The future is now - a new era of technology

[skepticalbip]

Even Star Trek's Enterprize needs McCoy between the technology and patient.

A problem Star Trek's Voyager solved by having a holographic AI be the doctor.

You are, of course, right. But then by the time of the Voyager series AI had advanced so far that the holographic doctor was orders of magnitude smarter and more insightful (maybe even more caring) than any human could possibly hope to be.

 

[dystopian]

You are, of course, right. But then by the time of the Voyager series AI had advanced so far that the holographic doctor was orders of magnitude smarter and more insightful (maybe even more caring) than any human could possibly hope to be.

holographic star trek doctor is best star trek doctor

https://www.youtube.com/watch?v=16LaAYMaLYI

 

[Togo]

False. Hook the computer up to a camera and sufficiently advanced software and it could. Or plug the patient into a future-tech full body medical scanner, probably letting it not only identify the source of the pain, but anything else wrong with you that you might not even be aware of. There is absolutely no fundamental reason why technology couldn't become advanced enough to completely replace doctors...

Yeah there is. Programming is static. Doctors are updated constantly, in every interaction. If you want your program to be as up to date as a doctor, then you need to update as often as a doctor, which means keeping up the same social and professional interactions as a doctor. Given that you're having to do all that, what exactly is the advantage the computer has, that a doctor using a computer index doesn't?

 

[dystopian]

Yeah there is. Programming is static. Doctors are updated constantly, in every interaction. If you want your program to be as up to date as a doctor, then you need to update as often as a doctor, which means keeping up the same social and professional interactions as a doctor. Given that you're having to do all that, what exactly is the advantage the computer has, that a doctor using a computer index doesn't?

Uhm, this is a pretty bizarre line of reasoning, I'm sorry.

First of all, programming is actually; depending on the type of software; updated far more often than a doctor's knowledge or skill. A doctor's knowledge is *not* "updated constantly, in every interaction". A doctor might receive new information about a patient that helps him make a more accurate diagnosis, but this is *not* equivalent to a software update: the underlying algorithms/database with which the doctor makes the diagnosis (his medical knowledge), remains the same. One can argue that repeated experience hones the doctor's skill and that this is a constant process; but the 'constant' part of said argument is questionable, and there's only so much that experience itself can do. Software, on the other hand, could (and often is) be updated on a daily basis, with these updates bringing improvements leading to increased efficiency, new knowledge for its database, and so on; whereas for the doctor there is a point of diminishing return: the experience he gains over time will at some point stop leading to increased efficiency as age and lack of updated knowledge become more of a detriment. Computers do not suffer from this problem, at least not as compared to humans.

Secondly, it is simply false to say that programming is static; even if we take human-written updates out of the equation. Self-modifying code is *not* science-fiction, and comes in a wide range of forms. We've had computers capable of optimizing their own programming since well before I was born, and this technology is constantly getting better and more refined.

Finally, why on earth would you think that for a program to be as up to date as a doctor, it needs to keep up the same social and professional interactions as a doctor? That would be horribly inefficient. It would just need to download the latest version of the central medical database, and it would instantly be up to do date, and be so far beyond what a single human doctor is capable of. Social interaction is not required. Social interaction with the patient would be fairly minimal, as scanning technology sufficiently advanced could bypass entirely the need for a patient to explain what symptoms they're feeling; however, assuming the scanning technology doesn't get to that point, the computer still has a distinct advantage over the human doctor: a human doctor can only hold one conversation at a time, and has imperfect memory of past conversations. A computer has no theoretical limit to the number of social interactions it can have at a time, and it has perfect recall over past such interactions which can be studied and included in its diagnostic routines.

So, no. You have not identified a fundamental reason why technology can't become advanced enough to completely replace doctors. Nor some fundamental aspect that makes human doctors just as good as a theoretical future medical machine.

