Applying the lens to your own life
The same lens that reads economies reads a working life. You are an agent in the system, not a passenger.
Where a working life stalls
The German conversation about technology and work is dominated by loss: jobs that automation will take, skills that will expire. The fear is not irrational — transitions are real. The trouble is what the stance built on it produces: waiting for certainty, which cedes every early decision to whoever moved first. These traps are personal versions of the national ones.
"Technology is something that happens to me." Reading every headline as a threat assessment, treating change as weather. The stance feels like realism; functionally it outsources your agency to whoever moves first.
SOURCES · Bankenverband 2026One qualification, one employer, one ladder — the career as a plan drawn at twenty-five and executed for forty years. It was a good script for a slow system. In a fast one, a single track is a single point of failure.
SOURCES · Bitkom 2022Waiting for the employer to send you to a course, for the certificate to make you officially ready, for permission to try. Short training measures with limited effect are what you get when learning is something done to you.
SOURCES · KfW 2020What the lens says
Consider the decision the headlines keep forcing. An administrator reads that AI will automate office work, and has to do something with the information. One available strategy is forecasting: work out which jobs are safe, retrain toward the predicted winner, wait for certainty. Its weakness is the same one that defeats the ministries and corporations of the earlier chapters — the labor market is a complex system, and nobody can reliably say which skills will pay in ten years. A forecast about the average case also says little about the one career you control. The other strategy is to run experiments at your own scale. Suppose the administrator gives one evening, three hours, to making the new tool handle the week's routine correspondence. What comes back is specific: the standard letters are usable after light editing, while the case-file assessments it gets confidently wrong. Three hours have answered a question no forecast addresses — which half of this particular job is moving, this year. Publishing a small piece of work, or volunteering for the project that touches the new system, returns the same kind of firsthand answer at a slightly larger scale.
The costs of the two strategies are worth comparing directly. If the alarming headline proves overblown, the experimenter is out a few evenings. If it proves right, the experimenter already knows the tools while the forecasters are still choosing a course. The book's core evidence is historical — technological change has overwhelmingly raised living standards, broadly and across generations, even as it displaced particular tasks — and the link below carries that argument. But the case for experimenting does not need to lean on history. It rests on the asymmetry above: a cheap experiment buys information about your own situation that no forecast, however expert, can contain.
This is what "rational optimism" means here — a strategy, not a temperament: act early, in steps small enough to be cheap, on the working assumption that the change can be used, and let each step replace a secondhand forecast with firsthand information. What does that look like as an actual weekly practice?
2016 Why rational optimism, the evidence → 2026 Agents and incentives, hands-on →
What to do instead
Publish the work — a post, a talk, a repo, a prototype — and let strangers react. A career loop that runs through one annual review learns nothing; one that runs through the public learns weekly.
M4 Make failure cheap.Keep one live side project whose total possible loss is a few evenings a month. It is your personal laboratory: the place where being wrong is cheap enough to be useful.
M8 Expect emergence.Choose the option that opens options. Between two roles, pick the one whose skills recombine — the future will ask a question your plan did not anticipate, and optionality is how you answer it.
Retire these
Doom-scrolling as career research— an hour with the tool teaches more than a month of headlines about it.The ten-year career plan— plan the next experiment, hold the direction loosely.Waiting for the official certificate before starting— the portfolio is the certificate now; the work is the proof.