Thoughts on IBM's watson?

[easy peasy]

This was posted on Reddit a few days ago:

IBM's Watson was tested on 1,000 cancer diagnoses made by human experts. In 30 percent of the cases, Watson found a treatment option the human doctors missed. Some treatments were based on research papers that the doctors had not read. More than 160,000 cancer research papers are published a year.​

NYT article link: http://www.nytimes.com/2016/10/17/t...et-on-watson-and-paying-big-money-for-it.html

Just curious on what your thoughts on the future of diagnostics might be?

Here are my thoughts:
How can one person, even the brightest and most motivated, read through hundreds of thousands of published research papers? There is just no way to compete with Watson on the ability to scan research studies and gain new knowledge at the same rate, or even anywhere near the same rate. It doesn't need to sleep, eat, take holidays or sick days, its ego doesn't get in the way or have any other weakness that affects us simple humans.

I'm very curious to see how things will change in the next decade. I've been fascinated by Watson from when I was a kid so this, to me, is kind of scary and exciting at the same time.

---

Members don't see this ad.

 

[Psai]

Just because something is published doesn't mean it's legit. You can pretty much publish anything if you are persistent. And having a regimen work in one patient doesn't mean that it will work in another. If a combination works in 60% of patients for 2 year remission, that means 40% won't have that same benefit. You can't just mine the data, you need to interpret it.

 

[SurfingDoctor]

This was posted on Reddit a few days ago:

IBM's Watson was tested on 1,000 cancer diagnoses made by human experts. In 30 percent of the cases, Watson found a treatment option the human doctors missed. Some treatments were based on research papers that the doctors had not read. More than 160,000 cancer research papers are published a year.​

NYT article link: http://www.nytimes.com/2016/10/17/t...et-on-watson-and-paying-big-money-for-it.html

Just curious on what your thoughts on the future of diagnostics might be?

Here are my thoughts:
How can one person, even the brightest and most motivated, read through hundreds of thousands of published research papers? There is just no way to compete with Watson on the ability to scan research studies and gain new knowledge at the same rate, or even anywhere near the same rate. It doesn't need to sleep, eat, take holidays or sick days, its ego doesn't get in the way or have any other weakness that affects us simple humans.

I'm very curious to see how things will change in the next decade. I've been fascinated by Watson from when I was a kid so this, to me, is kind of scary and exciting at the same time.

I think human pathology is incredibly complex. Far more so than the human brain can understand and synthesize. Computers do not have the creativity to dream up new ways to tackle questions, however they have the ability to organize a massive amount of data that humans can not. I don't think computers will replace humans in science and medicine because they lack the creativity aspect (at least so far), however I do think they will be integrated (and integral) far more so then they are now.

 

[DrBowtie]

This was posted on Reddit a few days ago:

IBM's Watson was tested on 1,000 cancer diagnoses made by human experts. In 30 percent of the cases, Watson found a treatment option the human doctors missed. Some treatments were based on research papers that the doctors had not read. More than 160,000 cancer research papers are published a year.​

NYT article link: http://www.nytimes.com/2016/10/17/t...et-on-watson-and-paying-big-money-for-it.html

Just curious on what your thoughts on the future of diagnostics might be?

Here are my thoughts:
How can one person, even the brightest and most motivated, read through hundreds of thousands of published research papers? There is just no way to compete with Watson on the ability to scan research studies and gain new knowledge at the same rate, or even anywhere near the same rate. It doesn't need to sleep, eat, take holidays or sick days, its ego doesn't get in the way or have any other weakness that affects us simple humans.

I'm very curious to see how things will change in the next decade. I've been fascinated by Watson from when I was a kid so this, to me, is kind of scary and exciting at the same time.

Yawn. Until this changes NCCN guidelines it's not going to hit clinical practice.

 

[rederza]

Watson is such a cute name. I can't wait for machine learning to really take hold. Also Watson will likely improve so much in the time it takes to bring one college student to a full fledged doctor thanks to the redundant and inefficient American medical education system.

 

[OnePunchBiopsy]

Still can't do surgery...

 

[easy peasy]

Still can't do surgery...

That's exactly what i was thinking. Surgery, psych, and PM&R are still untouchable by AI.

 

[zaztong]

Studied CS with machine learning. I've worked small contract work as a data scientist. I've also had the chance to work on a project using Watson's API.

My former startup partner is working on a product to replace call centers. He's one of the best people I know, and he's not going to be even close to replacing humans fully in call center customer service, which is relatively repetitive/few variations when it is based on a call script. Machine learning is a buzzword that sounds better than it actually is, especially with all the sci-fi media chiming in on AI hype. IBM is partly getting as much media attention as they can for increased sales from brand reputation.