 

[rousseau]

I remember a few years ago, a doctor I visited prescribed me a medication that I didn't need, which would have cost me *200 dollars per month* and would have led to a fierce dependency that I never would have been able to kick. Sure, there are doctors that are good, but there are also *many more doctors doing a bad job than you think*. Even if you had a doctor using a computer as a tool, as long as the algorithms, information, and probabilities were somewhat correct, that never would have happened to me.

The obvious problem with automation in this way is that there would be such a huge number of cases to account for that building a program that.. worked, would be hard.

 

[dystopian]

The obvious problem with automation in this way is that there would be such a huge number of cases to account for that building a program that.. worked, would be hard.

I would rather think the obvious to be true. The more cases there are, the easier it is to build a system that can make sense of the patterns and work from that to deliver the optimum solution in individual cases. Figuring out the patterns and solutions from a limited sample set would strike me as being more difficult for a program.

 

[credoconsolans]

False. Hook the computer up to a camera and sufficiently advanced software and it could. Or plug the patient into a future-tech full body medical scanner, probably letting it not only identify the source of the pain, but anything else wrong with you that you might not even be aware of. There is absolutely no fundamental reason why technology couldn't become advanced enough to completely replace doctors...

Yeah there is. Programming is static. Doctors are updated constantly, in every interaction. If you want your program to be as up to date as a doctor, then you need to update as often as a doctor, which means keeping up the same social and professional interactions as a doctor. Given that you're having to do all that, what exactly is the advantage the computer has, that a doctor using a computer index doesn't?

Programming would have to be updated because yes, depending upon where they work, doctors are constantly being subjected to continuing education. They are required to attend so many seminars, they are required to join medical organizations which send them new information and data, pharmaceutical companies constantly send them updates on medications and newest drugs, if they join other groups of doctors in their specialties, these other doctors go to them for consults on puzzling cases asking for advice, so doctors aren't static. Good ones aren't anyway.

 

[rousseau]

The obvious problem with automation in this way is that there would be such a huge number of cases to account for that building a program that.. worked, would be hard.

I would rather think the obvious to be true. The more cases there are, the easier it is to build a system that can make sense of the patterns and work from that to deliver the optimum solution in individual cases. Figuring out the patterns and solutions from a limited sample set would strike me as being more difficult for a program.

Yea, I mean, once you implemented the program it would work well, but actually getting a working program, as in finished product, built would be difficult because there would be an enormous amount of information to consolidate. It could be done, and would be effective, but it would be hard to make. Hell, if I had enough information I could probably build it myself, but it would take a very, very long time.

A program of this type would ideally be AI like, which is hard to do. Otherwise you'd be defining paths to every possible solution to every problem.

 

[skepticalbip]

I remember a few years ago, a doctor I visited prescribed me a medication that I didn't need, which would have cost me *200 dollars per month* and would have led to a fierce dependency that I never would have been able to kick. Sure, there are doctors that are good, but there are also *many more doctors doing a bad job than you think*. Even if you had a doctor using a computer as a tool, as long as the algorithms, information, and probabilities were somewhat correct, that never would have happened to me.

The obvious problem with automation in this way is that there would be such a huge number of cases to account for that building a program that.. worked, would be hard.

Yes. For the foreseeable future, the best system would be a large medical data base used to assist the doctor in deciding on a diagnosis and course of treatment. For some reason, I have more faith in a trained professional in the field than a logic tree written by a computer programmer.

I could see a frantic mother rushing her five year old son into an emergency room and presenting the symptoms to a robo-doc. The little boy is unsteady, has difficulty balancing and focusing and has a pronounced slur when trying to talk. General medical triage procedure is to address the most serious case first. The robo-doc sees that as a possible aneurysm so the boy is rushed off for an MRI or Catscan (or whatever they use to look for such things. I know diddly-squat about medicine).

Same case but the frantic mother meets a human doctor in the emergency room and presents the same symptoms. The doctor looks at the boys eyes and into his mouth (I don't know why doctors always seem to do this but the do) and notices the smell of Bourbon on the boy's breath. The doctor recommends a breathalyzer test or a blood sample for blood alcohol level. A much cheaper test to determine the problem for sure.

