javadocmd

You've not refuted the article so much as pointed out a corner case the author didn't address in which ML is a good fit. Your example, using ML to perform the medical coding function, is using a data source (in this case the EMR) for one of the purposes for which it was explicitly designed and for which it is (arguably) non-deficient. That is a realm not doomed to failure.

The realm doomed to failure is using a data source for a completely oblique purpose for which it is horribly distorted. Namely, the purpose of optimizing individual and public health by discovering guidelines and treatments, diagnosing illness, and delivering optimal care.

(Of course medical billing as an enterprise shouldn't even exist, but that is another topic.)

A bit like Captain Renault proclaiming his shock at finding gambling in Rick's cafe.

If we go around breaking windows we could really bolster the world's window-making capacity.

I wish the first thread I saw was this enthusiastic about Brother. I've got an HP delivering today because it sounded like the Linux driver support was more solid -- didn't see a mention of this proprietary cartridges nonsense. Mea culpa for not doing more research I guess.

You invented "Chutes and Ladders: Legacy".

If the hot startup trend suddenly became lean hospital systems, you'd see all sorts of articles about the ideal attending-to-resident physician ratio.

Worth noting that the author makes a significant claim without any citation:

> That vaping works as a way to quit smoking... seems clear.

Studies have been fairly limited so far, given the recency of e-cigarettes (relative to science's timescale). And the studies that we have seem pretty mixed: some showing benefit over other accepted smoking cessation methods, some not. Some showing higher relapse rates in the e-cigarette group. Nothing about that seems clear to me.

What they didn't report is the corresponding decrease in labor costs. If they save 1M in labor to lose 850K, they're still ahead.

> You know someone somewhere got a can of creamed corn.

But the point was you don't know that. If they get deluged with creamed corn, it might just sit in a warehouse, get shipped to a different community, or even get trashed. Donate a can if you prefer, just don't fool yourself with false guarantees.

If this had launched with Slack support two months ago, I'd have saved a few days development time. Interested to see the pace of partner adoption.

I would argue that even the ideal medical system would likely produce the same outcomes experienced by the author of the parent post. In our present stage of scientific and technological advancement, the following might as well be a Law of Nature: even the best of drugs will have varying effects on various people in varying circumstances.

But yes, there are plenty of inadequacies in our healthcare industry (speaking from the US), but that's a much, much broader discussion.

How can we do better?