Something I’ve been stuck on for several years. There’s the part of every task that you do to get it done, and there’s the part that you do to benefit from the process of doing it. We’ve made this pretty bimodal: with some tasks (literary reading, expressive writing, educational problems) that are primarily about the experience and others (most things at work) that are about the result. Some things are clearly both (sports) but I think there are some things that are invisibly both. It’s important because it influences where we apply technology. If there are things that we think are about the output that are actually about the process, then they shouldn’t be automated. A good potential example is note-taking: if it’s less about the notes and more about the taking, the mass adoption of LLM summarization may not serve us.
I’m always afraid that maintaining governance systems (taking notes, facilitating meetings, getting to agreement, processing information) is about “both”. What if low-tech manually maintained governance helps us keep our self-governance muscles in practice, maintaining the habits of running effective teams, exhibit and instills values of service (commitment to a team’s mundane work is a strong signal), create things that are socially valuable because they were costly to provide (a well run gathering as a valued ritual), and even evaluate our peers (observe who is responsible in their commitments and trustworthy in their notetaking)? If there’s anything to any of this, then governance technologies may be the last thing we need: the most counterproductive thing to invest engineering/automation/technical effort into.
There are a few obvious counterpoints, namely that those things can be developed (maybe better developed) by doing the more meaningful work that’s not automated. Maybe that’s right, or maybe it assumes too much learning transfer, and certainly is disregards the credibility and legitimacy of drudgery: in community houses cleaning the toilets was acknowledged as less desireable and more appreciateable work than cleaning the kitchens
There’s a great example in Ostrom’s work of technical improvements killing a commons
The more constructive version of the same question: what is the role of technology in tasks that shouldn’t be automated away?
p.s. turns out Brian Eno says the same thing about music in AI:
https://bsky.app/profile/karlbode.com/post/3ldf6zg65ic2b
