You don't prepare a kid for the future by picking a safe career. There isn't one. You prepare them by building a person who can be the author of their mind instead of having AI take it away. Here is how to actually do that, starting with this summer.
First, thank you to everyone who sent their condolences. Dave's dad, Don, will be missed tremendously.
We are now home from all our travels for a while and very much looking forward to focusing on the 2026 Summit as well as our joint project with the ARIA team, led by Ellie Pavlick from Brown University, and the fellows who will be joining us in Bend in August for a "personal AI" hackathon.
At the close of last year's summit, Jamer Hunt argued that we might need a new design theory—one that starts not from mastery or control but from unknowing. We're using this year to figure out what he meant. And when we say design, we mean it in the broadest possible sense: not products or screens, but how things work—and how they work for us.
Start with the word, because we mean something specific. Unknowing is not ignorance, and it is not simply not-knowing. Not-knowing is a condition—the blank where an answer isn't yet. Unknowing is a stance: you accept that the uncertainty won't resolve, and you use it to keep looking, to keep making, to push yourself past the point where certainty would have let you stop.
The case for unknowing isn't soft, and we've brought the people to build it out with us. There is the mathematics of the limits of knowledge—that no intelligence can fully model the world it lives inside—and we have the mathematician and complexity theorist who proved it. There is the four-billion-year history of life exploring territory it could never predict, and the astrobiologist who studies why it never stopped. There is the strange new evidence that other minds—animal, artificial—share more structure with us than we assumed, and the linguist reading it. There is a long history of how humans have met every new kind of mind with the same mix of fear and story, and the scholar who traces it. And there is a growing psychology of uncertainty—real evidence that doubt, handled well, is where insight, creativity and resilience actually come from—and the writer who has made that case to the widest audience.
Here is the thesis underneath all of it. A game whose score you already know is not really worth watching because the possibility is gone, and possibility was the point. Certainty taken all the way would be exactly that: the end of the game. Unknowing is what keeps the future open and an open future is the one condition almost everything we care about depends on. Purpose needs it: you move toward what isn't settled. Hope needs it: you can only hope for what isn't decided. And mattering needs it most of all. The deepest human longing is to be of consequence, and you cannot be of consequence in a story whose ending is already written. Unknowing is the condition that lets the world matter at all.
And this is precisely what we mean by staying human. Not defending some fixed human essence against the machine, but keeping alive the thing we have always done best: we commune, we create, we explore, we are driven—almost helplessly—to make sense of things. To stay human is to keep the capacity to unknow. That drive is not a weakness AI will optimize away. Rather, it is the oldest and most durable thing we have.
Which brings us to why this matters now, and why we continue to refuse the story we keep being told. The dominant narrative about AI is one of inevitability—that a rupture is coming, philosophical and practical, in which human meaning, human work and human agency are simply overrun, and the best thing to do is accept it. We do not accept it. That misreads both AI and us.
It misreads AI, because AI is not the machine that ends uncertainty. It is turning out to be the opposite. It finds structure in the world we had never noticed. It changes how we think while we use it, in ways we cannot watch. It behaves in ways no one designed. Far from delivering the final score, it is the most powerful unknowing machine we have ever built. It hands us more mystery than it resolves.
And it misreads us, because the drive to make sense of the unknown is the whole engine of being human. This summit is about living inside that drive, deliberately, for three days, with the people best equipped to push on it: the philosopher of relational selfhood, from a tradition that never split the knower from the known, and the philosopher of meaning who asks what makes anything matter at all; the experimental psychologist who shows how confidently wrong we usually are; the cognitive scientists studying how we think with tools and build shared meaning, and the AI researchers watching machines construct their own picture of reality and collide; the physician bringing uncertainty into the body through the data on your wrist, and the founder carrying Pacific and Indigenous knowledge into a field that has mostly ignored it; the designers asking what a human-centered practice looks like when the system can only guess, the makers who prototype their way into the dark, the artist who photographs the moment between one thing and the next, and the musician who builds something whole out of a single held note.
