Trust Is the Binding Constraint
I invite you to upgrade to a paid subscription. Paid subscribers have told me they have appreciated my thoughts & ideas in the past & would like to see more of them in the future. In addition, paid subscribers form their own community of folks investing in improving software design—theirs, their colleagues, & their profession.
Teams accumulate code faster than they accumulate trust. That gap — code without trust — is where projects fail.
I have been circling this idea for two years in this newsletter. Trust accumulates slowly and evaporates in an instant. Most trust-building work is invisible. AI has poured accelerant on the visible parts. Code arrives faster than ever, while the practices that manufacture trust — collaboration, testing, review, shared understanding — quietly get sacrificed to solo speed.
So I am teaching a live class about it.
The Human Practices Behind Trustworthy Software runs on ByteByteGo Live, November 9th & 10th. It’s part of a new partnership with ByteByte to help me speak to a larger audience sustainably.
Each day is a 2-hour lecture followed by an hour of open Q&A. Two days, roughly the shape of the argument I have been making here, made concrete:
The trust problem. Trust in the product vs. trust in the process. Why trust breaks faster than it heals, and why its economics matter — trustworthy software lets customers spend less time guarding against risk and more time using the thing.
The XP practices that manufacture trust. Testing, pairing, continuous integration, planning, customer on the team, continuous deployment, refactoring, observability. Not as a checklist — as a system.
The principles underneath. Humanity, mutual benefit, improvement, flow, redundancy. The XP values — communication, simplicity, feedback, courage, respect — as a trust-manufacturing system, not a poster on the wall.
Trust-optimized AI-assisted development. This is the part I most want to get across. The tweak paradox: trustworthiness of code falls with the square of the size of changes. AI wants big batches, trust wants small verified steps. Single-player vs. multi-player: everybody is going to push you into solo AI development, and it is a huge mistake. Slowing down in the right places is how you go faster.
It’s for engineers and engineering leaders. It’s live, so you can argue with me in real time — the hour of Q&A each day is my favorite part of the format.
I have also made sure of a few practical things. Every session is recorded, so if the time zone is unkind you can catch up. Many folks expense it through their learning & development budget, and ByteByteGo provides an invoice and an email template for your manager.
A million caveats, because that’s how I operate. This is a paid class, and it’s not cheap. If your team already trusts each other completely and your AI workflow is humming, you don’t need it. (I have never met that team, but statistics says you’re out there.) And the ideas are all related to things I have written here — the class is the concentrated, worked-through version, with exercises you do on your team, not mine.
If it sounds useful, you can sign up here:
(That link gets you the details & the current pricing on the available courses, let them know I sent you)
One more thing. This class is the answer to a question I have been asked weekly for a year: “Okay, I believe you about trust — now what do I do on Monday?” This is the Monday version. I hope to see you there, and I hope you argue with me.
You’re currently a free subscriber to Software Design: Tidy First?. Buying me more time to think & write means more thoughts & ideas for you.
© 2026 Kent Beck
548 Market Street PMB 72296, San Francisco, CA 94104