Skip to content

Search problems, genes, families, sources

Find a problem Families All problems Notation & genes Data & method About

Loading…

A living collection from Natural Artificial Labs at Dartmouth. Problems come from expert sources — research agendas, standards bodies, agencies — and every source is fetched and checked before a problem enters; independent readers re-tag samples blind so the genes can be trusted.

How verification works Data & method Bring us a problem Teach with the collection

Genome › Failure mode

Failure mode · gene 13 of 14

17 problems carry solved the wrong problem

failure:wrong-problem

⚠ independent readers often disagree on this tag — it passes an extra decision gate before use

The rarest failure gene and the trickiest to spot: the goal itself was wrong, so even flawless execution couldn't help. Programs that maximized how many devices got distributed, when the binding need was teaching anyone to use them. Redesigning the syringe when the real need was knowing whether the vaccine inside had frozen. Everything downstream of the objective was competent — the objective was the mistake.

Related genes

aimed at the wrong personwrong-stakeholderthe solution targeted a user or decision-maker who isn't the one who actually determines the outcome ignored real-world conditionsignored-contextthe solution assumed conditions — power, weather, skills, institutions — that the real setting doesn't provide measured a stand-inproxy-metricoptimized or certified against a stand-in that doesn't track the outcome it's supposed to protect · added August 2026 — under review for its first hundred uses the success created the problemsuccess-causedan earlier intervention achieved its goal, and that very success — by a specific mechanism — produced today's problem

How to recognize it

State the objective past efforts pursued and ask whether achieving it perfectly would have solved anything. Rule out the cheaper explanations first: right goal but wrong audience is 'aimed at the wrong person'; right goal but unexamined conditions is 'ignored real-world conditions'; right goal but wrong measuring stick is 'measured a stand-in'. Only when the objective itself was mis-specified does this apply.

Independent readers re-tag samples of the collection blind so tags like this one can be trusted — plain-words reliability numbers live on data & method.

17 problems · 9 fields — filter them, then share the link.

field: society & learninga laptop is not a lessonmeasured the wrong thinghigh field: society & learningshelters designed for winter, deployed in the desertmeasured the wrong thingmedium field: health & medicinemalnutrition measured by the wrong bodymeasured the wrong thinghigh field: the built worldrain gardens with no scorecardmeasured the wrong thinghigh field: society & learningsolar panels with no one to fix themmeasured the wrong thingmedium field: health & medicinethe gold standard is guessingmeasured the wrong thingmedium field: water & energythe heatwave takes the grid downmeasured the wrong thingmedium field: health & medicinediagnosed decades too latemeasured the wrong thinghigh field: health & medicinethe virus clears before the patient arrivesmeasured the wrong thinghigh field: food & the living worldthe toxin in three-quarters of african maizemeasured the wrong thinghigh field: food & the living worldthe cure that clears the waymeasured the wrong thinghigh field: health & medicinestockpiled ventilators that can't ventilatemeasured the wrong thinghigh field: digital & spaceno instrument knows what alien life looks likemeasured the wrong thinghigh field: society & learningthe tent becomes an oven by noonmeasured the wrong thinghigh field: food & the living worldcenturies of breeding above groundmeasured the wrong thinghigh field: food & the living worldbrains rewired by a changed worldmeasured the wrong thinghigh field: food & the living worldfour billion years of r&d, going extinctmeasured the wrong thinghigh