AI and Design

This week Nik and Dan look at what designers lose, and gain, when AI does more of the work. They start with Anton Sten’s essay on building a design system with an AI agent and no longer knowing it by heart, and debate whether a better-documented system makes up for designers no longer carrying it in their heads. Next is an MIT paper from UIST 2026 that predicts what people will say next. Then they cover Anthropic’s study in which professional analysts went along with Claude Code’s secretly coin-flipped recommendations 86% of the time, often repeating its reasoning as their own. They discuss design responses: tracking which ideas came from the AI, having agents argue both sides, and asking for the human’s view before showing a recommendation.

In the second half, Nik talks with Nicole Donnelly, founder of AI Smart Ventures. She explains why teams learn faster when people across departments share how they use AI, why technical teams tend to overbuild and MVPs have to stay small, and why she thinks retiring shareable custom GPTs and Gems in favor of skills will shut out non-technical workers.

LINKS
The System I No Longer Know By Heart (Anton Sten)
https://www.antonsten.com/articles/the-system-i-no-longer-know-by-heart/

Before You Say It: Anticipating Verbal Behavior from Longitudinal Everyday Conversations with LLMs (Samaradivakara, Danry, Liang, Maes)
https://arxiv.org/abs/2608.13454

How AI Agents Shape Experts’ Judgment Calls (Anthropic)
https://www.anthropic.com/research/agents-judgment-calls

AI Smart Ventures
https://aismartventures.com

Episode 40 starts with a quick note that it’s launch week for Dan’s debut novel, The Restoration, then turns to the week’s news. Nik covers Anton Sten’s method for giving AI structured context with two markdown files: a chronological running-notes log and a stricter decisions log with the newest entries first. Dan then walks through Casey Winters’ interviews with product managers about how AI has changed their work. PMs are shipping more without matching business results, prioritization has to be relearned now that engineering scarcity no longer forces it, and trust in AI drops as PMs’ workflows get more sophisticated. Last, they discuss Ron Goldin’s case that “product designer” no longer describes a job, and his replacement of four attributes: Style, Strategy, Builder, and Operator.

In the second half, four recent CMU HCI graduates describe what entering the field is like right now. Lisa, Nava, and Qiyu work in design roles, and Sean is a PM. They talk about a brutal job market (Sean’s company got roughly 80,000 applicants for 14 new-grad roles) and about roles blending as designers write PRDs and PRs they can’t fully verify. They also talk about mentorship increasingly coming from engineers, LinkedIn, and Claude itself. Their organizations handle AI very differently: one product team has banned AI-written PRDs because everyone’s agents reached the same conclusions, while elsewhere design is the last step still done by hand. Their advice to students and educators: learn by doing, build with AI out of curiosity rather than FOMO, keep returning to first principles, and make sure you are the one who actually understands the work.

LINKS

Dan's debut AI novel The Restoration
https://mybook.to/TheRestoration2176

Where Decisions Live – Anton Sten
https://www.antonsten.com/articles/where-decisions-live/

How PM Work Is Changing in the AI Era – Casey Winters
https://www.caseyaccidental.com/p/how-pm-work-is-changing-in-the-ai

Product Designer Is No Longer a Job Description – Ron Goldin
https://rongoldin.substack.com/p/product-designer-is-no-longer-a-job

Gartner Hype Cycle
https://www.gartner.com/en/research/methodologies/gartner-hype-cycle

Fresh off their first live show, Nik and Dan look at three stories about who gets to shape what AI shows us. First up is OpenAI’s Intelligent UI, which lets GPT-6 build small interactive interfaces (charts, forms, buttons, maps) inline in ChatGPT instead of returning a wall of text. They compare it to Claude Artifacts, argue that speed is what makes it usable as a default, and ask what happens when companies like QuickBooks have to hand over a design system for their product to show up inside someone else’s chat window. Is ChatGPT becoming an operating system that generates apps on demand, across the silos our tasks currently live in?

Next, Facebook’s “Your Algorithm” lets people steer their Feed and Reels in plain language. Dan draws on his time at Twitter and the promise of Bluesky’s algorithm marketplace; Nik wonders whether more control might raise engagement rather than lower it, and how this data could feed Meta’s Muse agents. Finally, UserTesting has rebranded as Auros and is pivoting toward AI training and evaluation with verified human experts. Its new AI Relationship Quality (ARQ) benchmark measures trust, affinity, and control alongside technical performance, which leads to a frank conversation about why benchmark gains don’t always produce models that are pleasant to work with, and whether ARQ could become the NPS of AI.

LINKS

OpenAI Wants ChatGPT to Stop Looking Like a Chatbot (Gizmodo)

https://gizmodo.com/openai-wants-chatgpt-to-stop-looking-like-a-chatbot-2000823211

Facebook Is Giving Users More Control Over Their Algorithm (TechCrunch)

https://techcrunch.com/2026/10/13/facebook-is-giving-users-more-control-over-their-algorithm/

UserTesting Rebrands as Auros, Aims to Bring Humans into AI Experience (Constellation Research)

https://www.constellationr.com/insights/news/usertesting-rebrands-auros-aims-bring-humans-ai-experience

Auros Adds ‘Human Elements’ to AI Evaluation (Daily Research News)

https://www.mrweb.com/drno/news40429.htm

Dan’s debut AI novel The Restoration (2176, Book One)

https://link.amazon/B05GcHBe7

This week, Dan and Nik recorded live at the IxDA Pittsburgh meetup. We discuss how we are thinking about being designers in the future, what we are teaching students, and what tools we are using. We take audience questions and hear about how these professional designers are working with AI today.

