Scaling Loomio, Functional Decision size?

I think the polling helps for mass numbers, the coordinating group deciding the polls and what decisions to make, i found with Occupy, it was only good for preconsensus groups already working in the same mindset. If a new idea or person joined in, it spiraled quickly into the defensiveness and an emotional unintelligent typical AOLforum. I don't think you can fix that with an algorithm! For our same page things, it very excellent to hash out formal proposals from sketch ideas.

You can't fix anything with algorithms, but they can definitely help to mitigate problems. Of course, they can also cause more problems than they solve ;)

This is a pretty interesting question. I am under the impression from readings from many years ago that humans just aren't good at discussions at that scale - people start to become anonymous, because no-one can keep track of more than a few hundred people. Anonymity allows youtube-comment-itis: everyone can be an arsehole and get away with it. I think for organising at this scale, you're better off looking into some kind of confederalism - split the group up into smaller groups than can reasonably manage their own decisions. You can split the group along arbitrary lines, and can have cross-over (e.g. region-based and interest-based affinity groups).

I am under the impression from readings from many years ago that humans just aren’t good at discussions at that scale

Humans aren't good at a lot of things, so we build technology to allow us to do those things better. It was not that long ago that humans were pretty terrible at communicating with other humans on the other side of the planet in realtime.

You can’t fix anything with algorithms

Completely agree, and I think this is one of those problems that you don't fix with algorithms. You fix it by improving the user interface.

The tools we have now (loomio included) aren't well suited for large scale discussions (in many cases they aren't even well suited for small scale discussions, but more for sharing random information). They don't address the problems you call out (youtube-comment-itis, conversations are hard to follow, etc), but I don't think these are unsolvable problems. We just haven't actually put much effort into solving them.

I think one of the major improvements will be changing how we think about discussion. A chronological sequence of posts just doesn't work at that scale. Discussions need to be able to flow organically without interfering with other threads of discussion. Instead of a linear sequence, a large scale discussion is more of a directed graph.

Tools like http://www.reddit.com/, and http://www.discourse.org/ have started to realize this and try to expose some of that structure, but I'm my opinion they don't go nearly far enough. You're still distracted by a lot of unrelated posts in different threads of discussion because they try and show you everything.

I think http://assembl.org/ is another step closer, but still under active development.

There was a bunch of related discussion in the Political use for Loomio thread.

The basic question of how big does Loomio scale is a great question, which the core team could probably give some opinions on here. I'll just opine on some non-technical matters.

Decisions can scale better than discussions, because some decisions require little if any discussion for the vast majority of potential participants. That's mostly a question of whether or not people are framing proposals that are truly within the shared interests of a large group. The shared interests of a large group are fewer than those of a subgroup.

If people can only be convinced to support something by way of intensive discussion, well then I think that doesn't scale well at all, nor IMO should it. I'm a bit radical here but I doubt I'm alone in that.

I have a pretty strong opinion on liquid democracy: I don't think people should delegate their votes to others on any long term basis. In fact I prefer for it to be on a decision-by-decision basis. Passive ongoing delegation of votes can lead to a lot of problems which, IMO, may end up looking an awful lot like the problems we already have.

@quentingrimaud might have some insights here.

Personally the two models of scaling that I'm interested in are delegative & deliberative.

See Wikipedia on delegative democracy and Aaron Swartz on the deliberative model known as parpolity

DemocracyOS and Liquid Feedback are exploring the delegative model, where votes can be passed between people to form blocs of influence. I can imagine that being pretty awesome, and pretty problematic too.

Loomio is currently far down the deliberative end of the spectrum. We've stayed away from the "hard" problems that come with scale (e.g. identity verification) and are working on the difficult "soft" problems like teaching people to engage with each other respectfully in pursuit of shared understanding.

In my opinion any model is always going to end up with a relatively small group of people deliberating together in pursuit of consensus, so that's the point I've put my energy into. It seems like a pretty sure bet.

When we're talking about nation-scale decision-making, the problem with either electoral or delegative systems is that the point of citizen engagement is entirely discontinuous with the actual crux of governance: negotiation, compromise and consensus-building. Voting leaves a bitter taste in my mouth, it feels like I'm giving up my autonomy in the hopes that some jerk who I don't even know is going to exercise it in my interests.

Conversely, the deliberative model is fractal, self-similar at all scales, so the grassroots participants have an insight into what operations look like at the uppermost or innermost decision-making body. Participation is an enriching experience for the individuals, and it draws out collective intelligence greater than the sum of its parts.

There are something like 25,000 Podemos members using Loomio right now, in 1,000 different groups. It's pretty easy for me to imagine Loomio 2.0, where all those groups are associated together into one network. Imagine sending a proposal out to all the different groups in the network and seeing distinct deliberations underway in each local group, watching points of agreement or controversy or insight or initiative spreading virally throughout the network, everyone participating in their full autonomy and simultaneously contributing to a massive collective roar, or a song, a unity of unmerged voices.

Thanks all for the quality feedback/input. I'll do some more research and post back at some point. I am developing a civic engagement platform and am looking at Loomio as part of the integrated set of tools.

Re: threading: https://trello.com/c/SAMuMFoc/24-reply-to-comment-one-level-of-threading

@richarddbartlett can you please turn that gorgeous comment into a blog post?

What about a version control system like github for collective documentation for functions such as public policy coauthoring, or group business plan writing for large groups? Not everyone would need to co-author, but hopefully many would vote on which versions they favor. This voting would steer the group toward the majority's version until some limit closest to consensus is met, perhaps. Maybe a version control wiki with a WYSIWYG editor.where the voting steers the direction of the documents evolution. You all know of functionality like that? I have seen some examples of people directly using github for co-authoring, but not with the voting intertwined. And for very large groups, I think liquid/delagative democracy makes a lot of sense. Yes, there is risk of people becoming complacent, but that is always a risk. At least once they start to see things going in a direction they don't agree with they can take their vote back, immediately.