Organizing a growing Social.Coop: an idea for a proposal

@Stephanie Jo Kent I haven't seen anyone vilify ChatGPT here on Loomio.

However, the inaccuracies of LLM summarizations make them no better than human summarizations. Therefore, as I stated in a previous message,

"I find no value in ChatGPT or any LLM summarizations of our conversations."

@Stephanie Jo Kent ChatGPT may be useful for some purposes (e.g. drafting emails or rewriting one's own resume), but I still don't want it anywhere near my democracy. Using it to "summarize" content isn't really a summarization, but a statistical model of what a human-written summary might look like. It's therefore extremely important that the author of the text being summarized take charge of making sure that the summary accurately conveyed what they were trying to say.

The current ecosystem of LLMs do not parse sentences in order to extract meaning (although there is other, less flashy, natural-language software that is designed to do that). Because of the way the models work, they can seem to produce high quality summaries of some kinds of text, especially if the text is structurally similar to texts in their training data, but there is no guarantee that they will capture all views (it doesn't know what a view is), nor is there any guarantee that they won't introduce extraneous content. This makes them extremely poorly suited to extract information from a long involved conversation in which people are trying to hash out our desires for the future as part of a decisionmaking process.

Posting ChatGPT "summaries" invites readers who don't have time to view the entire thread to assume that the relevant information is represented, and I think it would be extremely unfortunate for our community if our members started adopting stances on proposals based on LLM-generated summaries.

If we can't figure out a real democratic process for representing the diverse views in our community, then maybe the right answer is that the community is just too big. I don't think we're to that point, but if people start seriously talking about LLMs as a solution for extracting meaning from our conversations, then I'd be inclined to say that we've reached that point.

At risk of stating the obvious, the ChatGPT summary that was posted by @Jonathan Bean clearly illustrates the dangers of inaccurate information being introduced by LLMs. This entire paragraph:

We also talked about your career interests and skills, and how you might be able to leverage these to support Social.Coop and the transition movement more broadly. We discussed the potential value of AI technology in supporting your work, including the ability to conduct research and analysis, create content and messaging, and engage with members of the Social.Coop community and other stakeholders.

does not reflect any part of our conversation thread as of the time the ChatGPT text was posted, and in fact is clearly geared toward an entirely different style of organization from Social.Coop. Until Jonathan Bean posted the the ChatGPT "summary", there was no discussion of AI at all in the current thread. And we don't engage with stakeholders beyond our own membership. ChatGPT couldn't capture this because it doesn't parse text or extract meaning. There is no comprehension of what has been said. The summary it gave is the best approximation it can come up with for summaries of text that is semantically similar to our conversation.

Democratic conversations by cooperative organizations clearly do not form a large portion of its training data, and if even they did, I'd be concerned that the "summary" it came up with would be a better reflection of some other democratic conversation rather than the ones we are having here.

@Dynamic It is important recognize the limitations of the GPT.

I should point out that the mistake is mine for not providing the proper context for that snippet of our conversation about this thread. In the google doc I do talk to it about my career interests and use of AI as personalized sort of suggester for what and how to vote and just secretarial kind of work, Imagine that everyone of us had an AI programed by us to create reports personalized for use on how to best represent us but of course with our final approval on votes.

So the summary is a summary of my conversation with the GPT AI about this thread and the broader social.coop Loomio and organization. At this point it is in the spirit of experimentation just something to be aware of as the trends around us grow.

I apologize as I have been quite busy and intended to participate more but some things have developed in my life taking a lot of spare time. Maybe there is some tech out there that would help manage the complexity a little better than we are with the current tools. I wish there was a plug in that could analyze the Loomio and allow me to participate with less time investment.

@Stephanie Jo Kent The tech is developing exponentially faster and faster and some say the newest version of GPT4 is pretty much at the level of a minor AGI Algorithmic General Intelligence capable of almost any human intelectual task. I am cautious about this transition but am optimistic that it would be very hard to make an AGI that was smart enough to be used to dominate us but dumb enough not be a benevolent intelligence as it seems in time greater intelligence tends toward greater benevolence as benevolence is the most intelligent strategy in the long game. But there are plenty reasons to still be worried about it.

