Training the Fast Brain

The Setup

Most of my real work happens on autopilot.

Not just the obviously routine work — Slack triage, calendar choreography, the small administrative loop that runs whether I am paying attention or not. The complex work too. The judgments on a vendor problem, a customer escalation, an ERPNext data model. By the time I notice I have decided something, the decision is already in motion. Some part of me has already chosen.

A second kind of thinking exists in parallel. The slow, deliberate, sit-with-it kind. It produces better work when it gets the time, but it costs me disproportionately more. Slow thinking is not a switch I flip — it is a state I have to construct, hold, and defend against constant interruption from my own mind. So most days, most of the time, my fast brain runs the show.

The question is not whether I use it. The question is what it has been trained on.

The Thinking

Kahneman's System 1 and System 2 give me the vocabulary. Fast and slow. Automatic and effortful. What the frame leaves implicit is how System 1 ends up the way it does — and that part is worth taking seriously, because it is where the leverage lives.

The clearest way I have found to think about it is reinforcement learning.

A reinforcement learning model develops a policy — a function that maps a situation to an action — by being exposed to many situations and receiving feedback after each one. Good feedback strengthens the action, bad feedback weakens it. Over enough iterations the policy converges on something that looks, from the outside, like intuition. It does not reason from first principles in the moment. It has compressed thousands of past experiences into a fast lookup.

My System 1 is exactly this. It is a policy. Trained on every reaction I have ever had, every consequence I have absorbed, every time something worked and every time it did not. Most of that training was passive — I was not supervising it, I was just living. The reward signal was whatever my nervous system happened to register as success or failure in the moment.

This matters disproportionately for me. My fast brain is doing more of the load-bearing work, because my slow brain is more expensive to engage. Which means the quality of my day, and of my work, is downstream of the quality of a policy I am not actively training.

The Work

The work is not to audit fast-brain decisions after the fact. My slow brain does not have the budget to run a post-mortem on every reaction. The work is to install small policies my fast brain can run on its own.

A concrete one I have been practicing: before assigning a task to a colleague, run a stakeholder check. Who else touches this? Who needs to be informed? Who will block on it if they hear about it after the fact? Three seconds, maybe four. It is not slow thinking — it is a small policy shaped by enough past misfires that my fast brain now runs it as part of the decision itself.

That is how the policy actually updates for me. Not by my slow brain reviewing afterwards, but by my slow brain choosing what gets reinforced and then letting repetition do the rest. The training has two paths. One is feedback from past outcomes — the painful kind that adjusts the policy whether I ask it to or not. The other is deliberate: identify a situation, design a small rule, run it consciously until the fast brain takes it over.

Instincts are not something I have. They are something I install.

The Edge

The risk of treating my fast brain as fixed — as personality, as wiring, as something I just have — is that I stop watching what it is learning. The policy drifts in whatever direction my environment pushes it, and I experience that drift as a change in myself.

The opposite move is to take the policy seriously as a policy. Something with inputs and outputs and rules I can choose to load into it. Not a thing I can rewrite from scratch. But a thing I can shape, one small installed habit at a time.

For me, this is not optional. My fast brain is going to do most of the work. The only real question is whether it is being trained on purpose or by accident.