Training Taste, AI Systems Thinking, LLM Counter Defaults, and special guest Dan Maccarone

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

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.