But right now we can help shape the technology before less benevolent forces shape it.

I think I could probably ask it to make an AI powered Distributed Cooperative Organization platform at this point as I've seen many people simply engineering the prompt with natural language and getting fully functional code for entire apps to implement with very little modification.

I will write up some requirements and try to make it soon, hopefully we can solve this scalability problem with federated DisCOs assisted by algorithmic intelligences. So yes scary, but less benevolent forces are getting an edge and we might be left obsolete if we don't adapt. It is not the strongest that survives but that which is most adaptable to co-evolutions.

@Jonathan Bean

I agree that time is an issue, but I'd really prefer if we focused on social solutions rather than technical ones. I really feel like Loomio isn't currently working well for what we are trying to do.

Can someone help me understand why I received an email notification that I was invited to vote in this sense check but now have a Loomio message saying that I haven't been invited?

@Dynamic same experience!

@Nathan Schneider @Dynamic Ah, I think the problem is that this sense check was for people who volunteered to help with the rewrite. Once they have weighed in, we will make a proposal to the whole group.

@Matt Noyes

What is annoying is getting an email telling me I've been invited to do something, and then logging on getting ready to think through a proposal idea and finding out that I'm not invited after all.

@Dynamic Sorry, my loomio blunder - I tried to send the notification to the folks who had volunteered to work on this but it seems I sent it to everyone who was on the thread.

@Matt Noyes

Thanks!

Thanks so much for the team for working on this! I think this is really solid, and it seems like a good framework for producing more accountability and accessibility in S.c governance, building on the logic of affinity groups and a spokescouncil. A couple of thoughts for some refinement:

• I wonder if we can avoid WG elections altogether. This has not been something our WG have actively been doing, and some WGs were not elected at all. I generally think our membership is high-trust enough that we could invite WGs to organize in a variety of ways, leaning toward voluntary participation, with one point-person to the OC. I think we want to encourage more permissionless participation in WGs rather than set an expectation that it is an office one has to run for. (If WGs want to hold elections, they still can.) See context for my learning on this here.

• If you go for the above, then the recall procedure could simply be applied to all OC members, not just at-large. (Let's aim to simplify wherever possible!)

• In the interest of simplicity, I would simplify the At Large selection process: Just run a sortition process among the membership, with an option to opt out. Again, I think we need to be very sensitive to the time-cost of processes like this, and we are a high-trust network. I think we should try just a simple sortition process.

• For the roles, it seems to me that Membership and Treasurer are actually WG roles (Community and Finance, respectively). So why not just have those go to the representatives for those working groups automatically?

Again, my major concern through all this is to avoid undue complexity. We are a largely volunteer community, and we need to be careful about how much we ask of each other in terms of time for governance. (The more we have to govern, the less we get to toot!) But I do think this kind of coordination body will be an important step forward for all of us. I am so grateful for your leadership in moving this forward.

@Nathan Schneider

I think I agree with all this. I also think that before putting the proposal forward to the general community there should be some language added about why the OC membership would be capped at 9. The document seemed surprisingly specific about that, and it's not at all clear why.

@Dynamic I also felt this, but assumed - perhaps improperly - that this could be amended by the OC if/when it determined that another Work Group might be formed.

@Nathan Schneider "Just run a sortition process among the membership, with an option to opt out." My concern with this is that if the general body is not globally representative with respect to DEI sortition will tend to reinforce and naturalize that condition. (Roughly analogous to the social problems around Large Language Models scraping and normalizing settler-colonial legacy texts.) I encourage a structure that intentionally seeks to reduce the time between 'who we are now' and 'who we want to become'.

@Rich Jensen Sorry for the late reply -- I agree with Nathan's suggest to let WGs figure out how to recruit new members and to have recall apply to all OC members. For reasons Rich mentioned, I do think we should use juried sortition for at large members.