Until some far future time when technology gives us a system that is more versatile, insightful, and creative than humans, we should keep doctors in the loop to make the final decision.

 

[dystopian]

Until some far future time when technology gives us a system that is more versatile, insightful, and creative than humans, we should keep doctors in the loop to make the final decision.

I don't disagree with this; although I might take issue with the 'far future' part of it; although I guess that depends on how you define that term. I think we'll get there sooner rather than later; but I can't give you an exact timetable, and having humans involved seems like a good idea for the foreseeable future.

 

[skepticalbip]

Until some far future time when technology gives us a system that is more versatile, insightful, and creative than humans, we should keep doctors in the loop to make the final decision.

I don't disagree with this; although I might take issue with the 'far future' part of it; although I guess that depends on how you define that term. I think we'll get there sooner rather than later; but I can't give you an exact timetable, and having humans involved seems like a good idea for the foreseeable future.

Your are right. My personal definition of "far future" is "some time beyond the foreseeable future". Mileage will vary depending on the driver.

 

[dystopian]

Your are right. My personal definition of "far future" is "some time beyond the foreseeable future". Mileage will vary depending on the driver.

Well, now we need to define 'foreseeable future', I guess.

 

[Kharakov]

The doctor looks at the boys eyes and into his mouth (I don't know why doctors always seem to do this but the do) and notices the smell of Bourbon on the boy's breath. The doctor recommends a breathalyzer test or a blood sample for blood alcohol level. A much cheaper test to determine the problem for sure.

I'd think that any robo-doc would include sensors that swab the interior of the mouth, monitor exhaled gases, etc. because of the many things that can be detected via these things (especially exhaled gasses- kidney problems, diabetes, alcohol, certain toxins, etc.).

And truthfully, a human doc has to go through a checklist too- so a robo doc would simply perform the same calculations (compare symptoms to known maladies, compare various tests (such as alcohol detection), etc.), only a billion times faster.

 

[Perspicuo]

False. Hook the computer up to a camera and sufficiently advanced software and it could. Or plug the patient into a future-tech full body medical scanner, probably letting it not only identify the source of the pain, but anything else wrong with you that you might not even be aware of. There is absolutely no fundamental reason why technology couldn't become advanced enough to completely replace doctors...

Yeah there is. Programming is static. Doctors are updated constantly, in every interaction. If you want your program to be as up to date as a doctor, then you need to update as often as a doctor, which means keeping up the same social and professional interactions as a doctor. Given that you're having to do all that, what exactly is the advantage the computer has, that a doctor using a computer index doesn't?

With computerized medicine, every treatment is a test for the particular treatment decision, and therefore with every iteration (e.g. every patient that is given a certain drug for acid reflux) creates more data and greater precision in the big data that computerized medicine uses to treat people. Family doctors today don't provide this benefit to science at large and thus to all future patients.

 

[skepticalbip]

The doctor looks at the boys eyes and into his mouth (I don't know why doctors always seem to do this but the do) and notices the smell of Bourbon on the boy's breath. The doctor recommends a breathalyzer test or a blood sample for blood alcohol level. A much cheaper test to determine the problem for sure.

I'd think that any robo-doc would include sensors that swab the interior of the mouth, monitor exhaled gases, etc. because of the many things that can be detected via these things (especially exhaled gasses- kidney problems, diabetes, alcohol, certain toxins, etc.).

And truthfully, a human doc has to go through a checklist too- so a robo doc would simply perform the same calculations (compare symptoms to known maladies, compare various tests (such as alcohol detection), etc.), only a billion times faster.

I don’t disagree but such a robo-doc is still in the far future. As I said:

Until some far future time when technology gives us a system that is more versatile, insightful, and creative than humans, we should keep doctors in the loop to make the final decision.

