What if your real problem is that you have too much room to work with? David Epstein's new book, Inside the Box, says that's been the problem. We treat limits as the thing standing between us and our best work, but Epstein spends 300 pages showing it runs the other way as well: hand a team unlimited money and a blank page, and often what you get is bloat, not a breakthrough. Fifty words on a bet produced Green Eggs and Ham. The catch is that the constraint has to bind the problem and leave the solution alone. Look at what we did to the Chinese. We cut off the good chips, then banned the workaround chip Nvidia built to squeak under the rule, and DeepSeek answered by writing its own scheduling to route around bandwidth it didn't have, then putting out a frontier model for what it says was a rounding error, a number still under dispute. Nvidia still gave back roughly $600 billion of market value in one day. We handed them a constraint, and they built the workaround the well-fed labs had no reason to invent. Here are a few examples for the curious from the book. Block the familiar path and people invent inside the hour. An Olympic skeleton coach couldn't afford the biomechanics rigs the American team showed up with, so he told his athletes to do anything except what they'd always done, and two of them came back within the hour with the one-handed start, now standard in the sport. When something is broken, the instinct is to add, and the instinct is usually wrong. Leidy Klotz paid adults to fix a wobbly Lego tower and docked their pay for every piece they added; removing one bad piece solved it, and they added anyway. Capacity can be added anywhere, but the bottleneck still limits output. An NIH lab bought a sequencer that read DNA faster than anyone's, but discovery didn't move, because there weren't enough analysts, so the data just stacked up and sat. I've bought a sequencer more than once. Write the kill rule before you look at the data. When drug researchers had to start pre-registering their prediction and analysis plan in 2000, trial results flipped from mostly positive to mostly null. Same drugs. Less freedom. Make the trade-off visible and the argument settles itself. A director on The Incredibles spent months on background aquarium fish while main characters sat unanimated, so a producer put popsicle sticks on a wall, one per week of animator time, next to each character. Want more fish, go take a stick off somebody. A producer built that wall on purpose, because the fish would've eaten the whole schedule otherwise. Deadlines. Budget caps. Goals with a date attached. Guidelines you wrote instead of inherited. A constraint is coming either way. Build your own, and you know the reasoning behind every limit in it. Skip that, and somebody else's limit lands on you with no reasoning attached.