Creating tools for impactful discourse

Hello! I haven’t been here in a while, but when I saw this new discussion/collaboration tool, I thought of EdgeRyders. If you use it, I’d love to see what you do with it. is the paper describing A System for Interleaving Discussion and Summarization in Online Collaboration

Here is access to the tool if you want to play with it:

I posted it here because it looked like a compatible convo (“…how we grow as a team”) – but if there is a better thread, please feel free to move it.


Thanks Howard, I like the interface but I think I need more time for playing before fully grasping what it can do!
I’m wondering if there’s a minimum body of texts that are worth summarising like this… Unless we import different conversations in the same to-be summary…

1 Like

Thanks Howard for the pointer to Wikum, I’ll have a deeper look in a bit.

If you liked that tool, you might be interested in Polis as well. It also is about online dialog and also uses AI, but instead of building summaries they focus on building consensus. According to a recent Guardian article, it already has a role in the political process in Taiwan.

1 Like

@matthias, @howard_rheingold… had a lot of good publicity because of the Taiwan case. Years ago, I was in a demo with someone from the team. Observations:

  1. Unless they updated the tool substantially, it does not use “AI”. It uses standard clustering algos, based on geometric distance (Cartesian? Cosine?). It first shows where people are in an n-dimensional space, where each statement (tweet) is a dimension, then runs the clustering algo to group people. At this point (I imagine, it turns things around, assigning statements to the clusters they are closest (now in m-dimensional space, where each cluster is a dimension).

  2. It tends to pick up uninteresting statements, because it is based on consent. In the demo I was in, the most highlighted statement was “A referendum should not be used by politicians for short-term gain”. Agreed, but then what? There was zero learning value. In this online demo you see that things like “Football is too big to fail” are picked up.

  3. These statements can not be made more interesting by the process, because there is no way to reply to something, adding a distinguo, making a case for or against it etc. You can only approve or not approve.

  4. The voting process that picks up the “winners” is at risk of being mathematically unfair. This was the great argument made by @pietro on all kind of online “voting” (likes, retweets, etc.). I am not certain of the technicalities here, a lot depends on who gets to see which statements to agree/disagree with, and on the sequencing between voting and looking at results. In my demo that was real time: people could see “winners” emerging, and so they would agree with them even more. But maybe they do not do it like that in real use cases.

IMHO, Taiwan is a success story… but its driving force is Audrey Tang. It’s not every country that recruits a digital agenda minister from a revolution, and that minister turns out to be a prodigy with a IQ of 180, who was doing high math before she turned 6, and who also had to experience some difficulties in life (she is a trans woman). Scale that if you can!

With that said, if you are looking at a heated discussion, people are shouting at each other and you need to rhetorically highlight the similarities, then can be a great choice. I am personally more fascinated by more argumentative, interactive ways to collective intelligence… as is clear from my own work.


So the details of the Polis software are rather unimpressive :worried: But I remember that Richard Stalman once said something along the lines of “it’s a great piece of software if it completely changes how people do certain things”. So, if it has severe impact on the real world. By that metric, the Polis software is doing well … and if only because it was, perhaps by chance, the tool of choice for the government of Taiwan at a time when they were open to real e-democracy.

The question becomes, how can we propose or design a path for our set of consultation and sensemaking tools to do the same? And then try to follow that path? Because so far, to my knowledge, nobody ever listened to what our community said and which we presented in a refined form to so-called authorities.

Or the other way around: If software never has any real-world impact, it’s not worth doing it (certainly to me). But I can imagine situations where there would be real-world impact. For example when using Discourse and our sensemaking tools (somewhat modified) in participation and decision making processes in progressive cities. Where it would be clear what is the path from participation to action, and that there will be action. That would finally be interesting and rewarding to do (at least to me)!

Any ideas or contacts how we could get closer to this? (Maybe also @nadia and @amelia? I’ll split this into a separate topic if this turns out to be a fruitful discussion to have.)

1 Like

My brain is kind of cooked at the moment but would be happy to think about this in a month or so if someone reminds me.

1 Like

That’s alright, as this is a longer-term question anyway. I’ll remind you here :slight_smile:

1 Like

This is exactly what I’m hoping our next big project will be building towards — both my mentor/friend at UCL whose an anthropologist and our ER crew’s big care at the core of it is how to get this research to have real-world impact (something frustratingly difficult to do).

That’s the capacity of Graph Ryder that I really see, from the perspective of my research — how to get the knowledge and expertise that people doing their thing in their everyday lives and doing things creatively in their own contexts transmitted to these “so-called authorities”. And maybe equally importantly, their roadblocks and problems – how to get them the resources they need to keep doing the things they’re doing, or remove the obstacles in their way.

That’s why these grand, jargony presentations drive me crazy – I feel like the role of government isn’t to find some buzzy, pretty solution at a macro level, it’s to do the more tedious but in many ways simpler work of facilitating the already excellent work of the people deeply embedded in their own daily practices. This is the general theme of my part of the Regions Week talk, anyway.