AI-Driven Prototyping: Simon Willison on Abandoning Rigid Design Processes

A recent post by Simon Willison highlights Jenny Wen’s Hatch keynote urging teams to favor rapid, AI-enabled prototypes over lengthy, process-heavy design rituals. He argues AI lowers the cost of being wrong, letting teams iterate faster and experiment more boldly.

AI-Driven Prototyping: Simon Willison on Abandoning Rigid Design Processes

TL;DR

  • Shift from long upfront process to rapid, AI-enabled prototyping — favor quick, tangible prototypes over user research → personas → user journeys → wireframes before build
  • AI-assisted programming lowers the cost of building the wrong thing, letting missteps be discovered in days instead of months
  • Faster prototypes shorten feedback loops and reduce sunk cost
  • Lower cost of failure enables broader experimentation and bolder idea exploration
  • Working prototypes become the central artifact for more iterative design–engineering collaboration

Simon Willison’s short post on Jenny Wen’s Hatch Conference keynote, “Don’t ‘Trust the Process’”, flags a practical shift in product design thinking: move away from long, upfront process-driven workflows and toward rapid, AI-enabled prototyping.

What Wen argued

Jenny Wen suggested that the conventional sequence — user research → personas → user journeys → wireframes before any build — is increasingly out of step with a world where anyone can make anything. The recommendation is to lean into prototypes: iterate quickly, test ideas in tangible form, and use that feedback to guide decisions instead of committing to lengthy, pre-build design rituals. AI is a key enabler, because it makes prototypes faster and less expensive to produce.

Why it matters

Simon picks up the practical implication for engineering teams: AI-assisted programming lowers the cost of building the wrong thing. Where a misaligned design might once have meant months of wasted development, the same misstep can now be discovered in days. That change shifts the risk calculus—encouraging more experimentation, faster learning, and broader exploration of the problem space. For teams that already prototype frequently, the case for integrating AI into early-stage workflows is especially strong.

Implications for workflows

  • Faster prototypes shorten feedback loops and reduce sunk cost.
  • Lower cost of failure enables bolder exploration of ideas.
  • Collaboration between design and engineering can become more iterative, with working prototypes serving as the central artifact.

Context

The keynote took place at Hatch Conference in Berlin. Jenny Wen is Design Lead at Anthropic and previously led design at Figma. Simon’s reflection was posted on 24th January 2026 and connects Wen’s argument to trends in AI-assisted programming and prototyping.

Read the original reflection at Simon Willison’s site: Don’t “Trust the Process”.

Continue the conversation on Slack

Did this article spark your interest? Join our community of experts and enthusiasts to dive deeper, ask questions, and share your ideas.

Join our community