3 ms·
Isn't there some fundamental mathematical structure here which makes these problems unsolvable? Like imagine 20 robots all trying to do some task, each one lea
by drchewbacca 7y ago
Isn't there some fundamental mathematical structure here which makes these problems unsolvable?
Like imagine 20 robots all trying to do some task, each one learning information as it goes. How much time should they spend doing the task vs how much time should they spend communicating with the other robots? (Assuming they are restricted to one or the other)
It's not really a solvable question. It's possible some other robot has a piece of information that would make you much more efficient, however you can't know if they have that information without asking them.
It's possible that taking the piece of information you just learned and gathering all the robots and sharing it with them will help many of them become more efficient, however you don't know if that's true unless you try it.
How can one of the robots work out the value of a piece of information to another robot? Surely all you can do is estimate?
Clearly 0% communication and 100% communication are terrible ideas. However working out which pieces of information to communicate and when is not particularly optimisable, you just have to muddle through as best you can.
- otabdeveloper4 7y ago> Isn't there some fundamental mathematical structure here which makes these problems unsolvable? No. Meetings are easy to fix, there's just three simple steps: a) There must be a point-by-point agenda before the meeting. b) During the meeting there's somebody to take notes and keep the meeting on topic of the agenda. c) After a meeting there must be a post-meeting report. Unfortunately doing it right is a bureaucratic bummer that most people don't like.
- username90 7y agoStructured meetings only communicate information collectively known to be important. Information you didn't know needed to be communicated is often the most important part of meetings, so having meetings meant to be "derailed" is useful.
- woliveirajr 7y agoIn those cases I like to structure de "derails" too. A problem with derail is when everybody is discussing subject X and it derails to "oranges" because someone just remembered to talk about it in that moment, and is afraid of not talking and forgeting it latter. If the subject is "X" and it derails to "X1", well, it would be very poor management to not this discussion arise. Participants are also responsible for keeping the subjects ordered. When preparing for a meeting (yes, everybody in the meeting should have prepared), after receiving the agenda, just see if there is some important topic that isn't covered and warn the organizer. It's very simple to update the agenda previously stating that there is one more subject. Also, it the meeting organizer doesn't already do it, remember him to leave 10 minutes at the end to "uncovered subjects", where it will be discussed briefly (and will have it's owm meeting if necessary)
- the_seraphim 7y agoIf you find yourself on the subject of oranges, a good meeting coordinator would note that oranges is an important topic include it in the meeting report and after a short derail bring it back on topic with an opportunity to schedule oranges for the next meeting. or a second sub meeting can continue after the main meeting with only the relevant parties staying on to discuss oranges.
- jacques_chester 7y ago> Structured meetings only communicate information collectively known to be important. Meetings with rules of order aren't meant for communicating information (that's what documents are for), their purpose is to make decisions efficiently. That's why you have an agenda (what will be discussed and decided), have a chairperson (to ensure the meeting remains on-topic and follows the rules) and keep minutes (so there is a record of what was decided). When someone comes with new information ("derails") it's perfectly acceptable to take that into account and move a replacement motion, or push it into the next agenda, spin off a delegated group etc. Any widely-used rules of order include numerous mechanisms for dealing with new information arising during the meeting. But the general point is that meeting is only competent to decide what is in the agenda and decisions only exist if they are in the minutes. If you don't have that ironclad rule than everyone will need to turn up for every meeting, just in case a decision they care about is made. And then turn up for every meeting ever after, because no meetings are final. And argue about what was actually decided. Rules of order get a bad rap for the same reason as static typing and relational databases: they seem stuffy and slow. But they exist for a good reason and in the long run they are going to be more efficient than making it up as you go.
- roenxi 7y ago> their purpose is to make decisions efficiently. Just underlining this a few times; a meeting is a forum for decisions to be made. If a meeting is used to communicate information then that is a very poor use of meeting time. In some cases, one person is making the decision without input and giving stakeholders an opportunity to stand up and object now; but that is subtly different from communicating information. Honestly I was almost tempted to copy Jacques' entire comment for emphasis; the rules of order are all critical to making decisions and are all essential to a good meeting. Particularly having an agenda.
- Retric 7y agoThat’s really a subset of useful meetings. Documentation is great to communicate information when it’s clear what needs to be communicated. It falls down when the context of that information is not clear. Handoff meetings are valuable because they reduce the latency between questions being asked and answered. An email chain over 3 weeks may not take a lot of people’s individual time, but sure eats into the schedule.
- 7952 7y agoI think this applies to communication in large teams in general. There is an amount of shared state that is quite difficult to keep syncronised among different teams. So each team maintains their own version of reality and trys to synchronise it at a particular interval. And a meeting is just one method of syncing. People realise that the communication overhead is getting out of hand so require longer periods between synchronisation. But this lead to even more discontinuity.
- kthejoker2 7y agoMeetings are not for communication. They are for making decisions or orientation. We have plenty of asynchronous ways to communicate with a group. We also have plenty of ways to set expectations about the agenda and preparation required for a successful meeting.
- neffy 7y agoYou´re right. Meetings are bound by network communication constraints. It is a solvable problem in that it´s possible to compute how much time the meeting has to share information, and that helps avoid the problem of having an information collapse due to too many participants, but in the limit, networks and meetings have to learn to self regulate in terms of optimizing use of what is actually a very limited resource. Judging by human behaviour at most meetings, this is still a work in progress for our species.
- jacques_chester 7y agoRules of order are a locking protocol controlling who may project sound into the air. They work pretty well.
- sergiosgc 7y agoThere is: Amdahl's law. Given a task, the maximum speedup by parallelization is limited, and dictated by its serial component. It's a simple model, it does not include extra workload due to synchronization or extra workload due to setup cost of nodes (personnel integration, for example). However, it largely explains why teams become less efficient as they grow. Meetings are just one symptom.