Chats with Kent C. Dodds

Jamon brings a useful mix to this conversation: founder of Infinite Red, longtime consultant, React Native specialist, and now indie game developer. That perspective makes the episode unusually practical. He has spent years watching where projects go wrong when product thinking is weak: bad requirements, unclear stakeholder alignment, UX details nobody owned, and engineers optimizing the wrong thing too early.

The thread through the whole episode is durability. Product engineering is not just about shipping faster with agents or getting better at a specific tool. It is about understanding people, shaping better requirements, recognizing when the human side of the workflow matters more than the code, and making decisions that keep paying off as the technology changes around you.

Homework

Sit down with a non-technical person and watch them try to use a feature you built.

Write down every hesitation, workaround, double-click, or confusing step you notice, then use that list to reprioritize what you fix next.

Resources

Infinite Red

Jamon Holmgren — site

Night Shift Agentic Workflow

Gunship Origins on Steam

Guest: Jamon Holmgren

Company: Infinite Red

GitHub: @jamonholmgren

𝕏: @jamonholmgren

Host: Kent C. Dodds

Website: kentcdodds.com

𝕏: @kentcdodds

GitHub: @kentcdodds

YouTube: Kent C. Dodds

Video

Watch this episode on YouTube

Don's career makes this episode unusually wide-ranging: early computing, human error, aviation safety, Unix, Apple product decisions, digital cameras, color TV, and the long arc from usable products to systems that shape society. The through-line is straightforward but demanding: if you want better products, watch what people actually do, notice the workarounds they no longer complain about, and treat clusters of small usability problems like real product debt.

The second half brings that thinking into the present. Don and Kent talk about AI coding tools as force multipliers that still need direction, architecture, and supervision, then zoom out to Design for a Better World and the Don Norman Design Award. The result is a conversation about product sense that spans decades without feeling dated: the tools change, but the responsibility to understand people, systems, and consequences does not.

Homework

Spend time watching people do real work before you ask them for solutions; observation reveals the hidden setup, workarounds, and friction they now assume are just "how it works."

After a release, step back and fix clusters of small usability issues as a system instead of waiting for one confusing bug to become catastrophic.

Treat AI as a force multiplier you must instruct and supervise; stay responsible for the problem definition, architecture, and review.

Resources

Don Norman Design Award (DNDA)

Design for a Better World

The Design of Everyday Things

Nielsen Norman Group — Don Norman

United Nations Sustainable Development Goals

Guest: Don Norman

Company: Don Norman Design Award (DNDA)

Host: Kent C. Dodds

Website: kentcdodds.com

𝕏: @kentcdodds

GitHub: @kentcdodds

Youtube: Kent C. Dodds

Video

Watch this episode on YouTube

Will's path runs from designing bucket trucks to self-taught software engineering, education products, and database tooling, and that background gives this episode a distinctive lens: software is still a product people use with bodies, habits, emotions, and mental models. The conversation makes product sense concrete through examples like onboarding timing, course complexity, support workflows, and the small confidence signals that separate stable-feeling products from merely functional ones.

You'll hear why watching users work keeps surfacing across this series, how to tell broken experiences from merely unpopular ones, why user feedback usually improves polish more than strategy, and how product engineers can stay valuable in an agent-heavy future by understanding both the user and the constraints of the software medium.

Homework

Use AI agents more for gathering than executing: explore multiple solution paths, adjacent domains, and missing context before you ship.

Give agents richer context like user demographics, constraints, and likely mental models, then use your own judgment to evaluate what comes back.

Slow down long enough to question assumptions before implementation; use AI as a creativity and critique tool, not just a code accelerator.

Resources

Will King - site

Deploy Empathy (Michele Hansen)

The Mom Test (Rob Fitzpatrick)

Interface Craft (Josh Puckett)

Guest: Will King

Company: Crunchy Data

GitHub: @wking-io

𝕏: @wking__

Host: Kent C. Dodds

Website: kentcdodds.com

𝕏: @kentcdodds

GitHub: @kentcdodds

Youtube: Kent C. Dodds

Video

Watch this episode on YouTube

Aaron builds in public—Laravel roots, education, and now Solo, a terminal multiplexer–style desktop app for organizing agents and dev stacks. This episode is a practical tour of product sense for developers: watching people work, reading support email with empathy, cow paths vs. fences, and why the “right” architecture can still lose if humans go home furious.

You’ll hear how Aaron reasons from problem → solution when users ask for worktrees, when to duplicate UI affordances even when the model is “one,” and how introverts can still do discovery by treating outreach like an optimization mission—plus niche opportunities outside the Cursor clone gold rush.

Homework

When someone asks for a solution (e.g. a feature), slow down and ask what problem they’re really trying to solve—users often lead with implementations.

Practice user empathy: imagine someone stressed, trying to finish work; question “technically correct” UX that blames the user instead of protecting them (confirmations, back-button data loss, etc.).

If talking to people is hard, reframe discovery as a systematic search (spreadsheet energy, trusted partners, or domain friends)—or pair with someone who loves conversations.

Resources

Aaron D. Francis — X

Jobs to Be Done (Clay Christensen)

The Design of Everyday Things (Don Norman)

Guest: Aaron D. Francis

Company: Solo & Laravel education

GitHub: @aarondfrancis

𝕏: @aarondfrancis

Host: Kent C. Dodds

Website: kentcdodds.com

𝕏: @kentcdodds

GitHub: @kentcdodds

Youtube: Kent C. Dodds

Video

Watch this episode on YouTube

Dillon's path runs from internal insurance tools to Vercel Domains to Cloudflare's agent and dashboard work-always with the same through-line: care about the user, get real feedback, and invest in primitives so delighters don't collapse under bad foundations. This episode covers metrics and paging as a product habit, learning from customer escalations, scoping small when AI speeds up coding, and building cross-functional relationships (support, sales, finance) as part of engineering judgment.

You'll hear practical parallels with episodes on delighters and onboarding tension, plus why reviewing agent-written code still matters for system intuition when things break at 2 a.m.

Homework

Try hard and care a lot; more practically, focus on foundations and primitives.

Put good feedback systems in place so you know what's going on with your product and where it doesn't feel good-alerting and metrics, customer journey signals, or customer interviews.

If you have a customer support team, sit with them and watch them triage cases for your product; get to know support-they're sitting on a gold mine of product signal-and empathize with them like you do with users.

Kent's shorthand for the mindset Dillon agreed with: make pain painful-if your users are hurting, you should feel it too.

Resources

Cloudflare - Developers

Cloudflare Agents

Dillon Mulroy - site

Dillon Mulroy - GitHub

Guest: Dillon Mulroy

Company: Cloudflare

GitHub: @dmmulroy

X: @dillon_mulroy

Host: Kent C. Dodds

Website: kentcdodds.com

X: @kentcdodds

GitHub: @kentcdodds

Youtube: Kent C. Dodds

Video

Watch this episode on YouTube