Building with Intelligent Beings: Three Takeaways from the Artificiality Summit

Five years ago, I couldn’t have imagined it. Three years ago, it was a far-off dream. Now it’s Monday morning and I’m collaborating with machines that understand context, generate ideas, and help me ship faster than I ever could alone.

I return hopeful, a little nervous, and deeply encouraged to know there are people like me in the world. Lots of them. People who build, who question, who admit they don’t have all the answers but keep pushing forward anyway. People who see the promise and the peril in equal measure, and choose to engage with both.

The sophomore year test

Last year was exploratory for the Summit. Everyone wondered: would this be interesting, or just a bunch of nerds talking about RAG models and LLMs?

It was way more than that.

But this year counted more. This was the real test. Would speakers rehash what they’d already said? Would the novelty wear off? Would the magic fade now that we knew what to expect?

Not even close.

Talks were not repeated. Topics were revisited, but with crucial nuance. The thinking had evolved. We covered so many ideas I’m still processing them all.

And through it all, three insights emerged that will change how I work.

Communication with intelligent beings is about having the proper interface

Michael Levin spoke about communicating with different intelligences. When you discover the interface, you can make profound changes.

Watch this:

His work focuses on biological systems—cells, tissues, organisms—and how to communicate with them at their level. The principle translates directly to AI.

Finding ways to interact with AI is the same challenge. Understanding how to get what you want out of the interaction is about understanding how information flows and how computations determine responses.

I spend so much time tweaking prompts, adjusting parameters, trying different models. I’m hunting for the interface that lets me communicate clearly with these systems. Sometimes that means going deeper into how the model processes context. Sometimes it means stepping back and rethinking what I’m even asking for.

What changes for me:

Stop treating AI like a black box. Study how the system processes information. Learn its language. Build the interface that lets you communicate clearly.

The tools change. The need for clear communication doesn’t.

Uncertainty can be a gift

Maggie Jackson gave a fascinating talk on her work with uncertainty.

Her research shows something counterintuitive: uncertainty isn’t a gap to fill or a problem to eliminate.

It’s a gift.

When you sit with uncertainty—when you resist the urge to reach for the first comfortable answer—your brain becomes receptive to new data. Working memory elevates. Attention heightens.

Key ideas that landed:

• Makes your brain receptive to new data
• Elevates working memory
• Heightens attention

We treat uncertainty like a problem to eliminate. Rush to answers. Fill the gaps with assumptions.

Jackson argues the opposite: uncertainty, handled well, sharpens you.

What changes for me:

Be okay with not knowing the answer. Use that uncertainty as a focused moment to find it. Don’t reach for the first solution that presents itself.

When I’m building and hit a wall—when I don’t know if the architecture will scale, or if users will actually need this feature, or if I’m even asking the right question—that’s not a failure state.

It’s an invitation to pay closer attention. It’s when you’re most receptive to insight. The uncertainty forces you to look at the problem from angles you’d normally skip past in the rush to ship.

Tribes matter

In a world where we compete constantly—overtly or just through signaling—being part of a group that’s accepting and open is an amazing feeling.

I participated in a discussion about AI safety. I disagreed with much of what was said.

And I said so.

Out loud. In a room full of people who had spent their careers at the top of their fields in technology and science.

No one shouted me down. No one belittled me. No one dismissed my opinion as naive or uninformed.

They listened. They responded. They disagreed back, with specifics and reasoning.

We don’t all agree. Far from it.

But we recognize that we’re willingly in the same room, wanting to explore ideas honestly.

What changes for me:

I came away feeling equipped for the big questions coming. Not because anyone handed me answers, but because I saw how others wrestle with the same tensions. Supported by a community of smart people who also don’t know all the answers but have strong opinions and the humility to test them.

Building in isolation is possible. Building with a tribe is better.

The work doesn’t happen alone. I’m part of a community of builders who are smart, opinionated, humble enough to admit when they don’t know, and brave enough to voice disagreement when they do.

So what?

You should join us in Bend next year. Subscribe to the newsletter to find out when.

You probably won’t learn anything immediately practical. No one will hand you a framework you can deploy Monday morning. There won’t be a five-step process or a toolkit to download.

But you will walk away with a reframing of your big questions. The questions you thought were about tools will reveal themselves to be about communication. The uncertainties you’ve been avoiding will look more like opportunities. You’ll be more informed, more critical, and more encouraged as you try to navigate the new world we live in.

You’ll learn to sit with uncertainty instead of running from it.

You’ll find people who challenge your thinking while supporting your work.

And you’ll discover better interfaces for communicating with the intelligent beings we’re building alongside.

We’re building with intelligent beings now. That changes everything.

It also changes nothing about what matters: clear communication, honest uncertainty, and people who challenge you while supporting the work.