However, as of now, the human nose and eyes (of course, linked to the brain) are much better at quickly detecting and discriminating a wide range of sensory input than any of our technology. Our technology is better at looking for one specific thing in a sample in that it can detect trace amount below what the human nose can but unless the detector was specifically designed to detect that one thing then it would take hours of lab time to find it.

It has been quite a while, back to the stone age, since I took chemistry (required in my university for all science students) the final course in that series was qualitative analysis. Evaluating an unknown sample at the time I took the class, as I recall, required eight to ten hours of lab work. Of course if we were asked to look for only one thing, it would only take an hour or two. Today, it can be done more quickly but a test using liquid chromatography still takes several hours.

The human senses are amazing. What kind of technical sensor would we need to approximate them? We can easily tell that we are passing a bakery, an outhouse, a pigsty and identify which blindfolded. Technology, unless specifically designed for that purpose could only tell us that there are organic molecules in the air. And if we built a sensor to tell us which we had passed, it wouldn’t be able to identify a perfumery if we passed one, that would take a redesign to add that particular capability.

 

[Duke Leto]

However, my comment about doctors is this. A machine cannot 'get to know you', cannot take into account 'your quirks'. My doctor knows the medications I am on, which ones are more effective, what I am allergic too etc. If I was prescribed the 'best' medication as determined by a machine, my depression and anxiety would still be quite bad. If I was prescribed the 'most effective known antibiotic for this condition' I would spend my recovery time suffering from the side effects.

Well my only response to that would be to have you and bilby Google a variety of common phrases and concepts that can have multiple meanings and note the differences in your results. Or carefully note the differences in online ads you get from surfing to exactly the same web page.

Stating that it is impossible for a program to adapt itself to your quirks and peculiarities or to gain knowledge about them is a bit short-sighted when you realize that an example of a program doing exactly this on a daily basis is within your experience.

I can relate to the frustration in dealing with incompetent humans doing anti-depressant prescriptions by basically pigeonholing you into a set of if/then questions and having to deal with horribly made software that does the same, but software does not have the limitations you are defining as a fundamental property. It is possible for a large, well ordered database of personal preferences and quirks to match itself to you with as high a degree of fidelity as a human, and the machines and their programmers are getting better and better at delivering this.

 

[Duke Leto]

I'm concerned about the diminished job security because of this.

Happy are (will be) those who can adapt well to changing labor environments.
This probably will mean shorter majors, so that people will have more of a chance to have two or three different professions.
Having children will be even less practical/sustainable in such a situation.
Those who will have less ability to switch occupations will suffer and protest more. That is to say, most people will be in that position.
I'm expecting big social upheavals, just because of automatization.
Either that or professionals will start to become adjuncts of technology (the computer decides your diagnosis, the physician will just check on details and deliver bedside manners), and still, they won't be able to charge/make as much as they used to.

Well, what will eventually happen, assuming that we don't have a nuclear war over access to ever more limited energy and resources, is that macroeconomics will have to catch up to reality by realizing intervention is vital, especially if, and this is very likely if we defeat the energy crisis in the long term, deflation becomes a persistent societal problem.

If the future is like contemporary Japan, then eventually the following things will have to happen:

1.) We give up on a monetary system that mandates high real interest rates in a deflationary environment. That is we scrap the stuff that makes the zero lower bound obligatory: physical currency. (0% nominal interest rates are the floor when someone can just keep a sack of banknotes in a safe instead of depositing them at the bank, and when there is any deflation, 0% nominal rates are positive real rates.) That means some big changes in the way electronic transactions are processed. But since you have the benefit of negative real rates to juice the economy as needed on a consistent basis, it's a pretty clear net societal gain.

2.) Economics being the science of dealing with scarcity, and what we are discussing being a permanent scarcity of underemployed labor, you end up having to do what worked (or failed to hurt anything) in the 1930s-1960s, you have to start rationing work hours by setting a mandatory legal maximum of work hours per week and sticking to it. Diminished job security, or (what amounts to roughly the same thing), the absence of plentiful options of replacement employment (or economic incentives for employers to offer job security as a form a compensation, also actually the same thing) is not as much of an issue in that case. If the legal maximum work day is 10 hours instead of 40, then having four times as many potential employees as you need stops being a problem.

