Software engineers used to be the gatekeepers of code. That gate is about to swing open, and no one is sure who will be left to check the IDs. Mo Plassnig, the new CEO of CloudBees, spent his first weeks on the job laying out a future where human review of code becomes the exception, not the rule.
He took over earlier this month from Anuj Kapur. Plassnig is a founder of Codeship, the continuous delivery platform CloudBees bought in 2018.
Before that, he was chief product officer at Immuta, a data security firm. The core problem is simple: machines now write code faster than people can read it. Plassnig said the ability to write code is no longer a bottleneck.
End users will soon type what they want — their “intent” — and the system will generate not just the code, but the entire pull request. No human typing.
No human debugging. Just a prompt and a result. But the current workflow still depends on software engineers to review that code.
That worked when humans wrote it. It breaks when AI generates it in bulk. Plassnig is blunt about the math: the pace of code creation is overwhelming.
It will not be feasible to keep reviewing everything manually. He is not calling for a sudden overhaul.
“It will not be feasible to reinvent software engineering workflows overnight,” he said. Instead, the shift will happen over time. Existing tools, platforms, and processes will be re-engineered piece by piece.
Teams will move toward what he calls “agentic engineering workflows.” What does that look like? AI agents will review, observe, and deploy code that other AI agents generated.
The human role shifts from writing and checking code to defining the rules the agents follow. The gatekeeper becomes the rule-maker. This is not a distant theory.
The report notes that DevOps workflows are already being reshaped to handle code from both professional developers and so-called citizen developers — people who can describe what they want but cannot write the code themselves. That population is exploding. Plassnig’s background suggests he understands the pipeline problem.
He helped build Codeship, a platform that automated the integration and delivery of code. That was about moving code from a developer’s laptop to a server faster.
This is about moving code from an AI’s output to production with less human friction. The report does not say how long the transition will take. It does not name any specific customers or timelines.
What it does say is that the shift is coming and that the old way of doing things — human-centric tooling — will not survive the agentic AI era. For now, software engineers still hold the keys.
Plassnig’s point is that they should start thinking about what happens when the locks change.
