6 ideas
Companies that relentlessly optimize for efficiency strip out the slack, redundancy, and apparent waste that actually enables them to adapt to surprise. The leaner and more predictable a system becomes, the more fragile it is when conditions shift, because there is no spare capacity to absorb shocks or experiment.
Corporations stage the visible rituals of innovation — labs, hackathons, accelerators, chief innovation officers — while structurally avoiding the actual risk and uncertainty that real innovation requires. These performances generate the feeling of progress and protect careers without changing what the company actually ships.
The purpose of an experiment is not to succeed but to convert money into information you didn't previously have. Judged this way, a failed experiment that produces a clear lesson is more valuable than a safe project that confirms what everyone already believed.
Decisions should be sorted by whether they can be cheaply undone: reversible bets should be made fast and frequently to maximize learning, while irreversible ones warrant deliberation. Treating all decisions as if they were irreversible imposes the slowness of the rare high-stakes choice onto the many cheap ones, crushing the rate of experimentation.
Large firms increasingly outsource real invention to startups and then acquire them, because their internal incentives systematically punish the uncertainty and failure rate that genuine breakthroughs require. This means R&D spending inside incumbents is often misdirected toward incremental safety rather than the discontinuous bets that create new value.
Every failure a company prevents through control and process has a hidden price: the learning, capability, and option value that failure would have generated. Organizations track and minimize visible failures while remaining blind to the compounding cost of the experiments they never ran.