Design

direction.

Here's what they actually pay me for. Not the mockups. Walking into a room where everyone is politely agreeing about the wrong problem, and asking the question that stops the meeting. Design's defensible value has never been the artifact. It's translation: between what people need, what engineering can build, and what the business can sustain, done in a way that produces decisions instead of debates.

Production keeps getting cheaper and faster, and AI accelerates it again. What stays scarce is judgment. Knowing which problem matters, how to frame it for the room, and when to push past the comfortable answer toward the true one. That's the part that compounds, and it's what I train teams to get good at.

You've seen the meeting where everyone nods and nothing gets decided. Organizations rarely stall on bad ideas. They stall when alignment breaks down, and bad alignment is rarely about conflict. It's the critical conversations that never happened. The assumptions nobody challenged, the tradeoffs that got deferred, the problems that looked solved but weren't. I help teams catch that earlier. Not by adding process, but by creating moments of real clarity: a shared view of what's true, what's uncertain, and what a good decision looks like.

I've spent as much energy on people as on pixels. Mentoring designers, building the rituals and trust that let a team take real creative risk, and keeping us close to the people we're designing for. Research and insight aren't a phase that happens once. They're a habit, and the teams that stay curious about their users make far fewer expensive corrections later.

When I write career models, I write them as profiles, not checklists. Nobody grows evenly, and pretending otherwise flattens exactly the strengths worth building on. A designer should be able to see themselves in the model, spiky edges and all, and still know precisely what reaching further looks like. The one competency I hold as a floor rather than a profile is fluency with AI. Strength elsewhere can't average that gap away anymore.

I'll name my own biggest risk, since I know it better than anyone: doing the work for people instead of with them. It's faster my way, and it teaches nobody anything. So I run working sessions instead of sending finished files, and I show my process on purpose. Remote work hides how things get made unless you make the making visible. I'd rather stay six months ahead on the unstable tools myself, take the burns, and hand the team the version that works.

Culture is the quiet infrastructure under every good outcome. Get it right and the work gets braver on its own.

I never stopped being a maker, and I don't trust design leaders who did. Graphic design, illustration, identity, type: the hands-on craft that taught me what good looks like long before any brief. It keeps me honest as a leader. I can still sit at the table and make the thing, not just have opinions about it.

AI can generate a hundred variations in the time it used to take to make one. That's not the threat people think it is. The question has never been how fast you can produce. It's whether you know which one is worth keeping. Craft, now, is taste. It's judgment. It's knowing what good looks like before the brief, before the deadline, before the stakeholder presentation. That doesn't go away. If anything, it becomes the only thing that's fully yours.

The polish in my work comes from sequence, not heroics. Decide first: the requirements, the constraints, the one question the artifact has to answer. Then render, letting the system carry those decisions. Then prove it. Verify the contrast, check the edges, put it in front of someone. Clarity leads and fidelity follows. When that order flips, you get beautiful answers to the wrong question.

Photography, music, and writing aren't hobbies bolted onto the job. They're how I train the skills the job actually demands: noticing what's in front of you, listening to several signals at once, and thinking until you have something worth saying. Creative cross-training isn't a lifestyle choice. It's a leadership practice.

The way I run product work now: small teams working in parallel, coordinated by shared signals instead of handoffs. Every piece of work answers to three questions at once. Do people want this? Can we build it? Should we? When one of those goes quiet, the silence is the finding.

AI sits inside all of it, and it earns its place by reducing time to insight, not just time to output. It synthesizes research at a scale one person can't, drafts the variations, and pressure-tests ideas against real constraints. The team owns the interpretation, and every gate stays falsifiable. Soft acknowledgements are not criteria; if a check can't fail, it isn't a check.

Prompting is the smallest part of this. The real skill is knowing what to delegate, and being able to judge what comes back.