@Dynamic Good point about nine members on OC... the idea is to ensure that the OC is small enough to have high participation. I think the sociocracy target # for circle is 7.

I seem to have two accounts listed. My 'undecided' vote should be struck from the record. Sorry I don't have time to make this adjustment from my side before the impending vote.

Was right at the end of voting & writing a comment when it all disappeared. Guess i missed the time slot. Ive had a lot on and Im very tired. Quickly read thru the proposal. It appears to be longer - more complexity and work involved running all this. But also still quite similar to the original?

Im not sure theres really any changes that address the issues I raised. I wont elaborate at this time as it feels like I could be wasting energy and right now I have little.

Ive found it hard (both technically and finding the time and motivation) to engage with this process of review. Guess i could of done more, I tried a few times and found the document was missing, when it did reappear I found the ability to comment on the docs in Nextcloud is blocked to phones (i dont currently have a PC / laptop).

I cant read Spanish but as it was a sense check my vote is a down thumb.

@Darren Sorry for the late reply. The proposal is quite similar but we did try to incorporate people's comments and suggestions. In the updated version, linked above, you can see the previous version with people's loomio comments added in the margin. Sorry it's been a clumsy process. Definitely understand the need to protect one's time and energy.

On behalf of the Finance Working Group, based on our meeting today, I want to express our strong support for moving this process forward. We find that Social.coop's treasury has been continually increasing, due to our generous members, but we are not spending it and using it to provide those members with the service they are paying for. We believe it is necessary to have a more coordinated budgeting strategy, and that can only be done with more integrally coordinated governance.

@Matt Noyes @Matthew Cropp is there anything we can do to help you move this forward?

@Nathan Schneider Thanks for the nudge, Nathan! I will start the proposal today.

I'm a little confused to see references to voting in the doc? Does social.coop use consent decision making? Or is it majority vote?

@Sam Rossiter You can review the bylaws here: https://wiki.social.coop/wiki/Bylaws

We have a modified consensus process that combines majority vote with an allowance of blocks, which trigger a supermajority requirement. But different structures within the co-op could use different decision-making systems.

To slightly expand on my comment with my vote against this proposal. It feels like this will lead to further centralisation of control and activity to those with the ability and willingness to attend (sometimes closed) video meetings. Video meetings dont work for everyone (either their tech stack or they just dont feel comfortable operating in such a space)

The Community Working Group Ops Team (which operates via closed video calls) was supposed to deliver what was democratically decided in the CWG Loomio. With this proposal its now taking on the decision making power of the (abandoned - but it didnt have to be that way) CWG. I worry this is indicative of how democratic control in the coop is heading.

As i mentioned, it feels like the lack of participation reinforces the gradual momentum to closed centralised control. But also we often see low participation in more closed processes. Just most members arent aware as their operations are opaque.

I wonder if all these new proposed roles will be filled and how effectively people holding positions will operate (how good will comms in and out of this new body be, what levels of attendance and engagement there will be at meetings)

There wasnt a block option offered for this proposal. I would have been tempted to use it. Although maybe not as Im also not sure theres people willing to put in the efforts needed to organise and effective alternative.

@Darren Thanks for clarifying these concerns.

I do think, however, we need a bit more centralization in Social.coop. A number of initiatives that have needed to happen have simply gotten lost because of inadequate coordination among working groups and participants. (The participatory budgeting effort we were working on is an example!) As a working group member, I have often felt isolated and stuck, and unable to move forward without coordination that is hard to achieve without a shared space. I think we can address your concerns by ensuring the proposal includes:

• Opportunities for voice and feedback async, not just on calls (necessary for timezone justice as well)

• Ensuring all meetings are open to membership unless they focus entirely on confidential, personal, or legal matters that require limited access

• A clearer ladder of participation, so all members have more opportunities to get involved—something sorely lacking in the current structure

I hope we can operate out of a shared recognition that there are some real breakdowns in the current model, and focus our contributions on improving the efforts of those who are trying to make S.c governance work better for us all.

