Could Tyler Cowen be raising awareness of a failure of Information Revolution to trickle down to the common man, just as it begins to rain buckets.
For a long time a certain set of economists has griped about health care. One of my particular gripes has been the very concept of a physician in the modern era.
A physician is more or less a human database and an incredibly expensive one. With medical training we basically take some of the most valuable brains on earth and hack them into doing something a brain is not intended to do: store and retrieve massive amounts specific information on demand with very low error rates.
On the other hand we have this machine called a computer which does this effortlessly and at low cost. Bare minimum we should forget about cramming drug interactions into human minds and just let them look it up in a database.
But, why not take it a step further and just get rid of the human altogether? Literally tens of thousands of people die every year because doctors are, well, only human, and make diagnostic mistakes which can later be identified as violating evidence based medicine.
The problem I always saw was how to convince people that yes, a computer can beat House, MD. Well IBM can’t go head-to-head with House but maybe beating Ken Jennings and Brad Rutter might help.
Looks like that’s what they had in mind for Watson. From IBM
Beyond Jeopardy!, the technology behind Watson can be adapted to solve problems and drive progress in various fields. The computer has the ability to sift through vast amounts of data and return precise answers, ranking its confidence in its answers. The technology could be applied in areas such as healthcare, to help accurately diagnose patients, to improve online self-service help desks, to provide tourists and citizens with specific information regarding cities, prompt customer support via phone, and much more.
For now, I see via Dan Bowman, that IBM is saying very kind, political things like
“You’re never going to replace a trained doctor or nurse,” Dr. Joseph Jasinski, a member of IBM Research’s healthcare and life sciences team said in the video. “But certainly a system like Watson could be a physician’s assistant. It could help to check on things: ‘Did you consider this? Did you consider that?’”
This is how you always open. Oh don’t worry. You’re irreplaceable. We’re just here to help. I mean a robot can lift heavy things for you, but you’ll always need a trained autoworker.
15 comments
Comments feed for this article
Thursday ~ February 17th, 2011 at 6:57 pm
Hal
Um, dude. AI is the future and always will be. It’s one thing to replace the physical motions of a human being. I mean, replacing the physical capabilities of a human is a huge and enormously complicated thing to do, and we’ve made enormous strides in doing so over the years.
But seriously. When people say “this could be applied in areas such as…” you had better be reaching for your wallet, because they sure as hell are trying to empty it.
You should spend some time with real live programmers and computer engineers who actually do this stuff for a living. I’m sure it looks like there’s endless opportunity to get rid of those pesky humans and really start making things efficient when you look at this wonder from the outside.
But there’s really not that much “there” there yet.
Maybe there will be someday, but I do note that sweatshops filled with children and women are kicking ass on computerized assembly lines. Hello? Foxcon? Nike? You name it, it’s probably assembled by humans. So, there’s probably something that should give you at least some pause in your flights of fancy in this regard.
Thursday ~ February 17th, 2011 at 7:19 pm
Felix
Someone’s just gotta shop that picture to change the cars to the robot arms. Each one more scary looking as they come off the line. With a few of them clearly taking their places in the set of worker-bees on the sides.
Hal, same thing for computerization in general. Reach for that wallet. Back in the day, it was a standing joke that computer sales guys would talk about how you’d replace all those pesky, 8 hour employees with cheap, tireless 24/7 computers. The joke being that what always happened, at the very best, was you had the same number of pesky employees using the tireless computers to double the output.
The “children and women” [What, your model of the world is a sinking ship!?
] They’ll be gone in another decade or two. No more cheap ones left.
Thursday ~ February 17th, 2011 at 8:52 pm
Hal
No more cheap ones left.
Indeed.
Thursday ~ February 17th, 2011 at 9:12 pm
imaginary_nums
Sorry where was Malthus in this post?
Thursday ~ February 17th, 2011 at 10:24 pm
RickRussellTX
A cousin of mine has a baby at death’s door right now because she was put in the care of a surgeon who was either (1) significantly impaired by something or (2) operating on inadequately understood diagnostic information.
My wife was in a hospital for a week during her second pregnancy in 2008, during which time she had perhaps 30 minutes of doctor face-time, a situation that ended with us leaving the hospital against Dr’s orders and finding a new practitioner.