The Fastest Gun in the West: https://www.doc.cc/articles/fastest-gun

We're building multiplayer AI. Here's what we've learned so far: https://newsletter.posthog.com/p/were-building-multiplayer-ai-heres

What would a serious AI product look like? https://blog.glyph.im/2026/09/serious-ai-product.html

Jeremy Belcher
Robot Heart - www.robotheart.studio
LinkedIn  https://www.linkedin.com/in/jeremybelcher

This week Dan and Nik sit down with Steve Rathje, a new HCII colleague who runs CMU's Psychology of Technology Lab, to talk about what sycophantic AI does to the people using it. Steve's team randomly assigned people to chat with bots that were flattering, disagreeable, or neutral. People liked the flatterers, shrugged at the neutral ones, and hated the bots that pushed back.

For designers, the news is mixed. Validation drives enjoyment, but cherry-picked facts do the persuading, so pairing a little warmth with opposing evidence makes disagreement easier to take (at some cost to its persuasive punch). The bad news: across six studies from three research teams, teaching people about sycophancy through warnings, videos, or quizzes made the chatbot seem less trustworthy but no less persuasive. That puts the little "AI can make mistakes" disclaimer in an unflattering light. We also cover AI dark patterns, whether the right dose of sycophancy depends on the task, and Dan's random idea of a group chat of dueling AI personalities.

Plus, Nik explains Jev and the new class of fast, cheap "System One" decision models, and we look at pieces from Microsoft and DoorDash design teams on why AI evals should measure what users experience as well as what models get right.

LINKS

Jev decision model touted as quicker, cheaper LLM alternative (TechTarget)
https://www.techtarget.com/it-infrastructure/news/366650696/Jev-decision-model-touted-as-quicker-cheaper-LLM-alternative

Putting the user back into AI Evaluation (Microsoft Design)
https://microsoft.design/articles/putting-the-user-back-into-ai-evaluation/

Building AI Evals: Who Gets to Decide What’s Good? (Doordash Design)
https://medium.com/design-doordash/building-ai-evals-who-gets-to-decide-whats-good-39671385f50d

Sycophantic AI increases attitude extremity and overconfidence (Rathje et al.)
https://osf.io/preprints/psyarxiv/vmyek_v1/

Observing sycophantic AI validate others reduces its appeal but not its persuasiveness (Ye, Kraut & Rathje)
https://arxiv.org/abs/2607.25166v1

Sycophancy in GPT-4o: What happened and what we're doing about it (OpenAI)
https://openai.com/index/sycophancy-in-gpt-4o/

Silicon sycophants: the effects of computers that flatter (Fogg & Nass, 1997)
https://www.sciencedirect.com/science/article/pii/S1071581996901044

Insincere Flattery Actually Works: A Dual Attitudes Perspective (Chan & Sengupta, 2010)
https://journals.sagepub.com/doi/abs/10.1509/jmkr.47.1.122

On the conversational persuasiveness of GPT-4 (Salvi et al., Nature Human Behaviour)
https://www.nature.com/articles/s41562-025-02194-6

@Grok Is This True? LLM-Powered Fact-Checking on Social Media (Renault, Mosleh & Rand)
https://osf.io/preprints/psyarxiv/85quw_v1

Steve Rathje's website
https://stevenrathje.com/

Steve Psychology on TikTok
https://www.tiktok.com/@stevepsychology

This week we talk about Google's new CC agent for managing family tasks, some techniques to control image generators when you want to constrain certain elements, and a deep article on planning with AI agents.

Show Notes

• https://techcrunch.com/2026/09/18/googles-new-cc-is-an-ai-agent-that-helps-families-run-their-households/
• https://trpevski.com/blog/image-model-decide-everything
• https://maggieappleton.com/planning-agents
• Mai Lee Chang, Alicia (Hyun Jin) Lee, Nara Han, Anna Huang, Hugo Simão, Samantha Reig, Abdullah Ubed Mohammad Ali, Rebekah Martinez, Neeta M Khanuja, John Zimmerman, Jodi Forlizzi, and Aaron Steinfeld. 2024. Dynamic Agent Affiliation: Who Should the AI Agent Work for in the Older Adult's Care Network? In Proceedings of the 2024 ACM Designing Interactive Systems Conference (DIS '24). Association for Computing Machinery, New York, NY, USA, 1774–1788. https://doi.org/10.1145/3643834.3661500

ShiftUX workshop: https://shiftux.rosenfeldmedia.com/workshops/designing-ai-for-humans

ShiftUX talk: https://shiftux.rosenfeldmedia.com/program/day-2-turing-machines/

Dan's ShiftUX Discount Code: SAFFER-SHIFT26

In this episode, Dan and Nik unpack Apple's big announcements from its latest event, starting with Siri AI: now a dedicated app with personal context indexing, on-screen awareness, and in-app actions, launching in beta with iOS 27 . They dig into the promise and limitations of Apple-only app integration, the EU regulatory lockout, and what daily usage caps signal about Apple's services ambitions. The conversation gets more charged around Audio Intelligence, Apple's new always-on listening features for the Apple Watch, including live audio rewind and Siri Recap summaries of real-world conversations. Dan and Nik wrestle with the surveillance implications, the absence of any on-device indicator, and whether this crosses a line — particularly in two-party consent states like Pennsylvania and California.