This requires a political system amenable to such changes, one that doesn't regard a positive real return on capital as fundamental right of the universe and one that sees workers as an important constituency, but both of those become smaller and smaller obstacles in a democratic state the worse the problems you are concerned about get. A world where medical doctors are in lines at soup kitchens is one where virtually nobody is doctrinaire libertarian but the very few owners of capital.

Want proof? The imminent self-immolation of the Republican Party of Kansas as a direct result of adhering to doctrinaire financial/fiscal libertarianism ought to be all you need.

 

[barbos]

...uhm, yes? Was this ever really in doubt by anyone?

I'm much more interested in how often people were actually *right* in their future predictions. Everything you've listed was to some degree predicted before it became real. It used to be science fiction, and is now science fact. I'm interested to see which predictions will become true next. For instance, we're currently seeing several technologies from Star Trek becoming real; the 3d printing revolution certainly represents the first step to a true 'replicator'; we now have prototype handheld medical scanners, primitive tractor beam prototypes, and warp drives are looking more plausible than ever before; the future's definitely looking bright, at least when it comes to scientific/technological progress.

That's the funny part about it.

Most people are texting away on the rocket-ship in their pocket, completely oblivious to how incredible what they're doing is in the context of human history, or even world history.

But then, in the context of what may be to come what we've seen so far might be nothing.

Louis CK, is that you?

 

[NobleSavage]

Or alternately, it could result in fewer doctors. With technology aiding them, the remaining doctors would spend less time with each patient so increasing their patient load. More patients with each patient paying less could result in no change in the doctor's total income.

I'd rather we retain high-skilled doctors.
As someone working in a field where most of the computation is automated, I often notice it takes even higher skills and experience to notice there might be an error in the computation hypothesis and a need to crosscheck the computation with a different method.
I'd rather my doctor doesn't take blindly a computer's advice.
(Now, some memory helps like the computer scanning my file and the prescription in progress to warn of possible medication interactions, I'm all for it. But that's not the future, that already exists)

It's basically just a matter of who makes more mistakes. A computer or a doctor. Yeah, the software can make mistakes, but humans make A LOT of mistakes http://en.wikipedia.org/wiki/To_Err_is_Human

"Watson is already capable of storing far more medical information than doctors, and unlike humans, its decisions are all evidence-based and free of cognitive biases and overconfidence. It's also capable of understanding natural language, generating hypotheses, evaluating the strength of those hypotheses, and learning — not just storing data, but finding meaning in it.

As IBM scientists continue to train Watson to apply its vast stores of knowledge to actual medical decision-making, it's likely just a matter of time before its diagnostic performance surpasses that of even the sharpest doctors."

http://www.businessinsider.com/ibms-watson-may-soon-be-the-best-doctor-in-the-world-2014-4

I'd rather we retain high-skilled doctors.
As someone working in a field where most of the computation is automated, I often notice it takes even higher skills and experience to notice there might be an error in the computation hypothesis and a need to crosscheck the computation with a different method.
I'd rather my doctor doesn't take blindly a computer's advice.
(Now, some memory helps like the computer scanning my file and the prescription in progress to warn of possible medication interactions, I'm all for it. But that's not the future, that already exists)

TBH, My doctor has left the practice where he was because he was overworked, (and I presume underpaid), and I will follow him. He is a good doctor and I have discovered that once a doctor knows you - YOU FUCKING WELL STICK WITH THEM... It has taken Dr M and I nearly 3 years to get to know each other in terms of my medical, hormonal and psychiatric needs, I am not changing now.

Computers do not, and will never, know us as much as the human does. PERIOD!!!

"heavier-than-air flying machines are impossible" -- Kelvin, Lord William Thomson (a highly respected mathematician and physicist)

Have you seen the move "Her"?