Sascha Pohflepp's The Golden Institute (2009) imagines an alternate history in which Jimmy Carter wins a second term and renewable energy receives the scale and ambition of the Apollo program. In the interest of keeping Pohflepp's pathbreaking corpus alive, watch the film and revisit the project through his archive and this account of the work.

I'm curious to build on this work and develop further a world in which energy policy evolved from the moment of the pivot that is central to this work. Suppose we got Solarpunk instead of Cyberpunk?

Dubai Future Horizons brought together more than 350 people, joining from around the world, for a conversation with John Katsos and me about the kinds of people we might become alongside AI. John speaks to the moral muscle of keeping one’s own judgment under pressure; I talk about imagination as a human practice—the ability to see other ways of being and knowing. We also spend time with a risk that feels close at hand: ceding the difficult, uncertain part of thinking to a chatbot.

It’s a generous, wide-ranging hour, and a useful indication of the creative R&D work I bring to organizations: using research, storytelling, and artifacts to make consequential possibilities available for examination.

The AI Developer's Field Guide gives developers, engineering managers, and technical leaders a role-playing-game-inspired vocabulary for recurring failures in AI-assisted coding. O'Brien organizes developer archetypes, artifacts, and project-scale monsters to examine overconfident merging, ornamental architecture, fragile fixes, and unwarranted trust in automation.

IMAGINE THIS AT OpenAI

Applied Speculative Design, Generative Brand Experiences, Consumer Products

I had some peripheral insights on Black Mountain College over the years, but it wasn't until I started actively collaborating with Carl DiSalvo that I got a chance to be taken deeper into the whole vibe of the joint. Check out this recent post from Carl on that Substack just around the corner.

Black Mountain College ran from 1933 to 1957 in the Blue Ridge Mountains near Black Mountain, North Carolina, built around inquiry-led learning and a liberal arts education where the arts were central. Founded by John Andrew Rice after he was dismissed over academic freedom disputes, it drew major figures like Josef and Anni Albers, composer John Evarts, philosopher Theodore Dreier, and visiting faculty such as John Cage and Buckminster Fuller, along with students including Robert Rauschenberg, Cy Twombly, and Ruth Asawa. The school’s learning model was unusual: students moved between divisions by taking an open-note, multi-day exam followed by an oral test, with questions designed to probe how students think and investigate, not just what they memorized.

Sounds fun — but Carl makes the point that it's was a mess, which almost sounds like a joint like that should be.

One little aspect of the set up that I can relate to is the way the studies at any particular moment were based on who was there at the time. So..no generic History 101 or whatever. Reminds me of my time at History of Consciousness there at UC Santa Cruz, which wasn't the same kind of mess, but had a dynamic and fluid curriculum based on he faculty there at the moment/era.

Friend of the Blog and Patreon supporter Tim O’Brien shares Human Quotient, a self-declared 0–10 scale for naming how AI was involved in a piece of work. Tim rates his post HQ(10). Human Quotient distinguishes process from quality and gives the maker a way to describe how the work was made.

Curious to see things like this crop up — and of course wondering about what they become. Text and images are one thing, but what happens when we apply it to fine dining, music, pet DNA? And is there a near future in which the integration of AI into life is just like..hygiene?

Wasn't there a time when photography was painting's cheating cousin? (I seem to recall that from Prof. Berger's History of the Camera Image seminar back there at UW..

In a corporate setting, design fiction can help teams think critically about future products—yet it’s hard to keep the work genuinely speculative once business requirements kick in. Over more than a year, Mercedes-Benz collaborated on a near-future self-driving car scenario, generating stories and 24 artifacts like city maps and quick-start guides. The project revealed what made the process work, where it stalled, and how product-centric goals shaped the final direction. Instead of mainly provoking questions, the output leaned toward confirming and justifying innovation through specific, market-friendly implications.

Security and privacy problems keep showing up after products ship, often because teams can’t reliably predict threats and social-context risks ahead of time—especially with a shortage of security experts. The authors suggest a workaround: use “design fiction” grounded in science-fiction storytelling to help developers visualize future use, contexts, and failure modes. They point to a pilot effort that used simplified threat modeling plus story samples drawn from existing science fiction to support threat assessment for Health Internet-of-Things devices, with encouraging early results. The goal is a practical method that can be adapted beyond healthcare and IoT to other software domains.

Joseph Lindley and Paul Coulton argue that “imaginary abstracts” (fictional problem frames, prototypes, user studies, and findings included inside otherwise real research papers) can be extended into fully fictional papers. They compare this approach to “Game of Drones,” a paper that stays fictional from start to finish, revealing its fictional nature only at the end, and they weigh the merits and weaknesses of both tactics. The core claim is that fictional research—used tactfully and carefully—can strengthen HCI’s design discourse by offering compelling new methods for thinking about technology’s real-world implications.

General McCallum's org chart for operational clarity of the railroad network. Described as the first organizational chart (bit of a stretch to make that claim), but no matter..the point here is that this looks futuristic in contrast to all the usual right angles one typically sees.

The Office Hours had the good fortune to do a run of the workshop that Shreya Thakkar and Sneha Arvind are proposing for SXSW 2026 — and it was a fun structure and a great way to explore the implications of AI in learning. (cf Office Hours N°317)

The workshop is designed to help participants think through the implications of AI in learning and to create speculative artifacts that challenge what learning with AI should look like.

Over-reliance on frictionless AI can lead to cognitive downsides like overconfidence and weaker independent performance, even if it feels helpful. This SXSW EDU workshop time-travels to a 2045 classroom to create speculative artifacts that challenge what learning with AI should look like. Through debate and hands-on facilitation, participants reverse-engineer today’s critical product choices and leave with a facilitation toolkit and frameworks to anticipate unintended consequences—grounded in productive struggle pedagogy and emerging AI learning research.