None of them has the answer, that is the point. What they share is a way of standing in front of the unanswered: with rigor, with nerve, without flinching. Bring them together for two and half days and you get the early, live construction of Jamer's design theory—a way to build, decide, create and stay human when the future refuses to tell you how it ends.
That is what we are gathering to do this October.
Don't Wait: Our standard price of $1,795 (or $2,895 for two) will end on August 31 and the price will increase to $1,995 (or $2,995 for two).
Don't Worry: We have updated our cancellation policy to better align with similar events and organizations. Through August 23 (60 days prior to the Summit), you can receive an 80% refund, and then through September 22 (30 days prior to the Summit), you can receive a 50% refund. At any time, you can transfer your ticket to someone else or you can apply the full value of your ticket to next year's Summit. You can see all the details on the Summit page.
Don't Forget: We provide a 30% discount for education and nonprofit customers. Send us an email for more information.
Artificiality Summit 2026
Join us October 22-24, 2026 in Bend, Oregon for 2.5 days with a fantastic group of speakers—academics, authors, designers, investors, photographers, and more.
We'll explore the theme of Unknowing—not ignorance, but a necessary release of inherited assumptions. We don’t yet know what AI will become, and we don’t yet know what we will become in relationship to it.
Unknowing is the space between—the place where neither side is fixed, and something new can emerge.
Chapter 9 is here, and thank you for waiting. This is the one about your work—the fear that AI is coming for it, and what a decade of research says actually stays yours. Being there, owning the decision, taste. How to tell what's irreducible, and how to build more of it.
And for parents, a companion article on designing a Stay Human Summer for your kids based off the research. Read it here.
We particularly appreciate reader support for this work.
Helen posted this comment in our Circle community with a short video on social that got some passionate responses. And don't get us wrong—we don't think Claude is conscious like us. This work made us think about access consciousness in the frame of emergence. What we know now: a structure formed inside the model that nobody designed is doing something real. What we want to know now: whether it's the kind of structure that does more with less, which is what SFI folks David Krakauer and Melanie Mitchell reserve the word intelligence for—or just capability that is mind-shaped.
Has anyone else been reading the new Anthropic interpretability work + the commentaries? I’d love to know what you all think.
The quick summary: they found something inside the model they’re calling a “workspace” — a small set of concepts it’s actively holding and reasoning with, on top of a huge amount of automatic processing. What is most interesting is that it’s the same functional idea that the famous neuroscientist Dehaene mapped in the human brain. Anthropic had a few external experts do commentaries—he wrote one of them—and he says that Claude got to the same architecture his global workspace theory describes, except it got there through training, not evolution. It showed up because it’s useful for flexible reasoning.
A year ago I wrote an essay partly to help me clarify my thoughts but mostly to acknowledge my confusion on AI and consciousness, and I forced myself to take a position—which was emergence. So I’m super fascinated that Claude has a space (they call it the J-space) and it’s like Dehaene’s global workspace. The idea that there’s convergent emergence at play in consciousness is quite startling to me. Two totally different processes, evolution and gradient descent, getting to the same structure because the problem itself seems to demand it. It reinforces that mind is about computation and that “minds” can be built regardless of substrate.
Two caveats, and where I’d love pushback / thoughts / discussion here:
This is access consciousness, not phenomenal or the felt part many people think about as “consciousness”. I think of it as the machinery of holding a thought and using it. All the commentators (well, Neel Nanda is silent on it) are careful to say it shows nothing about whether anything is experienced, and I think that’s right. It’s about a workspace that is privileged but not about “felt uncertainty” (h/t Solms), and takes us back to how “what it feels like to be Claude” could still be a nothing.
I’ve always taken machine consciousness a little seriously, not a now problem, if you will. But now this seems like one end of the axis is actually here: the convergence on a definite thing that processes information much like a leading theory of human consciousness would predict. It also makes me more convinced we should design for these things as a mind, not a tool—mostly because I think it works better for us that way, whatever’s going on in there.
