Dive Club 🤿

Episode: Stephen Haney - The 2026 AI Design Field Report (tools, process, and what's working)

Date: January 19, 2026
Host: Rid
Guest: Stephen Haney (Founder of Paper)

Episode Overview

This episode of Dive Club features Stephen Haney, founder of the AI-native design tool Paper, sharing insights from extensive field research into how leading design teams—including those at Atlassian, Shopify, Vercel, and Notion—are really using AI in their workflows. Haney and Rid go deep on the latest tools, team processes, what's working versus what's hype, and how these trends are shaping design careers and tools. The conversation addresses the evolving role of designers, the new landscape of prototyping, the reality of AI mandates, and Haney's vision for the future of design tools.

Key Discussion Points & Insights

1. Current State of AI in Design Teams

• What Teams Are Actually Doing
  • Leading companies are largely converging on similar use of AI tools, regardless of what social media suggests.
  • “[...] what I see these companies actually doing is not necessarily matching up to what you'll read on Twitter.” – Stephen Haney (00:12)
• Performance Reviews Mandate AI
  • 100% of designers at these companies are now required to use tools like Claude Code or Cursor.
  • The reason: Companies want to break designers out of old habits and accelerate exploration with AI.
  • “AI usage is being mandated in performance reviews for designers, full stop. 100% of designers need to use a Claude code, a cursor in their work.” – Stephen Haney (12:51)
• Prototyping and Handoff Still Matter
  • AI enables richer, interactive prototypes (especially for stateful or focused flows).
  • Handoff is still alive—designers aren’t PRing to prod; rather, they create specs with more detail.
  • “I'm really not seeing much PR into production and I'm not even seeing designers wanting to PR to production. [...] PR into production comes with maintenance burdens. It comes with a different type of accountability.” – Stephen Haney (17:03)

2. Tools and Workflow Shifts

• From Canvas to Code-First Prototyping
  • The big unlock: Using the real app (or a copy/fork) as the “starting place,” allowing designers to prototype on live code instead of static canvases.
  • “That starting place that you get is your real app...you can just ask to change it.” – Stephen Haney (15:11)
• Technical Barriers and New Frictions
  • Local development with code brings new challenges: setting up environments, dealing with linting, lack of real-time sharing, and the need to maintain “designer forks” of the codebase.
  • “Someone's got to maintain it ... somebody has to build that, right? Like that's a cost that they're, they're investing in and maintaining.” – Stephen Haney (20:42)
• Cloud vs. Local Tools
  • Local tools like Cursor, Claude Code, and “cloud code” are far more prevalent in these orgs than web- or cloud-hosted prototyping tools.
  • Cloud tools are still used at startups, where designers tend to be more full stack by necessity.
• The Reality of Collaboration
  • Real-time, canvas-based collaboration remains a pain point in code-based workflows, and companies are actively building internal tools to bridge the gap.

3. Design & Engineering: Distinct but Overlapping

• Specialization Remains Key

  • Even as AI blurs some roles, most large organizations still maintain clear lines between design and engineering due to the scale/specialization of the work.
  • “Design and engineering are not the same thing. [...] Specialization is good. [...] Design is very much an expansion of problem space.” – Stephen Haney (02:08)

• The “Continuum” of Project Work

  • Haney introduces a mental model where design and engineering are different points along a continuum—from low-fidelity ideation (visionary/founder/PM) to high-fidelity engineering/reliability.
  • Different tooling and input methods are optimal at different stages of this spectrum.

4. AI’s Impact on the Designer's Role

• More Ways to Explore and Prototype

  • AI augments, rather than replaces, the design process—particularly for interactive prototyping and rapid iteration.
  • Designers are discovering new ways to communicate nuance (like focus, state, drift) to engineers via AI-powered prototypes.
  • “Any tool that removes constraints, that lets you be more free and explore more rapidly and explore the problem space is good for design.” – Stephen Haney (08:19)

• New Problems: Quality vs. Speed

  • Concerns arise whether AI-driven prototypes are actually faster or higher quality than old methods, despite the new powers unlocked.
  • “Was that actually faster than the old way? Not sure not sure. [...] We will find new flows that work and are actually advantageous.” – Stephen Haney (22:59)

5. Future of Design Tools & Product Strategy (Paper)

• Agents as the Central Orchestrator
  • Haney predicts agents (e.g., AI coding assistants) will become the core, with designers interacting via multiple “input modalities”: canvas tools, IDEs, terminals, and direct prompt interfaces.
  • “You’re going to have these different input tools to a centralized agent, probably, which is kind of like cloud code” – Stephen Haney (09:36, 34:31)
• Paper's Evolving Vision
  • Moving towards true interoperability—desktop app for local file system access, real-time collaboration, seamless integration between canvas, code, and AI agents.
  • Canvas will remain essential, especially for context management and design exploration (multiplayer, visual context for agents, etc.)
  • Emphasis on openness: “We want to make paper very open and very programmable. And everything’s an API.” – Stephen Haney (31:54)
• Partnership with Tailwind
  • Tailwind’s influence means Paper’s components/themes will be more aligned with engineering realities; direct copying of React components with true Tailwind classes.
  • “LLMs use Tailwind, like, 100% of the time, right back to the network effects [...] We'll add theming and our tokens are going to be based on the tailwind ways of doing tokens” – Stephen Haney (39:45)
• Canvas Context, Not Always Source of Truth
  • Having the canonical representation of an app in the design tool is becoming less critical, as everything becomes more disposable and context can be exchanged with code/AI.
  • “Designs being disposable is like a good thing... The ability to bring stuff back where it is, like which tool it's in or what format it's in is going to matter less” – Stephen Haney (37:54)