@Nathan SchneiderI agree theres problems but same as with the opening post to this thread I dont agree with the diagnosis presented. I think we have an adequate structure just need more member participation and better process/onboarding of new members to encourage and help facilitate that. Doing so would be far less work than organising and operating the structures described in this proposal and your elaborations and providing good comms into and out of them.

From my perspective the participatory budgeting initiative I tried to lead languished not from a lack of structure but a complete absence of anyone but myself prepared to participate in its development, or even discuss it, despite repeated invitations to the membership to do so.

I dont think its a good idea for me to do all the work on it alone. Combine that with a number of my interactions here on Loomio being met with unnecessary rudeness and me observing the progress of this move for greater centralisation, while having my concerns largely ignored, and you may get a better understanding of why I've found it hard to find the enthusiasm to move the Participating Budgeting initiative forward and what the real cause may be,

Possibly it needs people to volunteer and/or be voted into a "titled" position (of power? - but we already shared the power) for them to feel they have a need or responsibility to actually do something and that this structure will see more activity. I find that sad. I dont want to be all negative but predict that we'll find it hard to muster up active participants to fill all these new roles and take on some of the work of organising the co-op,.as we always have with existing roles. I'll be pleasantly surprised to be proven wrong.

As an outsider I have sympathy with both @Darren @Nathan Schneider positions. Darren I wonder if the process of getting a seat or specific in the OC makes a difference to how you feel about it?

If for example the working group members and OC roles were decided by election by consent would that allay some of your concerns?

Or is it the existence of verticality/ hierarchy that matters to you, irrespective of how that verticality/hierarchy comes about?

@Sam Rossiter The consent election stuff looks interesting. Not totally sure how that would work out in a project like social coop with a large membership where most people know little of each other. My gut feeling is that there would be a trend for a few visible/active folks to be repeatedly pushed for election.

My main concern with the proposal doesnt come from a complete hatred for all possible forms of hierarchy. Its that we are making a load of work to organise all this. More so if we are going to keep good lines of communication into and out of this group and pathways for participation open. Whos going to do all that extra work? Whos going to fill all these new roles? What makes this new group more able to make good decisions than the wider membership where theres a greater breadth of knowledge and experience? Are these few committed souls expected to spend even more time understanding and following all the various workings of the co-op so they can provide better informed decisions? Are we going to create something more opaque and harder to engage with where, as often has been the case to date, a few people get stuck carrying the burdens of running the coop with little awareness from other members or chances for the load to be shared?

@Darren @Sam Rossiter One thing I really like about the proposal is the use of sortition, so people don't have to run for election, which doesn't feel right in a community where I basically trust everyone to hold roles of responsibility.

Darren, I would love to hear ideas on how to better encourage people to participate. My own direct experience from working groups is that it is very hard to coordinate participation and action without intentional collectives spaces for doing so. Otherwise, I will tell you how things usually happen in this co-op, having passed a few proposals and achieved some changes in policy: a ton of individual initiative, and bothering others incessantly until they surprise you with their creativity and brilliance. Similarly, with the PB thing, my expectation would be, under the current framework, that nothing would happen until the bottom-liner proposes and sets up a basic framework into which others can engage. If we all wait for spontaneous collectivity, I am afraid, nothing will get done.

This has been a problem we found in the Finance WG. For months we have been saying "We have lots of money, please spend it!" And nobody is sure how to start doing that. Budgeting is not something that can be done ad hoc; it requires coordination.

That's why I think it is extremely important to have a central space for coordination, commitment, and identifying shared priorities.

[Posted originally in the wrong thread, Loomio on my phone is a lot. Lol

]Interesting notion. I wonder if "polities" could be split while "infrastructure" remained "commoned". In practical terms could this mean an experiment in common tech labor and revenues?

Persynally I am more motivated to defederate from Meta less because of a negative (threat) and more because of a positive affirmation:

I want to be in community with people who want something other than the social logic Meta offers.

Feeling that erosion..