11 points by hereticles 2 hours ago | 8 comments
- "normalised deployments," sounds even worse than tokenmaxxing
- I guess that tokenmaxxing phase has served its purpose, of forcing people's brains to go from "i will never use AI, it's scary and new and useless" to "i must learn to use it, this is the new normal, otherwise i'm getting fired".
And now they're able to move to the phase of "let's now learn to actually use it effectively".
- Kind of surprising. I had interpreted these tokenmaxxing initiatives at some of the larger companies as R&D investment, motivating employees to find the useful applications of a new technology with fairly broad applicability. Obviously some people gamed it, but I wonder if everyone gamed it or if opening the floodgates made for any useful discoveries.
- If any useful optimizations were discovered I imagine they belonged to people at the bottom of the leaderboards and hence weren't noticed by tokenmaxxing chad execs too dumb to have an ounce of creative thinking.
- Day 2 Amazon only cares about the KPI.
Innovation does not grow out of KPIs.
- > What happened next was predictable in hindsight. Employees began inflating their scores through tokenmaxxing: running meaningless tasks through AI agents to consume tokens and climb the rankings.
That was predictable in foresight.
- A metric was introduced. A metric was maximized. The metric stopped being a proxy for anything. A new metric is to be introduced and likewise gamed.