Lenny's Podcast: Product | Career | Growth
The design process is dead. Here’s what’s replacing it. | Jenny Wen (head of design at Claude)
This debrief was written and voiced by AI from the episode's audio.
🎧 AI Debrief
🎙️ Full Episode
Summary
Jenny Wen, Head of Design for Claude CoWork at Anthropic (formerly Director of Design at Figma), joins Lenny's Podcast to discuss how AI is fundamentally reshaping the design profession. The core thesis: the classic discovery → mock → iterate design process is effectively dead, not because designers chose to abandon it, but because engineering velocity driven by AI tools has forced the change. Designers who once spent 60-70% of their time on mocks and prototypes now spend 30-40%, with the rest shifting toward direct engineering collaboration and implementation. Jenny frames modern design work as splitting into two modes: (1) supporting rapid execution alongside engineers who are shipping constantly with AI assistance, and (2) creating shorter-horizon visions (3-6 months vs. the old 2-5 year decks) that point teams in a cohesive direction. She shares what a day at Anthropic actually looks like — heavy on Slack-monitoring for emerging prototypes, jamming with engineers, coding last-mile polish, and occasional traditional design work — all through a fully Claude-powered stack. The conversation covers hiring archetypes for the new design era (strong generalists, deep specialists, and overlooked 'craft new grads'), the value of managers returning to IC work to stay relevant, unconventional management philosophies like embracing psychological safety through humor, and the 'legibility framework' for spotting illegible ideas with hidden energy. Jenny is candid that AI will get better at taste and judgment, but argues human accountability for decisions remains the irreducible core of the role — for now.
Key Takeaways
The traditional design process (research → diverge → converge → mock → iterate) is effectively dead. Engineering velocity powered by AI has made it obsolete, and designers must let go of it rather than protect it.
highDesign work has split into two modes: (1) supporting rapid engineering execution in real time, and (2) creating short-horizon visions (3-6 months) via prototypes — not beautiful decks. Time spent mocking has dropped from 60-70% to 30-40%.
highDesign managers should seriously consider rotating back to IC work to maintain credibility and empathy. The design process has changed so dramatically in the past year that managing without hands-on experience means leading blind.
highThe three most valuable designer archetypes right now: (1) Strong generalists — near-block-shaped skill profiles across multiple domains; (2) Deep specialists — top 10% in one area; (3) Craft new grads — blank-slate learners unburdened by outdated processes.
highBuilding trust through speed is the new quality strategy: ship early with a 'research preview' label, respond visibly to feedback, and iterate publicly. The brand killer isn't launching rough — it's launching and going silent.
mediumAI will get meaningfully better at taste and judgment. Designers (and PMs) may be over-indexing on taste as a permanent human moat. The durable human value is accountability for decisions — someone still has to be responsible for what gets built.
mediumThe 'legibility framework' (from Evan Tana at SPC) is a powerful tool for designers at frontier companies: actively seek out internally illegible ideas that have unexplained energy around them — those are often where the most valuable product directions are hiding.
mediumNotable Quotes
“This design process that designers have been taught, we sort of treat it as gospel. That's basically dead.”
— Jenny Wen
Opening thesis of her Berlin conference talk and the central argument of the episode — the traditional design process is no longer viable in an AI-accelerated engineering environment.
“It's not just designers who are feeling like, oh yeah, we have to keep up with engineers. I think even engineers are like, how do we keep up with ourselves right now?”
— Jenny Wen
Illustrates that the pace of AI-driven change is outrunning everyone, not just designers — even engineers struggle to keep up with their own AI agents.
“At the end of the day, someone has to decide what is actually going to get built and what actually matters. Someone still needs to be accountable for the decision.”
— Jenny Wen
Jenny's core argument for sustained human relevance in design and product — accountability is the irreducible human function even as AI gets better at taste and judgment.
“The way you really lose trust around quality and releasing something early is if you release it early and then nothing ever happens. That is what degrades a brand.”
— Jenny Wen
Her practical framework for shipping rough products without destroying brand trust — speed of iteration after launch is the credibility signal, not perfection at launch.
“Given how much the roles are changing and what we're expected to do is changing, having somebody who almost has a blank slate and is just a really quick learner... doesn't have all these baked-in processes and rituals in their mind — that's super valuable.”
— Jenny Wen
Counterintuitive hiring argument: in a rapidly changing environment, early-career designers without entrenched habits may outperform experienced seniors.
Predictions
Design tooling will change significantly in the next year, currently trailing behind the transformation already seen in engineering tools.
— Jenny Wen · Confidence: high · Within the next year
AI will get substantially better at taste and judgment in design — designers may be over-relying on taste as a permanent human competitive advantage.
— Jenny Wen · Confidence: medium · Unspecified, but implied near-to-medium term
Generated UIs — interfaces dynamically created by models rather than hand-coded — will become increasingly common, blending the chat paradigm with traditional UI.
— Jenny Wen · Confidence: medium · Unspecified
Pure people management roles in design (and possibly broadly) will diminish — the future manager must also drive substantive work direction, not just career development.
— Jenny Wen · Confidence: medium · Already underway, accelerating
IDEs may become more useful to designers and PMs doing light implementation work as engineers shift to agent-based, command-line workflows and move away from traditional IDEs.
— Host (Lenny) · Confidence: low · Unspecified
Action Items
Audit your current design time allocation and deliberately shift toward direct engineering collaboration and implementation, even if it means less time on mocks.
Jenny's data: mocking/prototyping went from 60-70% to 30-40% of her time. Designers who haven't made this shift are likely blocking rather than enabling their teams.
If you're a design manager, take an IC rotation — even a few months — to rebuild firsthand understanding of how the tools and process have changed.
Jenny credits her IC year at Anthropic with giving her hard skills and empathy she couldn't have gained by managing alone during this period of rapid change.
Start using Claude Code or similar AI coding tools for last-mile design polish (CSS tweaks, layout fixes) rather than routing everything through engineers.
Jenny does this directly in VS Code — it speeds up execution and gives designers more agency over the final shipped product.
When hiring designers, reconsider your bias toward senior candidates. Evaluate 'craft new grads' — early-career designers who are fast learners without entrenched process habits.
Jenny argues these candidates are systematically overlooked but uniquely suited to the current moment of process disruption.
Shorten your design vision horizon to 3-6 months and express it as a working prototype rather than a polished deck.
With technology changing too fast to predict 2+ years out, long-horizon vision decks are becoming misleading artifacts. Prototypes communicate direction more honestly.
Adopt a 'research preview' launch strategy: ship rough but valuable, label it clearly, respond publicly to feedback, and demonstrate rapid iteration.
Jenny's framework for maintaining brand trust while moving fast — Claude CoWork is the case study.
Apply the legibility framework internally: pay attention to prototypes or ideas that generate energy but that you don't fully understand yet — investigate rather than dismiss them.
Jenny used this to spot the precursor to CoWork's skills framework inside an internal prototype that initially looked confusing.
Links Mentioned
Mercury — banking platform for startups and entrepreneurs
https://mercury.com
Orkes — enterprise workflow orchestration platform (open source Conductor)
https://orkes.io/lenny
Omni — AI-powered embedded analytics with a semantic layer
https://omni.co/lenny
Lenny's Product Pass — exclusive deals for Lenny's newsletter subscribers
https://lenny.productpass.com