Conversation

"Meta-rationality" is "figuring out what's actually going on in and around a system, and doing something about it," which is invisible and illegible to both MBAs executives and technical people whose education ended in rationalism.
Quote Tweet
Replying to @arachnocapital2
Yes, I intend to put a paragraph about your latter point in the thing I’m writing now. Upper management is desperate to make software development predictable and controllable, and they only way they know how to do that is to assign rational functional roles & responsibilities…
6
35
Replying to
This work is undervalued because it’s a labor market for lemons: the “seller” knows whether their metarational glue work is valuable and high leverage, but the “buyer” (manager as proxy for org) can’t tell easily. Same downward price spiral other lemon markers get.
2
9
Replying to and
Inability to distinguish quality work -> attracts grifters, lowering average quality and thus average price -> withdrawal of high skill players, due to low price -> further drop in average quality…This can be solved through the ability to distinguish quality work reliably.
1
9
Replying to and
Unfortunately, consistent evaluation is usually the hallmark of normal “rational” evaluation. Which is exactly why people are always trying to standardize glue work and drag it into the realm of the standardizable. Unfortunately that destroys most of the context-driven value.
1
11
Replying to and
I implemented something very similar to the “metarational” workplace you describe at Twitch in the early days. It has a lot of the upsides you describe. Unfortunately it has a lot of ways in which it collapses, particularly with greater scale.
2
2
Replying to
I would love to hear more about that! Consistent with confidentiality ofc. “An Everyone Culture” is about this, and there are a bunch of red flags in there as well. The question for me is how to make it work in practice and not just theory.
1
6
Replying to and
Thanks, this was really good! It’s a nice case study more in what I call “advanced rationality” than “meta-rationality” (a qualitative distinction, although the borderline is somewhat nebulous).
Image
1
Replying to and
This bit was funny because I had exactly this experience. Then when I was mentoring who was learning software dev, I set "implement your own ORM" as an exercise, both to demystify ORM magic and because it forces an understanding of the half-baked python object system.
Image
1
4
Show replies