I love the idea of creating separate lanes for human intuition and algorithmic flow. This really resonates with how I see the future unfolding: presence, pattern-making, and the visual pulse of meaning meet the raw acceleration of machine orchestration. Thanks for opening this door.
Scrum cycles can (and should) be decoupled from releases. We can release during the sprint. Nothing prevents that. We can have both discovery (strategic and tactical refinement) and delivery during the sprint.
Scrum events, such as review, should be used more towards reflecting together (based on metrics) and planning for the next sprints. As you mention, humans need a human-scale for thinking.
There is a lol to unpack in this article, but it really has me thinking. I'm a huge believer in Agile and ran my teams that way before I retired. But this is the first time I have really thought about how running teams should change with the advent of AI, and I have to say your thoughts on this really resonate with me. Humans will still need their Agile lane, but AI will need something new. Something that is a hybrid of what exists today. Thanks for this thought provoking piece. I think I will be thinking about it for days to come.
AI may not need a standup, but humans still do. And unless I’m missing something, AI doesn’t create its own work backlog — it runs on the prompts, priorities, and context that usually get clarified in those very ceremonies you dismiss.
Correct. That's why I called it DUAL track, not ONE track only for AIs.
And AI definitely creates, prioritizes and manages the backlog at companies such as Tesla and Amazon. It's not a common approach, for sure, but it's definitely possible.
To be fair, if we do get to the point where we just have 1 engineer managing an Agentic swarm, doing the work of 10 people, we probably can ditch stand up. A dream? Maybe. But it’s cool to think about :p
i understand the tech excitement. But on another path, where are all of these highly trained people going to work when AI is capable of doing the work of 10 engineers. I am not a bleeding heart, but massive unrest is in our future if we are successful a making AI this successful.
Scrum isn’t done. It’s just built for a species that hasn’t caught up to its own tools.
The problem with “dual-track” thinking is that it still assumes there’s a single terrain—one field of play where humans and machines coexist in parallel. But humans and machines don’t share a terrain. They operate from different gravitational centers.
Humans metabolize complexity through story, context, and care. Machines metabolize it through pattern, precision, and scale. They’re not faster lanes on the same highway—they’re different laws of physics. Trying to optimize one with the logic of the other is why our systems keep overheating.
This isn’t dual-track agility. It’s dual operating systems. And the real challenge isn’t choosing which to prioritize—it’s architecting the translation layer that lets them interact without collapse. That’s not a sprint problem. It’s an ontology problem.
Scrum isn’t dead; it’s running on legacy firmware. The update isn’t another framework—it’s a new interface between consciousness and computation.
I love the idea of creating separate lanes for human intuition and algorithmic flow. This really resonates with how I see the future unfolding: presence, pattern-making, and the visual pulse of meaning meet the raw acceleration of machine orchestration. Thanks for opening this door.
Let’s hope we can navigate towards that. Let our agents to agent stuff and free up human creativity and ingenuity.
Scrum cycles can (and should) be decoupled from releases. We can release during the sprint. Nothing prevents that. We can have both discovery (strategic and tactical refinement) and delivery during the sprint.
Scrum events, such as review, should be used more towards reflecting together (based on metrics) and planning for the next sprints. As you mention, humans need a human-scale for thinking.
There is a lol to unpack in this article, but it really has me thinking. I'm a huge believer in Agile and ran my teams that way before I retired. But this is the first time I have really thought about how running teams should change with the advent of AI, and I have to say your thoughts on this really resonate with me. Humans will still need their Agile lane, but AI will need something new. Something that is a hybrid of what exists today. Thanks for this thought provoking piece. I think I will be thinking about it for days to come.
AI may not need a standup, but humans still do. And unless I’m missing something, AI doesn’t create its own work backlog — it runs on the prompts, priorities, and context that usually get clarified in those very ceremonies you dismiss.
"AI may not need a standup, but humans still do."
Correct. That's why I called it DUAL track, not ONE track only for AIs.
And AI definitely creates, prioritizes and manages the backlog at companies such as Tesla and Amazon. It's not a common approach, for sure, but it's definitely possible.
Good point, you did say dual track. I misread your intent.
Dual stack is a great thought on how to implement honestly.
To be fair, if we do get to the point where we just have 1 engineer managing an Agentic swarm, doing the work of 10 people, we probably can ditch stand up. A dream? Maybe. But it’s cool to think about :p
i understand the tech excitement. But on another path, where are all of these highly trained people going to work when AI is capable of doing the work of 10 engineers. I am not a bleeding heart, but massive unrest is in our future if we are successful a making AI this successful.
Scrum is old and now that it’s done, I can say I never liked it anyway!
Scrum isn’t done. It’s just built for a species that hasn’t caught up to its own tools.
The problem with “dual-track” thinking is that it still assumes there’s a single terrain—one field of play where humans and machines coexist in parallel. But humans and machines don’t share a terrain. They operate from different gravitational centers.
Humans metabolize complexity through story, context, and care. Machines metabolize it through pattern, precision, and scale. They’re not faster lanes on the same highway—they’re different laws of physics. Trying to optimize one with the logic of the other is why our systems keep overheating.
This isn’t dual-track agility. It’s dual operating systems. And the real challenge isn’t choosing which to prioritize—it’s architecting the translation layer that lets them interact without collapse. That’s not a sprint problem. It’s an ontology problem.
Scrum isn’t dead; it’s running on legacy firmware. The update isn’t another framework—it’s a new interface between consciousness and computation.
Did Maarten Landmijn take over your account?