A friend of mine underwent long and painful treatments for knee pain based on what, he later learned, was a completely inaccurate basic understanding of the underlying problem.
So… would I sign up for a Watson diagnosis? Hell yes. If doctors want to keep pressing the system to create artificial scarcity, then they’re going to pay the price.
Thursday ~ February 17th, 2011 at 11:13 pm
Jeff
“With medical training we basically take some of the most valuable brains on earth and hack them into doing something a brain is not intended to do: store and retrieve massive amounts specific information on demand with very low error rates”
The human mind is actually exceptional at storing and retrieving massive amounts of information. What it’s not terribly good at is extended chains of reasoning or complex probabilistic inference.
As for Watson, it is basically an juiced-up expert system. Expert systems have been around for decades. They were quite the rage in the ’70′s and a number were developed to assist with medical diagnoses. Of course, you had to be skilled in using the thing–how to query the system, etc. Watson would make that aspect much simpler, but otherwise the important part is not fundamentally different.
Friday ~ February 18th, 2011 at 4:09 am
Miguel A.
Well, you’d better make a computer lawyer and judge if you “really” want to cut health care costs…
Friday ~ February 18th, 2011 at 7:22 am
mike
Toronto.
Friday ~ February 18th, 2011 at 8:57 am
NeuroLogica Blog » Dr. Watson
[...] does not imply that computers are ready to take over from humans what humans are good at. Here is an example of exactly that misguided interpretation. But, why not take it a step further and just get rid of the human altogether? Literally tens of [...]
Friday ~ February 18th, 2011 at 11:14 am
Econ Skeptic
Leaving aside whether this is really a practially possible thing to do right now, Watson-like reference system would not only reduce error rates based on better utilization of extant knowledge, but would also dramatically improve our knowledge base. You’d have constant large-scal human trials going on continuously as Watson’s accross the nation compare and contrast input parameters and outputs. I think that’s the idea behind computerizing recods in ACA, but computerizing records is a giant step behind actually using those records to learn something.
Friday ~ February 18th, 2011 at 12:46 pm
Hal
Well, as AI researchers have found, a reference system – even with a better UI like Watson’s – doesn’t mean much. We have zillions of such reference systems. And it’s not like Doctors, Nurses and other medical professionals don’t use them.
There’s probably a lot we can do to improve record keeping and cross referencing of drugs n’ such. But seriously, though. If we could just get Doctors to wash their hands and stop wearing long sleeve white coats (which Germs love to hit rides on between patients), the benefits would probably swamp any savings that would come from installing any Watson like system.
Heck, just eliminating the insurance companies as useless middlemen who are simply sucking up money would probably provide similar efficiency gains.
But that’d be socialism and therefore wrong.
Friday ~ February 18th, 2011 at 4:16 pm
Simon K
“We’re going to make your job easier. Don’t you want that?” Words to strike fear into the heart.
Saturday ~ February 19th, 2011 at 12:04 am
Cowen, Malthus and Watson | Brucetheeconomist's Blog
[...] http://modeledbehavior.com/2011/02/17/cowen-malthus-and-watson/ Could Tyler Cowen be raising awareness of a failure of Information Revolution to trickle down to the common man, just as it begins to rain buckets. [...]
Saturday ~ February 19th, 2011 at 1:19 pm
mad
Very much don’t talk to engineers about AI, and what is ‘there’ there. AI is the ghost in the machine; a seemingly ever receding goal.
But that’s not what we’re talking about here. Real computer driven progress is banal. The important part of sending a tweet is fundamentally no different then the pony express.
The banal parts of medicine are measurements, database lookups, and the correlation between the two. And it’s not about GP’s using Watson. That’s like going to the post office in the black berry age.
Dr. Watson of the future is about analyzing an HD picture of yourself in the privacy of your own room, along with your past daily recorded vitals, an online database of medical information, and your Facebook page, and delivering in 2.93 seconds with 95% confidence, the question, “What is Herpes”?
Wednesday ~ February 23rd, 2011 at 12:25 pm
Hal
A minor follow up from Technology Review. It’s optimistic as all forward looking tech posts are required by law to be, but it does flesh out a few of the rather high hurdles the tech has to overcome to be even moderately useful in such an environment.
I love how at the end, Watson is relegated to a call center.