Mildly funny analogy by John Cutler, niche audience, illustrating a failure mode that feels personally salient to me. Here’s how it begins:
Imagine if a restaurant behaved like your average product team. The kitchen is packed. Everyone is moving. Every station is busy. Prep lists are long. Meetings are constant. There is always something to do. Chopping, rearranging, documenting, planning, replating.
But plates rarely reach customers. When they do, they’re late. Or wrong. Or cold. Or oddly disconnected from what the diners said they wanted. Yet the kitchen isn’t “failing,” exactly. It never looks like a crisis. No one storms out. No one flips a table. Diners don’t riot. They just lower their expectations and stop coming back.
Inside the kitchen, though, the staff feels productive. Everyone is exhausted. Everyone is “at capacity.” Everyone can point to a dozen tasks they completed. They can even argue those tasks were important. And in isolation, many of them were.
But restaurants are not judged by how busy the kitchen is. They are judged by how consistently they deliver great food, on time, to the people who ordered it. Product development is strange because this feedback loop is muted. There is no instant revolt. A team can be unbelievably heroically busy without producing much that actually moves the needle.
That’s the trap: in software, effort is easy to generate, activity is easy to justify, and impact is surprisingly easy to avoid.
Yeah, this was the source of much personal consternation when I left my operations-heavy career path in industry to explore research roles, as much as I found the latter more intrinsically exciting.
It’s also what’s always back-of-mind w.r.t. the alignment-related work I’m most excited by, even though part of why I’m excited about them is how relatively empirically grounded they are.
Mildly funny analogy by John Cutler, niche audience, illustrating a failure mode that feels personally salient to me. Here’s how it begins:
(much more at the link)
Yyep. And it’s much much worse for research.
Yeah, this was the source of much personal consternation when I left my operations-heavy career path in industry to explore research roles, as much as I found the latter more intrinsically exciting.
It’s also what’s always back-of-mind w.r.t. the alignment-related work I’m most excited by, even though part of why I’m excited about them is how relatively empirically grounded they are.