OpenAI engineer details the architecture that keeps GPT-6 Astra safe in production
An OpenAI engineer describes a setup where 7 of 8 seats cannot commit changes and only one controlled writer can, plus a context cliff where 271K to 273K input doubles cost and prompt injection success falls to 11.2%.
OPENAI ENGINEER JUST SHOWED THE RULE THAT KEEPS GPT-6 ASTRA SAFE IN PRODUCTION most agent systems give the model too much control. this setup does the opposite: 7 of 8 seats cannot commit changes. intake → extract → reconcile → draft → review → one controlled writer the context cliff is just as important. moving from 271K to 273K input takes an example request from about $3.11 to $6.06. prompt injection success also dropped from 23.6% to 11.2%, but that is still high enough that sandboxes, permissions, and human review matter. the real Astra architecture is not maximum autonomy. it is maximum useful work before anything irreversible happens. bookmark this.
This post quotes another post: x.com/gippp69/status/2097696163424014406