Network Age Reflections: Ep. 21: “Let’s Get Personal...AI (feat. Owen Barnes)”
Some musings on a recent episode
Working on the Network Age podcast has been a real delight. In the show’s short life, I’ve spoken with a multitude of brilliant guests on the cutting-edge of Web3, crypto, and technology at large. As our audience has grown, I’ve also had the joy of connecting with listeners who find themselves as stimulated and challenged by these conversations as I have been.
With that in mind, I’m hoping to write up a short reflection on each episode a few days after it’s released. These reflections will serve two purpose:
To express the lingering thoughts, questions, musings that tickle my mind after each recording
To give listeners a chance to connect with us / each other—my hope is to build a real community around The Network Age of curious folk interested in a positive vision of the future
While we already have an Urbit group (~mister-hoster-dozzod-hocwet/network-age-antechamber), these posts are an invitation to support, argue, agree, jeer, mock, reflect, and finalize business deals in the comments. Don’t hesitate to say hello! (And, if you have other ideas for organizing a Network Age community—please, reach out).
Reflections
Owen Barnes was a wonderful guest on our last pod. Intelligent, articulate, and with a charming accent to boot! Here are some things that have been on my mind since we spoke:
The Little Man in Your Phone
Personal AI is going to be transformative. It’s easy to conceptualize AI optimizing across large networks (the electricity grid, international shipping, etc.), but increasingly I feel the average person will be most impacted by the AI that lives on their phone. It will curate our search results, organize our calendars, create custom notifications, silence calls from our mother-in-laws, return pants that are now the wrong size because you’ve put on five pounds (which you didn’t even realize, but the AI can see how many bags of chips you’ve been ordering), and show us everything we wanted before we even knew we wanted it.
As with so many new technologies, it is difficult to know where this falls between scary and exciting. Increasingly, what I want from technology is ways to optimize my physical life without spending more time in a digital world. I want my life to be frictionless—not in the sense of no problems, but in the sense of less bullshit. There is a value to some suffering, but there is no value to suffering from logistics. The optimistic version of personal AI means more time for yourself, your loved ones, and your passions. Dare to dream.
Data
Data! What a problem our poor personal data has become. Increasingly, privacy concerns have moved from fringe internet whackos to the mainstream discourse. Personal AI makes this discussion even more pressing. It’s great to have an assistant that knows everything about you, but not if that assistant is run by a company with financial incentives to exploit you. In this way, Urbit really is perfect for personal AI—a unified space where an AI can access / process the huge amount of personal data necessary to be useful without sending it out to a third party. It sounds silly, but Urbit honestly might be all that stands between us and dystopia.
Is Perfect Perfect?
I keep thinking about what Owen said about noisy restaurants. I hate noisy restaurants! I would like to avoid them as much as possible. The prospect of an AI that can, even without me asking, learn my preferences and send me to restaurants where I enjoy not only the food, but the ambiance is very exciting. Spotify playlists, but for restaurants.
Of course, this raises another question: do I actually want a perfect experience? If I never have to eat at another mediocre restaurant again, will that make me happy? My Spotify playlist is nice, but it never asks me to consider anything really new. There’s something to be said, of course, about the unexpected, the challenging, the strange—an experience whose value may never make it into my aggregated data. Grit and pearls, etc.
Who is Art For?
If art is for the consumer, perhaps we lose nothing from AI. If you can’t tell the difference between an AI painting and a painting made by humans, who cares? If you watch an AI generated movie that looks and feels exactly like a Marvel blockbuster, who is losing?
The artists, of course. Increasingly, I believe that art is mostly for artists themselves. I am a passionate reader, but I am open to the possibility that an AI could eventually produce literature that moves me as much as, say, my boy Knausgaard. But then what would Knausgaard do? (he’d probably be happier not writing). Certainly there are many who have found meaning in making art privately, but creating really powerful art is an incredibly difficult task. As much as we strive not to be egotistical or materialistic, there is a reason I continue to submit my little stories to little journals that no one reads (including the contributors). Because it makes me feel good. Because sometimes the only way I can make myself sit down to work on my novel is remembering that last week I got a $50 check from the Middle of Nowhere Review.
Perhaps we will see small communities organize around the appreciation for human-created art. Of course, with time, those too will disappear. Luddites, the kids will say, shaking their heads as they enter their pleasure pods.
Misc.
Can’t stop thinking about where signs are placed. Read a very long note posted at the cash register of my local grocery store. Caused a minor traffic jam. Not my fault! Bad sign placement!
AI coding. ~timluc-miptev thinks we’re far away. Owen Barnes thinks this is the future. Me? I spent an hour trying to make Codex write a snow simulation where the particles followed my cursor. No, we’re not there yet. But whenever we get there, it’s going to be very cool.
Okay everyone, that’s it for now. Thanks for listening, reading, and being a part of the community. If you have any thoughts on the episode (or anything above), please comment and don’t forget to subscribe for updates.



I think that's a good point—if AI speaks Hoon, it will be able to "just do more," though we're obviously a ways from that happening. I think, though, that if AI reaches the point where it can produce app level code w/o errors, it wouldn't necessarily be long before it can also easily solve deployment, integrations, etc.
AI seems to be at an interesting point where it is currently useful to learn how to manipulate it effectively, though it will eventually reach a point where it makes the user almost irrelevant. In this case, Urbit likely also wins because anything annoying about designing on Urbit will be abstracted away
I think there was a missed opportunity to describe how Urbit wins when AI can write code really well. Since an Urbit can “do everything”, the apps AI creates with it will be better. It won’t just make a JavaScript frontend and say “good luck deploying this or adding integrations.”