Superstruct: Botboss
I am motivated by rewards, not punishment; by attainment of hopes, not avoidance of fears; by promise, not threat; by thriving, not surviving. So, it has been deeply difficult and emotional. After weeks of agonizing, I have finally contributed something to Superstruct. Here it is.
Botboss
The Botboss strategy combines Prediction Markets with near-mindless games like the Google Image Labeler. Prediction Markets give cunning, wealthy, powerful sociopaths a larger incentive to tell us the truth, than their incentive to trick us into voting and spending against our own interests. But actual work is needed. For that purpose, mindless games can be used to motivate apathetic, unqualified slackers to get off their butts and save the world, or at least stop destroying it, by entertaining them while they follow orders and measure outcomes.
Bad news, then good news.
The bad news is, no plan for saving the world can succeed without reworking the incentives of driven sociopaths and apathetic slackers. All Superthreats stem from their interactions of deceit and laziness, respectively.
The good news is that a world in which both of them are productive, through ignoble means such as gambling and goofing off, would be a much more attractive world to live in than a world of miserable self-sacrifice!
Part 1 of 3: Prediction Market
Which solutions will help improve the condition of a Superthreat, which ones are an ineffective waste of resources, and which will make some other Superthreat get worse? As futurists we cannot afford the naivete so often seen in the technology press, lauding forthcoming innovations that never materialize.
Yet the issues that are pressing today are often the most uncertain ones, and the ones requiring the most technical expertise. Our Superthreats concern global systems of unprecedented complexity and subtlety. After we generate solutions, we must have a reliable mechanism to judge between solutions. We need a competition that reliably rewards the filtering of good solutions from bad ones.
That game is the Prediction Market. In our case, a Prediction Market would bet on a whole host of competing alternative plans, called "missions". In each one, they would bet on the relative likelihood of a spectrum of outcomes from their mindless drones carrying out the simple tasks in a given mission.
Part 2 of 3: Doing the Actual Work and Measurement
Information without action is useless. Someone has to carry out the various plans suggested by the Prediction Market and gather the resulting measurements to determine who wins the bets. This has nothing to do with coming up with solutions. In fact, they are absolutely not responsible for coming up with solutions, and I for one would feel relieved to be among them. Let’s talk about these people for a moment.
I see many around me, myself included, for whom the problems of the whole world are folded into the scope of what we are expected to care for, on top of each individual's own challenges. It is intimidating, and frankly impossible. The dumb ones become cranks with crank solutions. The smart ones realize they are unqualified and often struggle with suicide attempts as a result. Their apathy is the alternative to collapsing and coming unhinged.
A lot of these same people burn off stress and aggression by playing games. I am not talking about Alternate Reality Games, as those are too open-ended. I am talking about cleverly-designed reductionistic systems with a knowable set of rules and unambiguous scoring. These games are demanding enough (be smarter than a computer) but not too demanding (be a genius who saves the world from Superthreats or you die in real life).
Example number one: The Google Image Labeler is Human Based Computation. This project proved through demonstration that you can Botboss massive numbers of people into doing your work for you with the time they would have otherwise spent playing Solitaire.
Example number two: Suppose we had a flock of robots for sorting a landfill into different materials for recycling. The robots would probably be pretty dumb, and couldn’t do it without human supervision. A human Botboss of such a flock could wear glasses with an Augmented Reality system. Or they would just view a representation of it on their computer screen while tele-operating the flock from home. The Augmented Reality would overlay clockwork pirates and a rainbow-colored candyland on the scene, and instruct them in what kind of judgements are required from them and how it is scored. They could have as much fun as playing Pikmin. You don’t need to be a genius coming up with global solutions in science, engineering, economics and politics to play Pikmin. Landfills are just a lot of work that needs the attention of lots and lots of people for a huge amount of time.
So this is the basic philosophical design strategy we’re talking about. It works as follows.
Zillionics (see "Zillionics" in this essay by Kevin Kelly) are employed to obtain measurements to decide the winners and losers of the bets on the Prediction Market. Users volunteer to be spied on all day by chips embedded in a collectable trading card in their pocket, as they carry out competing strategies. The cards are basically Spimes, because their progress in real life is tracked on a game application through the internet. The printing on the card instructs the user which simple task to carry out, in one of the competing strategies. To judge this competition, dozens of different missions would assign roles to dozens of different types of players, measuring hundreds of factors of the physical and social systems, and carrying out double and triple blind tests (see "Triple Blind" in this essay by Kevin Kelly) on each other’s work, conditions, and the measurements themselves. The roles are specific, narrowly-defined, and measurable. Administering and mediating the game system is, itself, also a game.
By harvesting the data, a strategy is now proven to have a maximized beneficial net outcome with minimized side effects. It has become the “strong money” bet on the Prediction Market for that particular issue. There are now more mission cards distributed to win points by carrying out the various simple tasks of that strategy. Mission cards for less-effective plans are phased out, except a small number to measure how they work in the changed conditions that result from the success of the currently-effective strategy.
Two requirements of missions for millions of apathetic, unqualified slackers:
(1). Narrow the scope.
* the scope of the challenge presented to the player
* the scope of the work the game is quietly extracting in the background from the gameplay
* well-specified
* well-understood
* then stimulus and response can be quickly connected to carrots and sticks, to create the brain chemistry of addiction in games
(2). Hide the problem.
All world-changing solutions need to get done work that is boring or intimidating. The players should not have that pushed in their face. This helps a lot. For me, that is the central key that makes the concept so potentially wonderful! Disguise a boring and intimidating problem as something else. Theme the playing cards with clockwork pirates or something. Fool the brain. It's a beautiful mind hack! If they were willing to do boring and intimidating things, they’d be doing it already. Why else would it be necessary to turn it into a game to begin with? Banish the paralyzing overwhelm, the drudgery, and the gloom with a game--don't allow paralyzing overwhelm, or drudgery, or gloom into the game.
Part 3 of 3: Issues
A. Sabotage of plans to win on the Prediction Market. How to catch it.
B. Corruption of measurements to win on the Prediction Market. How to catch it.
Please contribute your enhancements of the Botboss strategy. Thank you for reading.
Comments
temujin9 on Nov. 1, 2008 8:02 PM
Using games to motivate real-world action from the apathetic and using markets to solve large information problems are both good ideas (although difficult to do right). Superstruct appears to be a collaborative science fiction storytelling game, so some authorial license is allowed. In the real world, however, I'd be quite leery of a plan that coupled two largely untested ideas together without good cause. It might be better to think of the prediction market and botboss strategies as separate modular components: if one proves unreliable, it can be replaced or reformed without impacting the other much.
temujin9 on Nov. 1, 2008 8:07 PM
Also: I appreciate a good RTFM as much as anyone, but I appear to be outnumbered badly in that. It's really easy to construct hyperlinks to Google Searches, and I've found that's a more subtle and acceptable way of saying it.
matt-arnold on Nov. 1, 2008 8:47 PM
Apologies. I got it out hurriedly. I'm adding links now.
temujin9 on Nov. 1, 2008 9:35 PM
No apologies necessary: it's an editorial criticism only.
temujin9 on Nov. 1, 2008 9:37 PM
Ah, one additional editorial: you only need to link the term the first time it occurs.
matt-arnold on Nov. 1, 2008 8:12 PM
Indeed. Unfortunately, I am assigned to make a "superstructure", which means layering one structure on another. This is my best satisfaction of that requirement so far.
temujin9 on Nov. 1, 2008 9:32 PM
Ah, gotcha.
Leave a Comment