The Technology Was Never the Problem: Forty years inside medical technology, and the machine that finally learned to ask

The best clinical knowledge in America is free to anyone who asks for it. What has been missing is something that can reason over it. That now exists, and I believe it is a real chance to change how medicine works.

I spent forty years finding that out the slow way.

I built brain monitors, anesthesia records, and the software that made infusion pumps talk to pharmacy systems. Every one of those machines worked. Almost none of them changed what happened to patients, and it took me until I retired to understand why. The connecting was only ever the plumbing. The data was always the asset. And every machine I built could calculate perfectly and understand nothing, so the entire burden of meaning sat on whoever happened to be standing in front of it.

All that time I was paying a private trade association, per machine, for permission to use codes the federal government requires a hospital to use in order to be paid. And in the same decade I was using a national medical vocabulary, built by the National Library of Medicine and given away for nothing, without once noticing what I had.

The data source is free. The access to leverage it is not.

That is the argument. This is the part I did not expect to be writing.

I did not stop at believing it. I built a working demonstration. It reasons over a simulated patient, in a browser, and it is running now. Chapter 22 is my own list of everything still wrong with it, because a claim this size is worth nothing without one.

The United States spends more on health care than any comparable country and ranks last among them for performance. I am not certain this fixes that. I am certain the opportunity is real, that it is sitting in front of us, and that the thing most likely to stop it is not the technology.

This is not a great man's life. It is an ordinary engineer's account of how he got there, and what he built when he arrived.