Families

19 problem families — different problems, same issues

A family is a set of problems from unrelated fields that get stuck the same way, such as a sensor failing in a pasture and one failing on a buoy. 385 of the 613 problems in the collection belong to a family so far. The rest stand alone for now. Each family is drawn with its own marks, and related families share marks.

it worked in the lab92 problems · 17 fieldsYour sensor works in the lab. Has anyone tested it where it has to work: in seawater, inside a patient, on a factory floor, or in a field with no electricity?drawn as the divide the problem lives in the rulebook was written for a world that no longer exists38 problems · 15 fieldsThe regulations were written for a world that no longer exists. In some cases the technology succeeded so well that its success created a problem nobody had written rules for.drawn as the thing that's missing or being lost the context you didn’t design for31 problems · 12 fieldsThe intervention was designed for a world with reliable electricity, trained technicians, working supply chains, and cooperative institutions. It was then used in a world that has none of these.drawn as two readings that don't agree the solution exists but nobody can afford it30 problems · 15 fieldsThe technology works and the economics don't. 'Too expensive' means eight different things, so find out which kind you face before choosing a strategy.drawn as stretched thin — waits, shortages, thin margins making one is easy. making a million is the problem30 problems · 15 fieldsIt works at lab scale. At factory scale living systems adapt, materials behave differently, and precision achieved by hand can't be mechanized.drawn as the divide the problem lives in · stretched thin — waits, shortages, thin margins the model was never trained on this29 problems · 15 fieldsThe model was trained on data that doesn't represent the conditions where it matters most, and nobody checked whose experience was missing.drawn as the thing that's missing or being lost designed for a world that doesn’t exist here24 problems · 13 fieldsMaking the technology free wouldn't solve these problems. The barriers are harder to see: no electricity, no trained technicians, no supply chain, and designers who never visited the place.drawn as stretched thin — waits, shortages, thin margins · the divide the problem lives in the chemistry itself changes at scale24 problems · 6 fields · newly namedThe chemistry works in the flask. At production scale heat and material move differently, catalysts fail by different mechanisms, and selectivity is lost.drawn as the divide the problem lives in the regulator demands evidence that cannot exist21 problems · 10 fieldsA regulatory framework exists, and it demands evidence in a form that doesn't match how the technology behaves.drawn as the rulebook disagrees with itself the answer is in a field you’ve never heard of20 problems · 10 fieldsThe pieces of the solution exist in different disciplines that haven't learned to talk to each other. Funding inside one field will not solve a problem that sits between fields.drawn as things kept apart — silos, systems that won't talk the data exists but can’t talk to itself19 problems · 11 fieldsThe data exists, and the systems that hold it can't read each other. No one has a reason to build the translator, because the cost is shared and the benefit goes to someone else.drawn as things kept apart — silos, systems that won't talk the wrong ruler18 problems · 16 fields · newly namedFind the number the institution acts on, and the outcome it is supposed to protect. How far apart are they, and who benefits from the gap?drawn as the rulebook disagrees with itself the missing yardstick17 problems · 14 fields · newly namedIf a technical committee published an agreed method tomorrow, the field would be unblocked. Which committee would that be?drawn as the thing that's missing or being lost designed for the wrong person16 problems · 7 fieldsThe intervention failed because it was designed to change the wrong person: fixing students when the institution was the problem, training patients when the device needed redesign, imposing solutions without understanding the setting.drawn as two readings that don't agree the infrastructure that generates bad science13 problems · 8 fields · newly namedThe infrastructure that produces knowledge has a quality problem that nobody is rewarded for fixing, and mandates have not fixed it. How many research conclusions rest on inputs that were never validated?drawn as two readings that don't agree built to be open. now under attack12 problems · 7 fieldsYou can add monitoring, threat detection, and encryption. The underlying system was designed before cybersecurity was a concern, and it can't be replaced without shutting everything down.drawn as the thing that's missing or being lost success's shadow11 problems · 14 fields · newly namedPick something that worked. What did the system stop watching on the day it started working?drawn as two readings that don't agree the theory hasn’t been invented yet10 problems · 10 fieldsThe theory needed to solve these problems hasn't been invented yet, and inventing it requires contributions from fields that don't share a language.drawn as things kept apart — silos, systems that won't talk · the thing that's missing or being lost we can see the danger but the system won’t act9 problems · 9 fieldsThe sensor works, the alarm fires, and the danger is real. No regulation requires anyone to act, so the signal is ignored until the disaster happens.drawn as the divide the problem lives in

How the families grew

Each bar is one family across the collection’s releases, oldest on the left. The newest families begin part-way along. Sizes are printed at the end of each row.

it worked in the lab

the rulebook was written for a world that no longer exists

the context you didn’t design for

the solution exists but nobody can afford it

making one is easy. making a million is the problem

the model was never trained on this

designed for a world that doesn’t exist here

the chemistry itself changes at scale

the regulator demands evidence that cannot exist

the answer is in a field you’ve never heard of

the data exists but can’t talk to itself

the wrong ruler