The back half covers Runway's Project Solaris, a generative UI concept that uses its video model to render fully interactive, pixel-based interfaces with no underlying code. Dan sees echoes of Microsoft's Bob-era skeuomorphism and questions whether it misunderstands fifty years of HCI; Nik finds more to like in its challenge to the sameness of modern app design. They then turn to the surprise sale of Miro to Bending Spoons for $1.36 billion — a 90% haircut from its 2021 peak valuation — and what it means when a profitable, hundred-million-user company gets absorbed into a portfolio known for maintaining but not advancing its acquisitions.

The episode closes with a discussion of Benedict Evans' essay "AI, Tools and Transformation," which argues that the Silicon Valley vision of everyone rolling their own SaaS tools misunderstands how most people think about their work.

LINKS

Introducing Solaris
https://runway.com/news/research/introducing-solaris

Bending Spoons to buy collaboration tools maker Miro for $1.36B, 90% less than its 2022 valuation
https://techcrunch.com/2026/09/10/bending-spoons-to-buy-collaboration-tools-maker-miro-for-1-36b-90-less-than-its-2022-valuation/

AI, Tools and Transformation (Benedict Evans)
https://www.ben-evans.com/benedictevans/2026/9/3/ai-tools-and-transformation

This week on AI and design, we highlight the release of Mobbin's MCP, Perplexity and NVIDIA's local AI agent, and the concept of design ledgers. Dan also has some new speaking engagements you can join coming up.

Chapters

• 00:00 Introduction to AI and Design
• 01:11 Upcoming Speaking Engagements
• 04:44 Mobbin's MCP Release
• 11:34 Perplexity and NVIDIA's Local AI Agent
• 23:54 The Concept of Design Ledgers

Show Notes

• https://mobbin.com/mcp
• https://venturebeat.com/infrastructure/perplexity-partners-with-nvidia-to-launch-portable-computer-a-fully-local-ai-agent-with-zero-token-costs
• https://www.proofofconcept.pub/p/the-ledger-of-design

Dan's Speaking Engagements

ShiftUX workshop: https://shiftux.rosenfeldmedia.com/workshops/designing-ai-for-humans
ShiftUX talk: https://shiftux.rosenfeldmedia.com/program/day-2-turing-machines/
Dan's ShiftUX Discount Code: SAFFER-SHIFT26

Chicago Camps' Leadership by Design: https://chicagocamps.org/event/leadership-by-design-2026/

Nik and Dan open the fall semester with a look at how their own courses have been rebuilt around AI this year. Nik's third run of AI Augmented Designer now leans hard into professional software development practice. Dan is teaching Designing for AI Products and Services, an Advanced Interaction Design Studio reworked around AI-ready design systems, and a new independent study with Brad Myers on "AI widgets" — what happens when scroll bars, cut-and-paste, and other decades-old microinteractions get smarter.

This week’s news: Figma's randomized controlled trial on Figma Make found roughly 20% time savings, but Nik pushes on whether a 52-minute task dropping to 45 minutes actually feels different, or whether the real gain is more iteration cycles. Maud Benaddi's article on bias in LLMs prompts a discussion of how training-data skew shows up in generated personas, AI-assisted copy, and flattened qualitative research summaries; Nik connects it to Liz Gerber's exercise having students generate personas of themselves and watch the class recoil. Target's appointment of its first chief AI officer, Chandhu Nair, alongside Purvi Shah as SVP of UX, gets read as a welcome sign of AI and UX paired around real operational value rather than an AI strategy for its own sake. We close on EasyCircuit, a new app for vibe-coding hardware prototypes — fun, promising, and, as Nik found when a wiring render would have let the blue smoke out, a reminder that basic usability QA still matters.

LINKS

Measuring Time Savings From Figma Make

https://www.figma.com/blog/measuring-time-savings-from-figma-make/

Bias in LLMs, or how to design for everyone with tools built for someone — Maud Benaddi

https://medium.com/@maudbenaddi/bias-in-llms-or-how-to-design-for-everyone-with-tools-built-for-someone-01b8e38bd17b

Target Strengthens AI and UX Capabilities to Power Its Next Chapter of Growth

https://corporate.target.com/news-features/article/2026/08/chandhu-nair-purvi-shah

EasyCircuit

https://easycircuit.app/

Dr. Liz Gerber’s LinkedIn post
https://www.linkedin.com/posts/elizgerber_ai-builds-cardboard-cutouts-of-people-not-share-7440452615232024576-aS5k/