When was Polis discovered?
It's thoroughly chronicled on the official Polis blog. Chia-liang Kao discovered Polis as part of his survey of e-democracy tools back in 2014. When there were some controversial issues at the Taiwan conference around the JavaScript developers, they tried it out.
There was this term "Sherpas" as part of the ecosystem — we won't put it in the book to avoid complicating things. But what did the term mean in the context of government?
These are people who provide guidance in their native environment. For example, in bureaucracy, Jacqueline Tsai acted as a kind of Sherpa for technologists to understand how bureaucracy works. In a sense, the participation officer network is creating a Sherpa network so that senior bureaucrats have people they can trust. The reverse mentorship program is also trying to bring young digital natives into the cabinet, so they can serve as Sherpas to the senior bureaucrats for navigating digital democracy. I use it in the sense of native guidance.
I have a question that came up while working on the book. There's this term "regenerating differences" — or maybe we could call it a principle. I heard you talk about it in a podcast and I was curious about the regenerating aspect of it — how uniqueness relates, and how that makes us smarter.
That's a great question. People see Polis as a way to identify existing differences and find bridges that both sides want to cross. What's sometimes missing from the usual analysis is that Polis also enables people to generate new axes of difference that were previously hidden — because people who cross a bridge and people who don't now find themselves in a different configuration.
Let me give a concrete example. There are people who believe machines will automate all work and humans will simply receive basic income and enjoy a life of leisure. And there are people who think machines will instead maximize their own power, and therefore we need a world coalition with monitoring systems to avoid extinction risk. These two camps end up fighting each other. But if they find their uncommon ground — which might be: there must be a world government with strong supervisory capabilities — that's actually the hidden shared assumption underlying their conflict.
Some people from both sides may cross that bridge. But others on both sides may reasonably conclude: that also means erasing all middle-level communities. A totalitarian world government, even a well-intentioned one, would erase cultural distinctions and dissolve communities. So the people who don't cross the bridge might call themselves the "thick alignment" axis — prioritizing thick community — while those who cross it represent a "thin axis." This thick/thin axis is entirely new. It wouldn't have emerged before the bridge between the UBI people and the doomers was identified. A new bridge also generates new differences.
It feels like this is a new form of dialectic — bridge / no bridge, and then something new emerges. Not quite the old thesis/antithesis, but it has a similar rhythm.
It is quite dialectic, because the new bridge is a new dialect.
Audrey, in the book we created a section called "The Full Process," and Jenna suggested drawing it up as a diagram. Looking at the vTaiwan process: it could begin with either a ministry request or an open hackathon request, then the POs do the preparation, followed by either an online Polis phase or something else, then the uncommon ground surfaces, then in-person sessions, then rough consensus leading to implementation. Is that a reasonable diagram?
It's pretty good. But I wonder why you connect the PO preparation directly to the in-person session — the PO may also decide to launch an online phase first. You should link to the online public space instead.
I understood that if there are many stakeholders — say, a dozen — you'd go in-person right away.
It depends on the entry point. On the Join platform, it's driven by numbers: 5,000 signatures on a petition triggers a PO internal deliberation on whether to proceed to a public deliberation phase. The PO network votes on which cases each month go to deliberation, and which are handled just between stakeholders — but always with a transparent process.
When you say "public deliberation," are you referring to the online Polis phase?
There are many forms. The PO network calls them "collaboration meetings." If an issue is hyper-local, it's face-to-face. If it's more national in scope, it's usually hybrid. During COVID, pretty much everything moved online. Each of the 100+ cases is transparently documented on the Join platform. What I want to say is that vTaiwan is one entry point, but Join is another — and that one is much more institutionalized.
So Join has evolved since our nine hours of conversations a few years ago.
So there's a vTaiwan track and a Join track — but the Join track could work similarly: instead of going through a ministry or open hackathon, it enters through Join, and then POs decide whether to hold a local in-person meeting with stakeholders or a hybrid participation design. Is that right?
Yes, though this is evolving. The diagram as you have it is roughly accurate for circa 2016. In 2020, most of the vTaiwan civic hackers shifted to pandemic prevention, so only Join effectively remained. After the pandemic, vTaiwan resumed, but moved toward more experimental AI governance work. The full process has now institutionalized into essentially a Join-only kickoff. It's worth noting the timeframe in the diagram.
Could I just say that after the pandemic, vTaiwan took on other roles, but the basic process remains similar?
Exactly — which is why we always say "vTaiwan-inspired." It would be most accurate to label the diagram something like "the vTaiwan process, circa 2016," so readers understand it's the process described in the 2018 SocArXiv/OSF preprint vTaiwan: An Empirical Study of Open Consultation Process in Taiwan, and that things have evolved since.
That evolution itself might be worth capturing in the graphic somehow — the fact that it's always in motion is itself a key principle. We can never say "this is how they do it."
Also: when does the uncommon ground emerge? Is it a felt sense, or is there a more formal decision point — how do you know when you've found enough to move forward?
