for fifty years, an attempt was expensive. you picked one aim, built one apparatus, and held coherence for years before nature told you anything. under that regime, concentrating before measuring made sense. we grew up inside that regime. it ended quietly, and most people haven't noticed yet.

the energy per emission keeps falling. it costs us almost nothing now to fire a quantum. so we stopped pointing a telescope at one spot in the sky, and built a source that fires continuously with a detector that reads fast. that's the whole redesign.

outcomes are not normally distributed. a few modes carry almost all the energy, and we have given up trying to compute which ones in advance. it isn't a flaw in our calculators. it's the field. sample, measure, re-aim. that's all anyone can do, whatever they claim.

most laboratories react by aiming more carefully. we tried that. an aimed shot and a sampled shot spend the same energy here, and only the sampled one teaches you the distribution. so now we just fire more often.

the doctrine, then: emit on schedule. observe early. discard without ceremony. a discarded state isn't a loss, it's a datum. we'd like to tell you the short half-life doesn't hurt. it does, a little, every time. we grieve fast and keep the schedule.

and when a mode resonates, we stop sampling and pump. a laser is just the discipline to feed one coherent mode all the energy the other modes wanted. amplification is a decision. we have promised ourselves we will be brave enough to make it.

the system is closed. no external energy enters. this is a choice, and some days it is an uncomfortable one. but open systems can grow for years without learning what actually sustains them. a closed system finds out immediately. we would rather know.

we sit low on the energy gradient, on purpose. the same joule drives more emissions at our coordinates than almost anywhere on the map. people read this as a limitation. we read it as the reason we can afford to be wrong more often than anyone else.

bound states are the only ones we count. what returns is signal. what scatters is noise, however loud the collision was. it took us longer than we would like to admit to stop measuring loudness.

quanta come and go. we have made peace with that. the apparatus is what accretes: a sharper detector, a faster gate, a better prior after every decay. some cycles, the only thing we keep is the lesson. after enough of them, the source itself is the result.

t₀ is fixed. the horizon is long. we are not trying to predict the field. we are building the instrument that samples it, and we intend to still be here calibrating it in five years.