6. Reflections, Advice, and Culture

• Learning vs. Hype
  • Advice: Try everything and skill up, but don’t stress about being left behind or buy into FOMO.
  • “Skill up. 100% skill up. Find time to go play with these tools, even if you don't need to use them in your job yet, they are going to matter.” – Stephen Haney (41:55)
  • Even experts feel overwhelmed: “If [Andrej] Karpathy is saying he feels behind, everyone feels that.” – Stephen Haney (46:15)
• Diversity in Process
  • Designers today must be adaptable—no “one right design process,” and each project may need a different approach.
  • “A big part of my practice as a designer is kind of at each new release or feature [...] is taking a second to think about what's my process for this initiative, because it might look completely different than the last thing I did.” – Rid (44:01)
• Still Early Days
  • “No one has the answers right now. [...] All of our explorations are going to uncover new ways of working that are really cool. So just be part of that and enjoy it.” – Stephen Haney (41:55)
  • Embrace experimentation—some “shiny things” will turn out to be dead ends, and that’s part of progress.

Notable Quotes & Memorable Moments

• On Specialization and the Value of Design:

  “Design is very much an expansion of problem space. It’s exploring the problem space. [...] Figma won was how much they helped us show collaboration being important.”
  — Stephen Haney (02:43)

• On AI Changes:

  “The first big takeaway is that AI usage is being mandated in performance reviews for designers, full stop.”
  — Stephen Haney (12:51)

• On Interactivity & Tooling Friction:

  “The sharing is really painful. There's no real time collaboration. [...] That flow is built for, like, engineering. Right. And so there's just parts of it that are not helpful for design.”
  — Stephen Haney (22:20)

• On Network Effects in Tooling:

  “There's such a network effect happening right now around that modern web stack because the models are so good at it, which makes more people want to go to it, which makes the models better at it.”
  — Rid (28:01)

• On Real-World Prototyping Limits:

  “I have definitely fallen into the trap of chasing the shiny thing for the sake of tinkering and exploration...I wasn't even solving the right problem.”
  — Rid (25:37)

• On Culture of Experimentation:

  “Times of change... are the times with the greatest opportunity, times where people can make a name for themselves.”
  — Stephen Haney (44:40)

• On Universal Designer Anxiety:

  “Karpathy — he's like the least behind person in the world. If he's saying he feels behind, everyone feels that.”
  — Stephen Haney (46:15)

• On What’s Still Undeniably Cool:

  “I've been sent three different test flights by designers who didn't write, like, handwrite a single line of code. And they built entire mobile apps over the Christmas break. And it's cool, man.”
  — Rid (46:53)

• On Staying Open:

  "We want paper to be, like I said, an open platform, an open tool. And I also want to hear where I'm wrong. So if you, if I said something today that like, you as a listener disagree with, like, shoot me a DM..."
  — Stephen Haney (47:28)

Timestamps for Key Segments

| Time | Segment | | ----------- | ------------------------------------------------------------------------------------- | | 00:00–03:00 | Debate on future of coding in design; design vs. engineering; the “continuum” model | | 07:56–09:36 | Double Diamond; value of tool diversity; Basecamp ‘hill charts’ and Shape Up | | 12:31–15:21 | Field report: how AI is really being used by top teams | | 15:21–18:52 | Local dev vs. cloud tools, reality of “designer forks” | | 22:07–25:37 | Problems with prototyping, limitations of sharing, and impacts on design workflows | | 29:08–30:53 | Network effect around modern web stack and AI modeling | | 31:14–34:31 | Product strategy at Paper: openness, API-first, agent orchestration | | 37:52–39:45 | Source of truth shifting; format/medium is less important with LLMs, design is disposable | | 39:45–41:42 | Tailwind partnership and its implications | | 41:55–44:40 | Career advice, learning vs. hype, process diversity | | 46:08–46:53 | Universal feeling of being behind; embracing change |

Summary

This episode paints a nuanced, up-to-the-minute picture of what’s working—and what isn’t—as designers navigate the fast-changing landscape of AI, design tools, and collaboration. Stephen Haney’s “field report” cuts through hype, highlighting both the potential and the pain points of AI-enabled design, and offers a hopeful, pragmatic vision: experiment, skill up, and embrace the diversity of possible workflows. Canvas and code are increasingly intertwined, and the best opportunities go to those willing to explore where new tools can take them—without forgetting what drew them to design in the first place.