Send us Fan Mail

Mark is joined by Luke Wroblewski, product leader, author and one of the clearest voices in interface design.

This conversation explores where interface design is heading as AI systems become more agentic, more capable, and more embedded in everyday workflows. Luke shares why he has long believed the goal is not better screens or cleaner layouts, but making the interface fade into the background so people can focus on what they are actually trying to achieve.

Luke explains how this shift is already happening. The prompt box may still be there, but the number of human inputs is starting to shrink as agents gather context, coordinate work, and pass information between systems. That changes the role of the person from direct operator to higher-level orchestrator, and it changes what good product design now needs to solve.

Mark and Luke also dig into the challenges this creates for UX and product teams. They discuss capability awareness, context awareness, and the overload of reasoning traces and system output that current AI products still push onto people. They also explore why designers need to get much closer to production, why static handoff culture is breaking down, and why this may be a golden age for people who genuinely love building products.

Fun Point: Mark recorded this conversation at 4 am!

What’s discussed in this episode:

00:00 - Introduction & The Future of Interfaces

02:30 - Dissolving the UI & Humanizing Technology

05:00 - The Shift to Agentic Workflows

09:12 - Object-Oriented Design vs. Pixel Manipulation

11:13 - Evolving Inputs: Less Typing, Voice, and Foot Pedals

13:54 - UI Trust & The "Pets vs. Cattle" Analogy

18:45 - The Three Core Challenges of Prompt Interfaces

23:20 - Context Awareness & Real-Time Sources of Truth

31:00 - The Evolving Role of Designers

36:40 - The Golden Age & Rapid-Fire Questions

Luke Wroblewski - https://www.lukew.com/
Luke Wroblewski is a product leader, author and long time voice in interface design. His work has consistently focused on making technology feel more human and reducing the friction between what people want to do and the systems they use to do it.

UXI (UX Institute) Original Podcast 2026.
www.uxinstitute.com

Support the show

UXI (UX Institute) Original Podcast 2025.
www.uxinstitute.com

Send us Fan Mail

In this episode of the UX Institute Podcast, Mark Swain speaks with Dan Saffer, designer, educator and a sharp voice on where AI product design is heading and where it is still falling short.

This conversation looks at the reality of using AI day to day, beyond the hype. Dan explains why so many AI products still feel like hard work, why the prompt box has become a lazy default interface, and how too much cognitive load is still being pushed back onto users. Rather than making products easier, many tools are still asking people to figure out the system for themselves.

Dan also shares how AI has become part of his own workflow, from research and writing support to building small tools for himself. But the conversation keeps returning to the bigger design question: if the models are already powerful, why does the experience of using them still feel clunky, tiring and unfinished?

Mark and Dan also explore what all of this means for designers, researchers and product managers. They unpack the rise of hybrid roles, the pressure on junior designers, the danger of letting AI do the thinking, and why understanding users, workflows and context may become even more valuable in the next few years.

In this episode, you’ll learn:

• Why AI tools still create confusion for most users, even when the underlying models are impressive
• What is broken about the prompt box as the default interface for AI products
• Why chat and voice are useful for some tasks, but poor for detailed refinement work
• How better AI products may combine invisible automation, graphical controls and conversational input
• Why AI is increasing the intensity of work instead of simply making people more efficient
• What design teams risk losing when they hand too much thinking over to AI
• Why junior designers are facing new pressure as production tasks start to disappear
• How hybrid roles across design, product and engineering are starting to emerge
• Why the biggest opportunity in AI may now be better product design, not better models
  
  

Dan Saffer

Dan Saffer is a designer, educator and thought leader focused on emerging product design patterns, AI interfaces and the future of design practice. His work explores how new technologies reshape workflows, expectations and the role of designers inside product teams.

Support the show

UXI (UX Institute) Original Podcast 2026.
www.uxinstitute.com

Support the show

UXI (UX Institute) Original Podcast 2025.
www.uxinstitute.com

Send us a text

In this episode of the UX Institute Podcast, I speak with Barry O’Reilly — author of Lean Enterprise and Unlearn, and co-founder of venture studio NobodyStudios — Barry joins us to talk about why he’s not chasing unicorns, and why “boring” AI businesses might be the smartest bet in the next decade.

Barry shares his journey from accidentally falling into programming in Dublin, Ireland to working on products such as Citysearch in the first dot-com wave, ThoughtWorks during the rise of Agile and Continuous Delivery, and finally into building a venture studio aiming to create 100 AI companies in five years. We dig into why most VC funds aren’t returning capital, why enterprise AI sales are a trap for early-stage startups, and how smart founders are using AI to build lean teams that still ship meaningful products.

We also get into why current AI UX has a long journey to go, how design will become the key differentiator on top of LLMs, and what product and UX teams might actually look like in 3–5 years.

In this episode, you’ll learn:

• How Barry went from engineering student in Dublin to author, advisor, and venture studio founder
• Why Nobody Studios focuses on health, wellness, e-commerce, and EdTech for AI opportunities
• Why “boring” but proven business models are often better than hype-driven AI ideas
• How tools like Lovable and similar “one-person billion-dollar company” platforms are great for prototyping but dangerous for production
• The studio’s strategy of building companies cheaply and selling early, instead of chasing unicorns
• Red flags in AI startups: vanity metrics, no learning loops, brittle tech stacks, and vague “proprietary data”
• Why early-stage founders should avoid long enterprise sales cycles and niche down to a very specific use case
• How Barry uses his own AI stack (meeting copilots, scripts, automation) to dramatically increase personal and organisational productivity
• Why UX and product design on top of AI will likely be the biggest competitive advantage in the next 5–7 years
• How roles for designers and product managers may shift away from pushing pixels in Figma toward higher-level problem solving and LLM-powered workflows
  
  

Barry O’Reilly

Author of Lean Enterprise and Unlearn, entrepreneur, advisor, and co-founder at Nobody Studios, a venture studio on a mission to build 100 companies in 5 years.

• Barry’s books: Lean

Support the show

UXI (UX Institute) Original Podcast 2025.
www.uxinstitute.com