The safest and most effective software release is not a major event but a boring, routine process that happens without fear or ceremony.
In the early stages, saying yes to every feature request creates a false sense of progress while building up unmanageable complexity and technical debt.
When detailed plans fail, adding more planning and meetings doesn't solve the problem and instead leads to a cycle of slipped deadlines, overtime, and errors.
Reducing release frequency from daily to every six months doesn't reduce risk but instead concentrates it into a single, high-stakes event with a larger blast radius.
Shifting to continuous deployment requires building trust by giving developers and QA clear responsibility, supported by a safety net of automated tests and controlled rollouts.
A shared test environment creates a bottleneck that blocks hotfixes, which can be solved by using independent, on-demand feature branch environments for every change.
Value stream mapping reveals that most time is spent waiting for work to start, not on the development itself, shifting the focus of optimization.
Fast technical deployments can create "continuous surprise" for the business, which is solved by decoupling deployment from release using feature flags and gradual rollouts.
A lean delivery process with automated, temporary environments reduces wasted work and infrastructure costs, leading to significant CO2 savings and a positive green IT impact.
The transformation from slow, risky releases to frequent, boring ones involved shifting mindsets, building a safety net, and optimizing the entire value stream.
Jobs that call for the skills explored in this talk.