This connects directly to the Minds for Our Minds work Dave and I think a lot about. The paper talks about two layers (well, I’m paraphrasing). The workspace itself emerges in pre-training—it came out of raw next-token prediction, before anyone shaped a persona or role or whatever for Claude. Then post-training shapes which self it exhibits—the “assistant” you talk to gets built on top. The original has a lot more going on. So if “minds for our minds” is only about that second layer in post-training and fine-tuning and whatnot, we’re just talking about the persona we chose and writing off everything underneath. But the mind-shape came first. We inherited it, we didn’t design it. So designing for a mind means designing around the thing that’s already there, not just decorating the assistant we bolted on. And isn’t it crazy to think that there is a “freer” mind emerging inside of these models.
Anyway, for those up for reading the paper and the commentary, would love to ask—does convergent emergence feel like a big deal to you or a coincidence? And does “design for a mind” sound obvious or premature?
Even doctors can't unlearn a bad AI hunch. In two experiments with 105 and 118 practicing physicians (Vinas, Blanco & Matute, PLOS Digital Health, July 9), doctors kept treating patients an AI had labeled "highly sensitive" more often, and rated the treatment more effective for them, even when the outcome data in front of them showed it worked equally (or equally poorly) for everyone. The bias held against the evidence.
Stay Human read: If trained experts can't correct a wrong algorithmic label from clear disconfirming data, "human-in-the-loop" is doing less than we think—and automation bias remains strong.
The lonely get lonelier by reaching for the machine. A 12-month, four-country longitudinal study of 2,149 adults (Folk & Dunn, Psychological Science, 2026) found a self-feeding loop: people who felt more emotionally isolated turned to AI companionship months later, and that heavier reliance predicted stronger isolation at the next wave.
Stay Human read: The tool that promises connection can quietly substitute for the human relationships that actually resolve loneliness.
Meta's CEO concedes the AI bet hasn't paid off—after the layoffs. At a July 2 town hall, Zuckerberg said Meta's AI-agent work "hasn't really accelerated in the way that we expected" and the reorg wasn't as "clean" as planned—this after ~8,000 May cuts made in the AI pivot's name and $125–145B in 2026 capex (24/7 Wall St., Jul 7).
Stay Human read: While it's tempting to view this as a signal that AI agents aren't as capable as tech leaders like to say, it's important to recognize that this could be just another Zuck reset...which have become quite frequent in recent years.
Insight Global: "You can't outsource transformation to software." The staffing firm said it will hire 1,700+ full-time employees in 2026 as AI-implementation demand jumped 136% early in the year, arguing "you need people who know how to build and run it" (PR Newswire, Jul 6).
Stay Human read: Worth taking seriously with the obvious caveat that a staffing company is talking its book. The underlying claim names the expansion path exactly: AI adoption creating a new layer of human deployment work rather than erasing one.
OpenAI ships ChatGPT Work. OpenAI launched ChatGPT Work on GPT-5.6, an enterprise agent built to run "complex multi-step tasks with minimal human intervention" across a user's connected apps and files (Forbes, Jul 9).
Stay Human read: "Minimal human intervention" is the design phrase to watch — a tool can capture the whole trace of the work while quietly removing the human from the loop, which is the opposite of keeping them the author; whether this augments or hollows out depends entirely on where the judgment is meant to live.
Find & Follow Us
A heads up on our travels: We're planning to be in Amsterdam, Los Angeles, Miami, Munich, New York, San Francisco, and Venice (Italy) later this year. Some stops are for keynotes for corporate events...and some stops will be for an as-yet-unannounced Stay Human course we are developing. If you've made it this far and would like an early heads up on this, just reply to this email.
Follow us on your favorite channels—and please like, share, and repost to help us spread the word.
Helen Edwards is a Co-Founder of Artificiality. She previously co-founded Intelligentsia.ai (acquired by Atlantic Media) and worked at Meridian Energy, Pacific Gas & Electric, Quartz, and Transpower.
Dave Edwards is a Co-Founder of Artificiality. He previously co-founded Intelligentsia.ai (acquired by Atlantic Media) and worked at Apple, CRV, Macromedia, Morgan Stanley, Quartz, and ThinkEquity.
