Incentives and the Cobra Effect

Discreet users, or discrete users? (I initially assumed you meant discreet, i.e. not wanting to share login details, but the comment read oddly as a result.)

Something I hope Netlify discovers...

Quite frequently. Personally, I think part of the issue is explicit vs implicit noise in the system. There's always noise in the system, so there is always reasons to reject or criticize something. Very few things have unobjectionable global optima. But if noise is explicit in the system, people tend to take a bit more care. If it is implicit, well people act like the metric is doing exactly what it is intended to do and can't be hacked.

Momentum is a pretty powerful force.

[Edit]: I thought I'd add a clear example because I was reminded of it. When I tell people that return free filing taxes could really simplify taxes for people at first agree and then start pointing out a bunch of edge cases. So I argue that if we could just get the 15 million Americans who use the 1040 EZ then that's a big step in the right direction. It never goes well and it is never considered that just autofilling everything you normally put in by hand significantly reduces time regardless of how you do your taxes. Maybe I'm just bad at making cases, maybe people are hard of hearing, maybe both. (this is not the argument I make in full, obviously)

I've observed that too and it raises some interesting psychological thoughts.

A linguistic distinction may help if we say that incentives are not quite the same thing as motives.

Motives can be intrinsic, coming only from one actor. "I am motivated to learn the piano". (See Pink, "Drive" etc)

Incentives are a relation between two actors, one (leader/manager) who sets the incentives to shape the behaviour of another. The 'incentivised person' is acted upon and is the target of the incentives.

That means the leader has to have a working model of both the environment (and other parametric factors) and of the group or individual they want to incentivise.

If you challenge that internal working model you're not just playing with logic and reason about parameters, you're challenging someone's perception - and that's much closer to the ego.

There was a good article posted earlier by Christine VanDeVelde Luskin on intrinsic and extrinsic motivations in the Bing preschool. Kids always know when you're playing games at "incentivising" them. They know it's a "bribe", which lessens your integrity, and they act up in response.

In my opinion the manipulator has a bit of a binary choice. Either be absolutely up-front and tell people "I am doing X as an incentive for you to behave as Y, and I expect accordant behaviour", or, be so clever as totally hide your manipulation. The former requires the overt exercise of power, or at least risks the vulnerability of total honesty.

I think nost perverse incentives come from people treading the middle ground, being deceptive, thinking they are being clever nudging others. But those others see right through the game.

It's super frustrating! Someone says, "Something needs to be done about this problem, and X is something, so X needs to be done." Then if you say X will backfire, you're taken as supporting the problem. It's a real rhetoric challenge.

You are absolutely right. Good intentions are not enough. We should all keep repeating this at the tops of our voices.

Yeah they sure are extra super stupid. oNE Would think they could induce the stupid, but I've been hitting myself in the head with a fry pan all weekend and I still not the stupids!

Not just perverse incentives, or incentives in general. People have a hard time thinking about second-order effects of all kinds and tend to think quite linearly. They're treating you the same as auras and astrology because they are unable to tell the difference.

Did that work out well for the business? Maybe that’s what they wanted? It’s a serious question.

> Rather than just trying various policies in simulation, we argue about them dogmatically until the end of time.

That may be because what people argue about is very often not what they actually care about, because what they care about is outside the Overton window. You can't really argue that we should just harm black people for being black in order to make miserable white people feel a little better about their lives in comparison, so instead you argue about drug laws and being "tough on crime". You can't really argue that the poor and sick should just die on the streets (although this did get an applause at a Republican convention) so instead you argue about people taking advantage of the welfare system. You can't really argue that men should have utter dominion over women, so you instead argue about killing unborn babies. These things move the Overton window in the direction you want, so that next time you can argue the more extreme version.

Who would want to simulate these things? That would just open up the possibility of being proven wrong. Wrong or right has nothing to do with it: they just want to hurt and control people.

> After a lifetime of witnessing tax cuts grow wealth inequality and burden the poor and middle classes, I view people in favor of them with concern. I wonder if they're able to extrapolate cause and effect to understand the ramifications of their choices.

But they know that tax cuts grow wealth inequality. That's what they want.

> And they look at me the same way, I imagine as an astonishing waste of potential

They don't look at you at all, unless they think they can make a caricature of you to parade on the news and use as a straw-man punching bag. Otherwise they barely know or care that you exist at all. They certainly don't give a shit about your "potential".

Obviously everyone knows exactly which half of people I'm talking about.

Also, perverse incentives arise spontaneously all the time. Business types love talking about this, they will talk your ear off about it, but only from the angle that looks to exploit them as tactics or strategic moats with studiously suppressed consideration for the first order effects let alone second or third. If you so much as hint that you think maybe there ought to be a countervailing force somewhere, suddenly COBRA EFFECT HOW DARE YOU COMMUNIST CENTRAL PLANNING ECONOMIC ILLITERATE FILTH! It leaves me with the distinct impression that nobody really believes in the system, let alone "warts and all," and instead just grabs the nearest self-serving principle off the shelf to rationalize whatever they want to do.

Why is that a clever idea? If you can make the landowners pay a fee when a rhino gets killed on their land, why not just make them pay a fine instead? But, it's South Africa so the thinking there seems a little muddled recently.

Or it's an Indian trope "look how stupid the British are" ?

Really? What about "look how clever the Indians/ natives are?" Not everything is racism. A similar result occurs every time something like this is tried regardless of race. Countless examples including in the U.S. and yeah people will run scams like this even if the scam is convoluted and twice as much work as doing the honest thing.

Im inclined to believe this theory based purely on how miss leading most written history has shown to be in terms of racism.