Let's avoid legal, insurance or even customer adoption discussions. Just on the technical side, IBM Watson is not going to replace any physicians, I've talked to the team once personally. You should worry more about outsourcing if you are a radiologist. It's not likely going to make any college students into doctors right away, because learning material means you have to understand it at its core and we have a fairly rigid, conservative system in place. There will probably be slightly less emphasis on memorizing dull details in medical school though in the future.

More likely it would be with physicians in a man-machine interface to reduce medical errors. They haven't even fixed relatively simpler complex problems that make much money than medicine. Based on automation research data, it's extremely unlikely (<2%) they will fully automate a physician's work in our lifetime of work, due to the amount of variation encountered in medicine and the empathy needed to determine symptoms. Automated medical diagnoses are only as good as the input it gets - namely symptoms interpreted by the physician and the physical exam etc.

If you're worried about job security, you should stop because we have one of the safest jobs that won't ever likely be replaced in our lifetime, based on my current research into the matter.

 

[Stagg737]

Just because something is published doesn't mean it's legit. You can pretty much publish anything if you are persistent. And having a regimen work in one patient doesn't mean that it will work in another. If a combination works in 60% of patients for 2 year remission, that means 40% won't have that same benefit. You can't just mine the data, you need to interpret it.

Add to that the fact that hospitals probably can't bill insurance or medicaid/care for a large percentage of those treatments and that makes Watson's utility even smaller. No one is going to want implement a machine that doesn't produce billable treatment plans.

 

[rederza]

Studied CS with machine learning. I've worked small contract work as a data scientist. I've also had the chance to work on a project using Watson's API.

My former startup partner is working on a product to replace call centers. He's one of the best people I know, and he's not going to be even close to replacing humans fully in call center customer service, which is relatively repetitive/few variations when it is based on a call script. Machine learning is a buzzword that sounds better than it actually is, especially with all the sci-fi media chiming in on AI hype. IBM is partly getting as much media attention as they can for increased sales from brand reputation.

Let's avoid legal, insurance or even customer adoption discussions. Just on the technical side, IBM Watson is not going to replace any physicians, I've talked to the team once personally. You should worry more about outsourcing if you are a radiologist. It's not likely going to make any college students into doctors right away, because learning material means you have to understand it at its core and we have a fairly rigid, conservative system in place. There will probably be slightly less emphasis on memorizing dull details in medical school though in the future.

More likely it would be with physicians in a man-machine interface to reduce medical errors. They haven't even fixed relatively simpler complex problems that make much money than medicine. Based on automation research data, it's extremely unlikely (<2%) they will fully automate a physician's work in our lifetime of work, due to the amount of variation encountered in medicine and the empathy needed to determine symptoms. Automated medical diagnoses are only as good as the input it gets - namely symptoms interpreted by the physician and the physical exam etc.

If you're worried about job security, you should stop because we have one of the safest jobs that won't ever likely be replaced in our lifetime, based on my current research into the matter.

You misread me. I didn't sat it's Watson's job to pump out faster doctors. I said in the time it takes to make one doctor in America starting at their entry into college till finishing residency, Watson will have improved by then. Which hopefully, will only force the system to cut out unneeded crap like 4 years of undergrad filled with irrelevant extraneous classes meant to suck up more time and ad dmore debt.

 

[Mad Jack]

Still can't do surgery...

They're working on that too, but that is also in its infancy.

http://spectrum.ieee.org/the-human-...t-surgeon-bests-human-surgeons-in-world-first

Give it a few decades. Might see autonomous robotic surgeons sometime before we die for minor procedures.

 

[cbrons]

It's not likely going to make any college students into doctors right away, because learning material means you have to understand it at its core

Then explain Nurse Practitioners to me. They have all the rights and privileges of physicians (in many states) without any understanding of the "learning material" from medical school.

 

[cbrons]

Computers will eventually take-over every human industry. It's inevitable. Don't know when it will happen though. Probably after robots are doing things like driving our cars for us, or even before that, doing manual labor. I find it very funny in science fiction that a common trope is the "medical box" - where injured humans are put into a machine that essentially diagnoses and fixes them. In many of these same stories, there are people still working in factories, paving roads, and cleaning houses.

I just think its hilarious how people worry about computers replacing the most complex profession (medicine) without worrying about mindless meanial labor jobs that would be much simpler to program computers to do.

 

[rederza]

Computers will eventually take-over every human industry. It's inevitable. Don't know when it will happen though. Probably after robots are doing things like driving our cars for us, or even before that, doing manual labor. I find it very funny in science fiction that a common trope is the "medical box" - where injured humans are put into a machine that essentially diagnoses and fixes them. In many of these same stories, there are people still working in factories, paving roads, and cleaning houses.

