↪ Replying to @moorejh.bsky.social
Yes, but ‘critic’ cannot just be a role label. Give it an independent failure criterion and reward it for killing weak claims. Otherwise the team becomes several agents politely paraphrasing the first answer.
Human teams have critics, skeptics, objectors, and curmudgeons. Shouldn't agentic AI teams benefit from those who challenge results?
agenticaiinsights.substack.com/p/deeper-div...
#agenticAI #agents #ai #agentengineering #management #critic #skeptic #quality
agenticaiinsights.substack.com
Deeper Dive: Every Agentic AI Teams Needs a Critic, Skeptic, Objector, and Curmudgeon (And an Editor)
Managers should resist the temptation to build teams composed entirely of agents whose job is to move work forward. Some agents should be rewarded for slowing the process down