Yes, please. I'm not able to scroll through the entire conversation

Here's a recap of our conversation. You can copy it in one go.

We started with what the job market looks like for designers now. Bigger teams are becoming less necessary because AI lets one person produce at high pace and quality, which effectively makes everyone a design team of one. In that world your value comes from outcomes. When I hire, I look for people who are hands-on, who can operate at both the strategic and execution levels, and who have high craft. As one Design Futures discussion put it, there's nothing defensible or permanent about your position, your role, or your salary... there is but this: do great work.

You then made the case that design was originally a holistic, one-person craft that got fragmented into production roles as companies scaled, and that AI lets us return to holistic design. I agreed. AI coding agents make development so much faster that the loop has flipped, and designing software is better when you work in the medium. If the job is making mockups, it's at risk, but the difficult work is, and always has been, maintaining coherency and intention across a product so it works for people. My wrinkle: faster output also produces more features that lack UX refinement and thoughtful integration, so someone still has to see the whole.

Next, you argued that raising holistic design thinking across the industry makes quality at scale possible. I agreed it helps, but broad literacy alone can still produce products where nothing made sense together. What's new is that the quality bar can live inside the tools. A design lead and front-end lead define the grid, fonts, colors, spacing, and components once, and everyone's agents are "snapped to" it. The formula is to raise the floor with literacy and hold the line with steering.

You pointed out that education rarely raises the bar and hands-on learning wins, which AI provides. I agreed: asking people to read the manual first doesn't work, and with AI, rather than learning then doing, people observe then refine. The caveat is that practice only helps if what you practice against is good, so pair hands-on AI work with solid encoded design systems.

Finally, we discussed taste and Jared Spool's chicken sexing example. Trainees can reach 95% accuracy, but it requires a lengthy process of mentorship. So taste is hard to instruct but can be developed through reps plus feedback from people who already have it. AI speeds up the repetition, but mentors with good taste still need to be in the loop telling people when they got it right.

Two tips for next time: earlier question and answer pairs collapse on this page, so you can scan and reopen them without long scrolling. And each answer has a Share action that gives you a permanent link you can bookmark or drop into a doc.