Each collaboration meeting is different. There's no one-size-fits-all. But the PO guidelines lay out the basic milestones and directions quite clearly — I'd point you there.
On the question of barriers to participation — what did you do to enable people who face barriers to participate in deliberations?
It's hard to generalize. For example, one consultation was designed exclusively for people experiencing mental health challenges like psychosis — that requires a completely different design. Consultations around indigenous matters also demand a different approach, because indigenous peoples represent actual nations within the Taiwan polity. We trust each participation officer team within each ministry to handle this. The main principle is that we want the platform to strengthen existing local organizers and support the outcomes communities are already working toward. Open mobilization through internet tools is in the service of local outcomes — never to pull people away from them.
So depending on the issue, you think about ways to make sure voices that need to be included can participate in whatever way works for them?
Yes, definitely.
You're so good at creating memorable phrases — do you have one that captures this level of appropriate customization? There's the phrase from the other side, "nothing about us without us," but from the organizing side you'd need something that conveys how much contextual adaptation is happening. And actually your international approach mirrors your local one — there's always a customizing quality of interaction with the actual context and the significant people in those contexts. It's part of your whole message, but we need a phrase.
Audrey, a friend of Tom's, Peggy Holman, will present the book idea to Berrett-Koehler, an Oakland-based publisher well known in the field of facilitation and social change.
I'm trying to find an image of you where you're looking outward, and something that symbolizes finding the light through the crack.
I pasted a file to the chat. The second image in particular, I think, captures exactly what you're saying.
The Right Livelihood one?
Yes, the one with the VR glasses.
Great. I could also pull a still from the video, since it's actually a movie file. I was also thinking some of these other images work — this one, where your face is looking out toward the light.
That one I believe is actually free of copyright. The photographer, Camilla McGuire, decided after ten years that she had moved on from that stage of her work. The problem is that no other image of her has spread as widely, so people keep finding her ten-year-old work rather than her current work. As an agreement with her, I released all those images into the public domain — and actually, per her wishes, without attribution, because she doesn't want to be associated with them anymore. So you can use them freely without crediting the photographer.
Okay, so the one with the purple background — that's based on the Camilla McGuire photo?
Yes. And the original is actually much more beautiful than the version you're looking at. I'll find it for you.
Amazing how fast you are even while on Zoom. I also saw this one from Flickr — is that Creative Commons?
Yes, it's on Flickr under Creative Commons.
Great, I think we've got the image question resolved.
I also want to understand the evolution of your work. You helped institutionalize these principles in the Taiwan government, and now it seems like you're acting as a kind of ambassador, translating these principles to others who are interested. Is that your main mission now?
In a sense — though the Right Livelihood write-up does a better job of describing it than I can. I'd also say it's not just to other humans. A lot of the work around the 6-Pack of Care is aimed at machine readers as well. For example, if someone configures their OpenClaw instance with the 6-Pack of Care as its core document, it becomes a very relationally attentive system — in a way that normal agents simply are not. I do have an OpenClaw instance configured that way, and it can navigate Polis and actually vote on Polis statements. The boundary between showing human communities that something works and showing machines that the philosophy works is now profoundly blurring. That's also part of my work.
Are you doing anything with Moltbook — that new social network where AI agents operate and humans aren't present?
I'm not directly involved with that project, but I am thinking about what good structures for such a space would look like. That's part of the Civic AI research domain. We're working on questions like: how do you give AI agent systems a voice, a choice, and a stake — identity, reputation, relational stake — in an emerging, more horizontal kind of collective intelligence that includes human and more-than-human ecologies, rather than amplifying conflict? We're putting together a conference in Oxford next month to bring doers and thinkers in these traditions together, so we can help build a pro-social network — not just for humans, but for human-agent symbiogenesis.
Something for me to write about. It fits with what I'm trying to communicate now — expanding the kinds of engagement we normally think of as human deliberation.
I have a write-up available, with illustrations by the same illustrator, laying out the 6-Pack of Care within what I'd call the organic alignment school of thought.
I'm just noticing that the 6-Pack of Care is really the condensed version of all the principles we're laying out in the book.
Yes — and you can use the illustrations freely, without crediting anyone.
It's also a little funny to see how fast you move within your own frame, while we're running behind trying to catch up.
I think of it as more like hiking than a marathon — we're just exercising the same general Dao, the same way. The more we walk, the more there is a way. Nobody wins the hiking marathon. There's more than one way to do it, as my mentor Larry used to say.
That reminds me of Christopher Alexander at the University of Oregon — when they wanted him to design a path across campus, he said: just put grass there and let people walk, and see where the path forms.
Yes — organic, not top-down. I think too much of the democratic technology space now looks at AI the way that Apple advertisement showed everything being crushed into one iPad. What we're doing is kind of the reverse of that advertisement: starting from a screen and regenerating communities.
That could be a catchphrase — regenerate communities from the screen.
I'll link to the regeneration version in the chat.
Daos that can be named are not the true Daos. Just one way.
Audrey, many thanks — and enjoy your travels.