I just think its hilarious how people worry about computers replacing the most complex profession (medicine) without worrying about mindless meanial labor jobs that would be much simpler to program computers to do.

In Austin, many google self driving cars can be seen navigating around the streets so your future may not be so far off.

 

[zaztong]

Then explain Nurse Practitioners to me. They have all the rights and privileges of physicians (in many states) without any understanding of the "learning material" from medical school.

NPs generally take clinical practice, biochem, micro, physiology, anatomy, pathophysiology and clinical pharmacology in pre-nursing/BSN/MSN programs.

 

[cbrons]

NPs generally take clinical practice, biochem, micro, physiology, anatomy, pathophysiology and clinical pharmacology in pre-nursing/BSN/MSN programs.

Joke classes

Sent from my SM-N910P using SDN mobile

 

[kbeitz]

NPs generally take clinical practice, biochem, micro, physiology, anatomy, pathophysiology and clinical pharmacology in pre-nursing/BSN/MSN programs.

Are you a robot that compared the course schedule between nursing and medical school before concluding no significant difference p>.05?

I see right through your ruse, trying to convince us that you're a human. Moderators, ban this toaster!

 

[sazerac]

I would be quite surprised if Watson affects clinical medicine in any meaningful way over the next several decades of our careers. They recently had Watson take USMLE Step 1, and it only scored a 222. Completely worthless.

 

[DrBowtie]

I would be quite surprised if Watson affects clinical medicine in any meaningful way over the next several decades of our careers. They recently had Watson take USMLE Step 1, and it only scored a 222. Completely worthless.

Link?

 

[chipwhitley]

I just think its hilarious how people worry about computers replacing the most complex profession (medicine) without worrying about mindless meanial labor jobs that would be much simpler to program computers to do.

Well first of all machines have already replaced a huge number of manufacturing type jobs, but what's the incentive to spend a lot of money designing and testing a machine if you can get a poor kid in Bangladesh to do the job for 25 cents an hour anyway? There's not much money to be made there. However, you can make a TON of money by reducing the need for workers that cost 6 figures to employ (hundreds of dollars an hour). Basically a goldmine for the first company to figure out how to do it.

 

[sazerac]

Link?

Link to value of the step1 score:

http://forums.studentdoctor.net/threads/step1-222-value.1212463/#post-17959341

 

[zaztong]

Well first of all machines have already replaced a huge number of manufacturing type jobs, but what's the incentive to spend a lot of money designing and testing a machine if you can get a poor kid in Bangladesh to do the job for 25 cents an hour anyway? There's not much money to be made there. However, you can make a TON of money by reducing the need for workers that cost 6 figures to employ (hundreds of dollars an hour). Basically a goldmine for the first company to figure out how to do it.

Just because there's relatively a lot of money doesn't guarantee much. There's a lot more money to beat the stock market than replacing doctors. You'd be a trillionaire or better.

In case you haven't noticed, they haven't beaten the stock market yet - not even close.

 

[cbrons]

Well first of all machines have already replaced a huge number of manufacturing type jobs, but what's the incentive to spend a lot of money designing and testing a machine if you can get a poor kid in Bangladesh to do the job for 25 cents an hour anyway? There's not much money to be made there. However, you can make a TON of money by reducing the need for workers that cost 6 figures to employ (hundreds of dollars an hour). Basically a goldmine for the first company to figure out how to do it.

It's still more plausible to program machines for mindless rote tasks vs. complex jobs which require higher-level abstract thinking/creativity.

Also, there are 100x more people doing menial labor than rich executives. Programming machines to do every bit of the manufacturing would cost a lot in the short-run, but save an enormous amount in the long-run. Aside from the fact that machines could have many-fold times the productive capacity, human capital has its own uncertainties and inefficiencies which machines do not have.

 

[cbrons]

I would be quite surprised if Watson affects clinical medicine in any meaningful way over the next several decades of our careers. They recently had Watson take USMLE Step 1, and it only scored a 222. Completely worthless.

good troll post 7.5/10.

 

[kbeitz]

However, you can make a TON of money by reducing the need for workers that cost 6 figures to employ (hundreds of dollars an hour). Basically a goldmine for the first company to figure out how to do it.

It will always be more lucrative to help healthcare workers do their jobs than replace them. Much easier to sell the pants than mine the gold.

 

[musicalfeet]

Studied CS with machine learning. I've worked small contract work as a data scientist. I've also had the chance to work on a project using Watson's API.

My former startup partner is working on a product to replace call centers. He's one of the best people I know, and he's not going to be even close to replacing humans fully in call center customer service, which is relatively repetitive/few variations when it is based on a call script. Machine learning is a buzzword that sounds better than it actually is, especially with all the sci-fi media chiming in on AI hype. IBM is partly getting as much media attention as they can for increased sales from brand reputation.

