The Amplitude Problem — Summary

Priya keeps a small cactus on her desk that she has not watered in three weeks. It seems fine. She cannot tell whether this means cacti are resilient or whether this one is dying in a way she has not learned to read. She has approximately the same relationship with the forty-seven papers open in her browser tabs, each one about maternal health interventions in districts like hers.

She is not struggling with misinformation. Misinformation has a shape you can learn to recognize: missing sources, implausible claims, obvious bias. What fills her screen is worse. It is information. Properly sourced, internally consistent, published in places that used to mean something. Forty-seven papers, and she cannot find the signal.

There are approximately 1.2 million recipes for chicken biryani on the English-language internet. Nobody finds this distressing, because the feedback loop is built into the domain. You cook it. You taste it. You know. This is what filtering looks like when it works: imperfect mechanisms adequate to a domain where reality provides feedback on a human timescale.

Most of what we learned about managing information abundance was learned in domains like this. Search engines, review systems, social networks: all biryani-tier solutions that assume someone in the chain touches reality and reports back. That assumption held for a long time. It is breaking now, and the break is not where most people are looking.

Before AI writing tools, the global corpus of serious-looking research was constrained by production cost. Writing a paper was hard. Getting it reviewed was slow. The friction was not a good filter or a fair one, but it kept the ratio of signal to noise within the range of human cognitive capacity. That world is over. Not because AI introduced falsehood, but because it removed the production cost that kept volume proportional to effort. The floor rose. Every paper now looks competent. Every abstract is well-structured. The markers that used to correlate, however imperfectly, with someone having actually done the work no longer correlate with anything at all.

The amplitude problem breaks differently across three tiers. In the kitchen tier, where feedback is personal and fast, existing filters mostly hold. In the clinic tier, where reality provides feedback slowly, distantly, and through institutions, the old heuristics have been rendered inoperative without replacement. This is Priya’s tier. In the abstract tier, where there is no ground truth and the membrane was always the scarcity of qualified evaluators, volume overwhelms the membrane without penetrating it. The signal is still there. The people who could identify it no longer have enough hours in their careers to find it.

The word matters. This is not a volume problem but an amplitude problem. AI amplifies signal and noise with equal fidelity. In the old world, the difference between knowledge and belief was partially legible in the effort required to express each one. Writing carried traces of contact with reality. AI erases those traces by making the surface quality of all writing converge. When everything reads like it was written by someone who knows what they are talking about, the reader’s ability to distinguish expertise from fluency collapses.

What we have not lost is truth. The papers describing interventions that actually work still exist. What we have lost is the set of ambient, imperfect, socially constructed mechanisms by which a thoughtful person could find truth without already knowing it. The heuristics have not been disproven. They have been rendered inoperative by a change in the production environment.

What Priya needs is not a better search engine or a smarter ranking algorithm. Not an AI that summarizes papers for her, because a summary of noise is still noise, just shorter. She needs a way to ask a different question: not which of these is right, but which of these was here. Which one carries the traces of someone who stood in a clinic like hers, counted patients like hers, watched what happened over months in conditions she would recognize.

She picks up the small plastic watering can she keeps behind her monitor and gives the cactus a little water. Not much. Overwatering, she has read, kills more cacti than neglect.

Somewhere in her tabs, there is probably a paper about that too.