"Solve For"

Shoot me now

It was 5 years ago at some kind of brunch put on by Facebook that I noticed that everyone was using the phrase “solve for”. As in: “we are solving for business challenge x”.

And I hated it on the spot. And I have continued to hate it with a passion close to the homicidal.

I did wonder if it was just me. Then I found out that it wasn’t. Thank you, Jerry, much appreciated.

But I think it’s important to explain exactly why I hate it so much. And why I think these two words highlight so much that is wrong with how technologists approach the world.

As Jerry says, “solve for” is a perfectly valid phrase to use in the context of maths (yes Americans, that is a plural, come at me). Maths uses a lot of algebra which typically involves more than one variable - x AND y (and maybe more if you’re feeling all multi-dimensional and kinky, I don’t judge your polynomial polyamory). So you need a way of referring to the variable that you want the student to tackle - e.g. “Solve for x if x-3=y and y=0”. In this example it should be bloody obvious which variable you want the student to tackle but in many other situations, it helps to remove ambiguity.

I have no problem with this version of “solve for”. That is absolutely fine.

But lets be clear when you use this phrase. You use it when you are tackling a maths problem. And maths problems are these weirdly artificial things. They are not necessarily easy. But they are a particular type of hard. “Prove the integral of 1/x = ln x” is a different kind of problem to “Your two sisters hate each other and want to sue each other after your father’s death and are dragging you into the whole thing”. Lets call these Type A and Type B problems. There are many of types of problem in the world but just go with me on this for now.

Computer scientists are trained to solve Type A problems. Quant types with a heavy background in maths get very good at solving Type A problems. So much so that I suspect that they tend to think of the world as consisting of mostly of Type A problems. But it doesn’t. The vast majority of human beings who have ever lived cannot tell you what the integral of 1/x is and that hasn’t been a showstopper for them. Type B problems are far more common and immediate.

N.B. I do not want to bad mouth maths. It is very useful. But its power is not absolute.

What worries me about the people who rule our world is that they think that everything is really a maths problem. Whereas most of our troubles are not only not mathematical, they are not even problems.

“What are you talking about?” you say. “I got 99 problems”. (Yes, all my imaginary conversations are with Jay-Z).

Well, maybe. But problems can be solved. What we have more of are predicaments, which cannot be solved, only managed.

Climate change is not a problem. It cannot be solved once and for all like x if x-3=y and y=0 can be. It is an interconnected set of issue that are unlikely ever to be completely resolved. There are better and worse outcomes but there is no right outcome.

If there is a right outcome, you have a problem, not a predicament. But mostly there are no right outcomes. Just less wrong ways of being.

So our technologists who want to “solve for privacy” or “solve for intelligence” are bringing a knife to a thermonuclear war. We think that the world consists of little maths problems where our big brains can find the right answer. But it doesn’t. It consists mostly of predicaments that we can only live through and hopefully make a shade less worse.

So don’t tell me you’re going to “solve for” anything. That just tells me that you don’t know what you are talking about.