Ready-to-mind

Technologies of coordination

I think cryptocurrencies have mostly caused financial and environmental harms, and facilitated a lot of immoral activities, but you have to give the Bitcoin protocol some credit for trying. The creators found a completely novel financial and political possibility and built the technology to make it possible. They showed how it was possible to maintain a decentralised ledger with very low trust assumptions.

In my last post, I listed the purposes of computers - and the one that stuck out to me most was the use of computers for collaboration or coordination.

Examples of technologies of coordination:

I’m inspired by some ideas from Ivan Vendrov on his substack Nothing Human. He believes that we have a shortage of collaboration and asks the question, how can emerging generative AI technology help us collaborate better? He paints the image of the use of LLMs to facilitate a discussion of 100 people. I like this idea. In a video conference call of 100 people, most people would have to sit idle and listen, and even those who did get to speak would probably feel like they had to be very careful or feel undeserving of that much airtime. The key issue is that only one person can speak at a time.

A common strategy to make a meeting like this go well is to have breakout rooms where smaller discussions take place, and then representatives from the groups summarise their discussion for everyone else to hear. This gives more people the chance to speak (as people are speaking in parallel in breakout rooms), and improves the overall relevance and quality of what people hear (since only the most important ideas are broadcast to everyone).

Imagine if we had a new meeting format where 100 people could engage in a discussion mediated by an AI agent. Everyone would speak in parallel with an AI agent, but these agents would respond with views that represent the rest of the members in the call. The AI could for example curate towards maximising disagreement to elicit the best arguments. This would maximise how much time people spend speaking and listening to the most relevant ideas for them to respond to.

More design is necessary to make this idea more palatable and right now it is a bit too cognitively and theoretically motivated, but it points to new possibilities for human collaboration enabled by technology. A text version of the same idea seems very practical. A company could arrive at an important decision through a focussed period of a text chat interface version of the above.

Imagine if this could scale to 1000s or millions of people, and be a mechanism of understanding people in your country, or sensemaking between two sides of a conflict.

Since computers are tools for the mind, the use of computers to boost collaboration assumes some cognitive limits that prevent better collaboration without the use of technology.

Where might we humans have limits that prevent us from coordinating more effectively?

It would be very interesting to make a detailed taxonomy of these human limitations and common issues in collaboration, and then to match them to technologies that augment our ability to collaborate. Any unmet needs could then spin out research questions about how to design better collaboration tools.

#gentle-computing