Thread
The Twitter debacle is at least partly explained by Karl Popper's 'Clouds and Clocks' theory. I wish it was more widely understood, so here's more.
thread...
thread...
Popper hypothesized that there were two kinds of problem in the world: cloud problems, and clock problems. Let's start with clocks...
Clock problems are ones with a discernable end state that's easy to identify. You make a clock, it tells time, job done; you know you've fixed it. Engineering is mostly like this. Doesn't mean that the problem is *easy*, just that it's simple to understand what done looks like.
Cloud problems are the opposite: there's no end-state, and for any problem solving you try, it may not be easy to understand if you're done. Relationships, psychology, politics, design. They're messy; indefinite, you may have an effect but not really 'fix' anything...
because the problem is messy and infinite (by now I hope you can guess where this is going)...
the real issue comes when you don't know whether you're solving a cloud or a clock problem. Engineers are prone to reducing the social world into a 'clock', even though it isn't one; doing so is actively harmful.
(very extreme outlier example) terrorists are frequently engineers by training: they believe that you can 'fix' hearts and minds and politics with a stick of C4!
Even worse is that there's often a smug superiority to it; a belief that because you're able to fix any kind of clock, that your approach is somehow better in *every* domain.
And this is what explains the bias of *you know who*. Electric cars and rockets are clock problems; Twitter is a cloud problem (with some clock stuff behind the scenes that fools you into thinking it's a clock).
This is why you'd eject entire teams of people working on policy and moderation without a second thought to the side-effects.
(side note) I've encountered this bias quite often in software organisations (being on the 'cloud' side, i.e. product + design!)
Another example: think of a time when you've filled in a form, and your situation doesn't fit the options you're given? That's clouds vs. clocks. Messy human reality != pre-engineered options.
In more general terms, being biased by your profession is called 'deformation professionelle' by the French; aka 'if your only tool is a hammer, everything looks like a nail'.
all of the recent threads by people who actually know their shit about moderation hint at this same issue e.g. @yishan
Complex social problems cannot be reduced or fixed in the way some engineers expect, given their existing expertise. Clock solutions don't solve cloud problems (but they can be a part of it).
This is not to say there's anything inherently wrong with the clock approach; we have wonderful gifts and achievements because of it. It's just not everything.
NOTE: I'm also not claiming that there's no bias elsewhere (see deformation professionelle).
(Thread over) - h/t to @rorysutherland - I think I learned about this from him, but my memory is fuzzy!
FYI I'm writing a book that on presenting design work. Explain work easily, deal with difficult conversations, get influence in your org.
Story before Screens: a practical guide to presenting high-stakes digital product design
Sign up for updates: storybeforescreens.com/
Story before Screens: a practical guide to presenting high-stakes digital product design
Sign up for updates: storybeforescreens.com/
I just realised something else of note: this (again partly) explains the failure of Tesla's Autopilot. Messy human driving behaviour (a cloud problem) is not easily reducible with a clock solution.
the clocks approach, right here:
Some are taking this thread *very* personally: so let me clarify: clock thinking is not exclusively a problem for engineers.
It's not a thread about engineering; but a style of thinking that happens to be a common bias in that profession, which *may* explain EM. Nuance people!
It's not a thread about engineering; but a style of thinking that happens to be a common bias in that profession, which *may* explain EM. Nuance people!
