful thought rather than enthusiasm or alarm.
Much of the public conversation about AI rewards speed, novelty and certainty. These letters are meant to be a quieter place to think about what is changing, what is being gained, what may be lost, and what needs to remain human.
Free. Unsubscribe at any time. No pitch on arrival.
~Manoj
Human Judgement in the Age of AI
Clear. Capable. Human.
