[[Common knowledge is almost impossible to do without synchronous communication or a hell of a lot of redundancy]] 💬 open ended design rant ❤️🔥🧵
Why text-based tools can't produce real common knowledge
16 tweets · August 2021 · 47 likes · 2 retweets · read on Twitter
To be really clear: common knowledge is not just "everyone in the group knows X" or "everyone in the group knows everyone in the group knows X". It's: 1. everyone in the group knows X 2. everyone in the group knows 1 3. everyone in the group knows 2 4. <repeated ad infinitum>
The recursive infiniteness of this sounds exhausting but in practice once you actually hit *infinity* it suddenly becomes simple again. If you're in a scene where there's some secret and you're trying to track who knows it & how to speak, and someone blurts it out, it collapses.
Say you're with friends, & you know some of you are going to a party, but you don't know if Bob is invited so you don't the mention the party in case Bob feels left out. Then Chad says "can't wait to dance with you all tonight!" and Bob goes "yeah!" Instant relief & clarity.
But suppose Bob goes "uhh... what?" Common knowledge breakdowns, when people act on a basis of thinking everyone's on some same page, and they're not, tend to produce distrust/betrayal, whether social shame like the party, or literal death if trying to coordinate battle tactics
Group chats are the closest tool we have for common knowledge in text-based mediums, but they have a lot of gaps. They solve the secret problem (above) because once something is in the group chat, you know it's not a *secret*. But they don't let you *ensure* common knowledge!
If I post X to a 3-person DM and later I see the other two people both have reacted or replied, have we achieved common knowledge of X? Not remotely!
For one, they don't know that I've seen their responses. But even before that: whoever responded second doesn't know if whoever responded first saw *their* response.
In practice, this somewhat resolves iteratively: as we keep chatting we assume we've all seen the previous messages, but stuff DOES get missed, especially as the group gets larger. It's no match at all for the obviousness of seeing everyone in the room react to X at once.
"So!" I declare triumphantly, "let's make a Common Knowledge module for Roam or Slack or whatever, where once the whole group reacts to some content, they all get a notification saying they've all reacted to that content!" Problem solved?
Better, probably, but not quite. You see, how do I *know* that everyone has received & read that notification? I don't. So we'll just make a button to acknowledge the notification, and once everyone has done that, we'll... send another notification? ...shit. 😅
I want to be clear that this is not a strawman critique of a proposed solution—this is what my brain generated while attempting to solve it, and it's still as close as I have to an answer that doesn't involve having everyone's attention in the same place at the same time.
If an organization has enough consistent shared habits (via discipline or otherwise), I suppose this can be a good approximation too. If everyone knows that everyone knows (×∞) that everyone gets & reads the same morning briefing, then you can treat that as common knowledge.
I suspect that in the days before decentralized/personalized media, it was effectively common knowledge among certain groups that everyone would have read the same newspaper that morning. Now that only happens for the biggest items, and barely/fuzzily
Some remote teams may be able to have such a practice, but don't underestimate how long it takes to build the trust that everyone consistently reads the briefing, and how quickly that trust is destroyed (if one person skips, even secretly, *they* know the system is untrustworthy)
I think in practice the solution is to make good use of live meetings for this purpose: Have attendees bring in things they want to get common knowledge on, & then just speaking those things creates comknol (assuming everyone can tell everyone at the meeting is paying attention)