Moxie Marlinspike: Agile is killing software innovation

Agile’s Impact on Innovation and Software Quality

• Many see “Big-A Agile” (Scrum, points, ceremonies) as turning engineering into an assembly line where devs just close tickets, reducing creativity and holistic product thinking.
• Some link the rise of Agile with buggier, slower-feeling software, though others argue overall responsiveness has improved with better hardware and practices.
• A contrary view is that innovation feels slower mainly because the “easy unknowns” are exhausted; software is maturing, not being “killed.”

Work Breakdown, Knowledge Loss, and Code Ownership

• Breaking work into tiny tasks can strip context: the planner gains deep understanding, but implementers receive only fragments and must rediscover design “theory.”
• Several argue for unit or module “ownership” so one engineer or small team maintains the mental model; random ticket assignment and constant context switching are seen as highly counterproductive.

Management, Corporations, and Incentives

• A recurring theme: companies optimize for short-term profit and speed, not quality, leading to “good enough” hiring and process-heavy control.
• Developer productivity is seen as nonlinear; interruptions and ceremonies destroy flow, which may explain perceptions of “10x engineers.”
• Some argue corporations and power-seeking managers, not Agile itself, are the root cause; Agile becomes just another tool to centralize control and micromanage.

Agile vs Scrum and “True Agile”

• Multiple comments distinguish the Agile Manifesto (lightweight, people-focused, adaptable) from heavyweight Scrum/OKR/process cultures.
• There is pushback against the “no true Agile” defense: critics note that the dysfunctional version is what most people actually experience.
• Certifications and rigid frameworks are frequently described as cargo cults, even “pyramid schemes,” contrary to the manifesto’s spirit.

Estimation, Metrics, and Taylorism

• Story points, planning poker, and forced estimates on unknown work are widely mocked as performative, feeding dashboards rather than decisions.
• Some see this as a modern form of Taylorism: pseudo-scientific measurement that doesn’t fit complex, creative work.
• A minority defends managers’ desire for estimates as reasonable, but others argue that deep uncertainty makes precise numbers dishonest or useless.