With the airbox?

I see.

So the infectiousness is.

Yeah.

Should we talk about mis and disinformation? Vicki? Robert, I feel like, again, this is something that I could talk about for 6 hours for sure, but I think we should get into the key parts of our conversation and then potentially loop back. So you showed us the talk ...

Yeah.

Yeah.

Interesting.

Yeah.

And so when you isolate that piece, you find the area of common ground and you say in the other piece, people are going to have to keep talking.

This has come up for us before.

Perceptions of marriage.

Interesting.

Right?

And I’ve been thinking about that in civil society, about how there might be a civil complexity that is important to democracy. And so when you think about the social sector here, does that resonate for you? And b, what are the set of actors that you think are really ...

May I give you a framework that we’ve been thinking about? And again, this will come up again in the dis- and misinformation. But we’ve been talking to all of these different civil society organizations and very much coming to this same, I think, place. And I’ve been thinking about ...

And there we were showing up. And that made a huge difference for our research. But with the ojibwe, no legitimacy, the history and prior work of the university with that community had so eroded the trust that my principal investigator of the project had to go for 4 hours ...

So I worked in Minnesota, in the United States, during the summer when the state government shut down. They decided to not fund the state government. And at the time I was working with the university of minnesota and going out to work with people who were from the ojibwe ...

I think this gives us actually a nice transition into the mis- and disinformation questions. But before we get there, I just want to give you a little bit of my own research experience that I think is relevant here.

Interesting.

Yeah. So it’s also an important source of power.

It sounds like it’s legitimacy from the way that they present, credibility from the work and the track record, and then also funding from having a successful social entrepreneurship stream.

Interesting.

Yeah.

Social entrepreneurship, right.

Thank you for that very beautiful rephrasing of my question. I really appreciate that.

That’s correct.

Exactly.

Yeah.

Because once they’re in the foundations, there are structural factors that keep them multiplying throughout our technologies.

And so, if you’re relying on helping the helpers, you’re relying on underfunded nonprofits to bring in this sort of general conversation, what are you doing to mitigate the sort of structural inequalities that we know perpetuate in every system?

If you have returning citizens, there is a subset of the population that has an interest in not having those folks have a voice in these models. And there are many people who have a real strong interest in not having these systems work for everyone.

I want to push back, of course, in a lot of this work, this idea that people’s needs from these models that they’re composable. Right. Or that you can bridge between two communities and there will be some sort of shared reality, how do you meaningfully engage with power in ...

Yeah.

We know Glen quite well.

But as an urban planner, I know that if you want to reach hard to reach populations, if you want to reach returning citizens, you can’t just put a policy on a website. And so, as I understand v taiwan, right, you have humans who go to taxi drivers or ...

I know, and I want to pull on that thread next because it’s fascinating that it’s the top score, but can I ask just maybe two questions? I’m really interested in the social technology or social infrastructure that goes with these technologies, because when I think about vTaiwan, you know, ...

Go for it.

Go ahead, Robert.

Right.

Yeah, that’s not what we’re getting.

Interesting.

No pressure. I’m just curious. It’s really interesting to think about these two features also as a metaphor, right, in policy, right, of low barrier to entry and composability, and sort of what that opens up in terms of who can participate and how.

Is that a known set of preconditions? Like, is there a theory or research on those two conditions being key to open source, or is that an insight that you’ve had?

Yeah, it got real deep real fast.

Purely on the fine tuning layer?

Like Anthropic.

Ah, thank you.

LoRA?

It’s an insight in this context.