Nye's Digital Lab is a weekly scribble on creativity in an age of rapid change.

This week, I'm reflecting on SXSW edu. I went to Austin to be on a panel. I came back with something harder to name.

Back from Austin.

Seth Piezas invited me. He runs AI for autonomous vehicles by day and chases creative projects by night. I've learned so much from him, as one of those people who can't stop building, regardless of what the job description says. He spent time at Pixar, orbited tech companies of all kinds, and has been thinking seriously about where technology and creativity collide. When Seth asks you to do something, you show up.

Simon Mowbray was there too. Chief Creative at Where The Buffalo Roam, a commercial house out of Los Angeles. Simon has been doing computer graphics longer than most of my students have been alive. Compositing, proceduralism, programming, and all of it, art. The guy's awesome.

The three of us did a panel on AI and creative purpose, and I thought it landed.

I got to say some things I'd been holding on to about adaptability, about what students actually need to thrive in acceleration, and about where education and creativity might go. Simon and Seth have a real love of the creative craft.

I feel honored to be on stage with brilliant creative humans, but the panel wasn't the story.

For the rest of the conference, I wandered.

I had a conversation with people from UPenn and MIT and heard talks from angles I hadn't considered. And the thing that hit me, almost embarrassingly obvious in retrospect: there is an entire discipline dedicated to applying the scientific method to how people learn.

I've quoted the book Make It Stick:The Science of Successful Learning for years. I've talked about interleaving, retrieval practice, deliberate repetition. But I've been doing it the way someone reads a nutrition label; I know the words without really understanding what I'm measuring.

My mind has been blown by the data the scientists showed me. Testing, reading scores, code check ins, emotional intelligence exercises. There are so many ways to seek out learning rates.

And now, the question that won't leave me:

are we actually measuring learning?

How do we know if our students are actually learning? They might think they are, or they might seem like they are engaged with the content we create. But are they developing critical thinking and actually encoding adaptive skills in those young brains of theirs?

Employers know what a student can actually do on day one versus what they can recite for a final. The gap between those two things is where most of higher education is currently living, and AI is about to make that gap impossible to ignore. As a result, I am making learning science one of my core skills to master. I am genuinely committed to be the best I can at it.

And...despite some absolute gems, I have to be honest...
Some of the talks disappointed me.

Not because the people weren't smart. They absolutely were. But too many of them were still arguing about whether AI is cheating, still treating it like an ethical dilemma rather than an infrastructure event. While the game developer conference was probably running sessions on how to automate half your game pipeline, SXSW EDU had panels debating whether candidate applications should be allowed to use these tools at all.

Many many organizations face crippled procurement problems, ethical design issues, and a deficiency of people who have the right mindset to teach AI literacy.

Oh, dear education industry, we are stalled. And the institutions that are stalled are going to get hit hardest as the wave impacts us.

I spoke with one woman who told me her school had 13,000 students when she started. Now it has four thousand. That's not a trend. That's a collapse in slow motion. She was still there, still teaching, still caring. But many around her weren't.

The institution around her was hollowing out. It became a stark reality to me that the landscape is in crisis. The enrollment cliff will be just the start, it's the value of the system itself that's at risk. I can't tell what that evolution will look like from here.

Luckily, the steak and beer is excellent in Austin.

There was one genuine source of hope, and I want to name it directly.

The Gates Foundation funded projects. The community nonprofits. The people who were figuring out how to get AI into the hands of high school kids in low-income neighborhoods, in rural areas without reliable internet. Not as a luxury. As infrastructure. As leveling.

These weren't the big keynotes. These were the small rooms where someone had built something scrappy and real and was watching it work. I was amazed to see a black high school student using machine learning to crunch complicated data sets to help civic organizations.

What it made me feel is... complicated.

Proud of communities. Worried about everyone else. The institutions that were supposed to be the towers. The schools are supposed to see further and protect the knowledge.

And the most alive work I saw was happening around and between them, not inside them. It seems that the institutions won't allow for it, but communities of humans can actually evolve with AI. The battle between institutional capitalism and the networked "something-next" is becoming visible.

And... One last thing I'm sitting with.

Quietly, without fanfare, a major exponential jump is coming. The big conference talks mention it as an aside as not to upset the massive infrastructure spends to the frontier business model.

But, make no mistake, it is coming.

One-shot learning.

The ability for AI systems to generalize from a single example rather than requiring massive training datasets could accelerate the development curve by an order of magnitude. We're talking weeks, not years.

Which means everything we're trying to lock in right now might already be obsolete by the time we implement it. The acceleration will obliterate fragile systems in ways we can't predict.

I know how that sounds. And my nervous system is processing on overdrive if it's true. The only rational response is to stop building for permanence and start building for adaptation. Don't optimize the curriculum for what AI can do today. Build students who can update themselves when AI does something none of us predicted.

The only constant is change.

I keep saying that. I'm starting to actually believe it at a cellular level.

Austin was good. The brisket was excellent. The city smells like live music and cedar and ambition. I met people I'll think about for a long time.

But... I came home different. The world is now ... different. We'll see what happens when I head back to the classroom.

Hey! That’s it for this time. I do this every week, if you vibe to the ideas I express, consider subscribing or sharing with friends. If you like tech-detoxing with a book like I do, I crammed some of last years best essays into a printed collection.

This essay was built from my notes from the conference, ingested into my personal knowledge base, and experimentally created by a messy collection of agents crawling over my Obsidian / Claude Code infrastructure. Final sweetening was done with Claude Sonnet 4.5. Images snapped with an Iphone.

For more info visit: https://nyewarburton.com

We’ll see you next time.

Seth Piezas: https://collab.sundance.org/people/Seth-Piezas-1732118036

Simon Mowbray: https://wtbr.tv/artists/simon-mowbray

SXSW edu: https://sxswedu.com/

The Math Narrative Project: https://www.mathnarrative.org/

One Shot Learning: https://www.geeksforgeeks.org/machine-learning/one-shot-learning-in-machine-learning-1/