I’m not a YouTuber or an AI performance researcher, I have 0 need to benchmark different models on the same task.
I’m a developer, I work for a company, I need to move tickets across a board, if my tests pass, I couldn’t give a flying “french connection” which model wrote it.
I don’t need to spend 4x the tokens just to make progress at the same speed I’m already going at.
Did you honestly think your customers would read “parallel agents” and think anything other than “4 different tasks running in parallel”? Hasn’t that been the ENTIRE point of multiple agents since their inception as a concept? Does any company hire multiple engineers to work on the same tickets in parallel just to see which engineer is better?
Come on…