Comment by DanielBMarkham | Hacker News Reader

Comment by DanielBMarkham | Hacker News Reader

Parent Full thread

As one wag said about a famous politician many years ago, "That guy's got a hundred ideas a day! But only one or two of them are worth a damn" So yes, I got some ideas.

I hate going so far down in a thread. I will try to be as meta as possible.

I think the type of work we're talking about, identifying the problem and making progress here, is meta-analysis. That is, I think the first thing to do is to describe the categories of things that are creating the systemic interlocking dysfunction. It is definitely _not_ to look more and more closer at the current problem. The only thing you're doing there is subtly reinforcing the problem in everyone's minds. As Einstein said, you can't solve a problem like this using the same old ideas and terms that got you here.

So you get a small number of literate, smart, and historically-minded people. You try to come up with half-a-dozen or so similar systemic dysfunctions that existed and were resolved throughout history. That's a conversation. Probably will take a while. Based on that conversation, you come up with 3-7 overall concepts that seem to categorize the types of misconceptions that exist in all of these examples. As an example, perhaps you decide that definitions of incentives, mission, conflict, and identity are areas in which in each example participants had one idea on one side of history and another on the other (as the dysfunction was resolved).

Once you've got that group of commonly-incomplete assumptions, then and only then can you take your mental blinders off and come back to the present situation and bounce what you've learned up against what's happening now (and the tech involved). At this point you'll have a framework for looking for behaviors and tech that might make a difference. That may be as far as you can go because of the before-after problem, but it gives you places to explore and support and some idea of the types of disruptions to look for that might break us out of this.

E-mail or IM me if you'd like to talk farther. As much as we love to gesticulate, rant, and moan about this problem online, actually making progress is real work and not just something we do with a morning cup of coffee typing into a text box into the void. :) (Although that's a lot of fun too!)

I suppose we could start by finding a historian who can enumerate all such systems, starting with the Greek city-states and ending with the 2008 mortgage meltdown or thereabouts. Then for every system that fits our criteria we can find a historian or an economist or another specialist. Then we could create a sort of forum with both experts and thinkers, trying to shake out commonalities and validate them against the entire set.

That’s a plan. But I have to say I feel like you’re trying to boil the ocean to make a cup of coffee here.

EDIT: are we talking about Facebook still?

Yes we are still talking about Facebook.

In regards to boiling the ocean, it is a hugely-complex system involving billions of people all interacting in millions of various dynamic subgroups using technology sometimes in ways even the designers can't anticipate. There's that.

No, this historian idea would not work. We need to look at these examples from a CS perspective, not from a historical one. Those guys wouldn't have a clue. You'd end up with 50 essays and a lot of nothing. Put a different way, what we want is a collection of generic classes that we could instantiate that would describe all of those systemic failures. That's it. You could do that in an afternoon, a week, or a year. I don't think it is as long as a year.

But like all coding practices, the key here is a small group and agreed-upon code. This isn't a sample-the-universe problem. Perhaps a way of looking at it is that yes, it's an ocean. It's an ocean of oceans we have to boil. Tech, people, and all of that beyond anybody's ability to grasp. But because it's so huge, we're free to then come at this in much the same way as mathematicians began reasoning about types of infinities. You don't have to know the infinity or collect all of the information required in order to effectively reason about it.

I'm not talking about a big-bang, monolithic solution to social media. I'm talking about a few commandos arriving on a strange beach, taking a look at their surroundings, coming up with some heuristics, then testing them. It's not a long, drawn-out complex process. In fact, just the opposite. (But it is a fundamentally different kind of problem than computer folks are normally exposed to. We're looking for quick set of heuristics to test, not an accurate or complete definition of the overall problem and solution. We use CS knowledge and complex systems understanding to make informed categorical guesses that we can then test)

Hope that makes sense. I'm outta here.

EDIT: You may be asking for a solution to the FB/Social Media problem. I am saying that the best I think we can do is come up with a workable way to score possible solutions. That's it. Since it's scoring, we could use the Great Pumpkin or something more formal. I'm voting for more formal.

Reply on news.ycombinator.com