Let's avoid legal, insurance or even customer adoption discussions. Just on the technical side, IBM Watson is not going to replace any physicians, I've talked to the team once personally. You should worry more about outsourcing if you are a radiologist. It's not likely going to make any college students into doctors right away, because learning material means you have to understand it at its core and we have a fairly rigid, conservative system in place. There will probably be slightly less emphasis on memorizing dull details in medical school though in the future.

More likely it would be with physicians in a man-machine interface to reduce medical errors. They haven't even fixed relatively simpler complex problems that make much money than medicine. Based on automation research data, it's extremely unlikely (<2%) they will fully automate a physician's work in our lifetime of work, due to the amount of variation encountered in medicine and the empathy needed to determine symptoms. Automated medical diagnoses are only as good as the input it gets - namely symptoms interpreted by the physician and the physical exam etc.

If you're worried about job security, you should stop because we have one of the safest jobs that won't ever likely be replaced in our lifetime, based on my current research into the matter.

Off topic, but are you saying friend is basically creating those automated machines everyone hates talking to in customer service? No offense to your friend, but those things suck majorly. I used to humor the machine and talk to it, now I just press 00000000000 x 10 to get to a real person.

 

[zaztong]

Off topic, but are you saying friend is basically creating those automated machines everyone hates talking to in customer service? No offense to your friend, but those things suck majorly. I used to humor the machine and talk to it, now I just press 00000000000 x 10 to get to a real person.

No he didn't invent them, he's trying to make it better and more human-like

 

[Cara Delevingne]

I'd imagine Watson would make physicians' jobs much easier and less error-prone, but you definitely need a person to assess patients, communicate with patients, and carry out procedures.

In fact, instead of being afraid of how Watson might make physicians "obsolete" - by having lesser-qualified people who know enough to operate Watson replace physicians - I think it'll allow physicians to make patient visits MUCH shorter and/or more efficient, which should result in an overall win-win. Physicians see more patients, make more $ (or heal more people, however you count a "win"), and patients don't have to wait months to see doctors anymore to get something potentially serious checked out.

 

[Psai]

I'd imagine Watson would make physicians' jobs much easier and less error-prone, but you definitely need a person to assess patients, communicate with patients, and carry out procedures.

In fact, instead of being afraid of how Watson might make physicians "obsolete" - by having lesser-qualified people who know enough to operate Watson replace physicians - I think it'll allow physicians to make patient visits MUCH shorter and/or more efficient, which should result in an overall win-win. Physicians see more patients, make more $ (or heal more people, however you count a "win"), and patients don't have to wait months to see doctors anymore to get something potentially serious checked out.

Patients already think doctor visits are too short. Seeing more patients isn't a win either, the more you see the worse you feel.

 

[TheoLeo]

Instead of blindly improving computers to slowly replace human labor and abilities, one should just stop for a minute and ask if we really want it and if we really need it. In the end, we end up losing a lot of jobs and only a few people at IBM will then pile up all the cash that once went into the pockets of nureses, physicians, lawyers, cleaning staff etc.

 

[zaztong]

Instead of blindly improving computers to slowly replace human labor and abilities, one should just stop for a minute and ask if we really want it and if we really need it. In the end, we end up losing a lot of jobs and only a few people at IBM will then pile up all the cash that once went into the pockets of nureses, physicians, lawyers, cleaning staff etc.

It's not as simple as a win-lose machines-humans. There are new jobs created and old ones replaced, goods can be cheaper as a result. Just don't end up in the old loser pile and adapt to new jobs if you happen to be unlucky enough. A lot of manual labor jobs are gone because of automation, but everyone has a higher quality of life.

 

[Newhorizons21]

It's not as simple as a win-lose machines-humans. There are new jobs created and old ones replaced, goods can be cheaper as a result. Just don't end up in the old loser pile and adapt to new jobs if you happen to be unlucky enough. A lot of manual labor jobs are gone because of automation, but everyone has a higher quality of life.

- said the med student from the ivory tower.

Pew’s America’s shrinking middle class shows that middle-class household income has declined throughout the population, while at the same time the gap between low- and upper-income households has grown, demonstrating a significant increase in income inequality across the US. A major contributor to economic decline and inequality has been the plunge in manufacturing jobs and wages.

http://www.globalresearch.ca/povert...dle-class-declining-manufacturing-pay/5525182

Being poor in the United States is so hazardous to your health, a new study shows, that the average life expectancy of the lowest-income classes in America is now equal to that in Sudan or Pakistan.

http://news.harvard.edu/gazette/story/2016/04/for-life-expectancy-money-matters/

These white Americans have been hard hit by the loss of manufacturing jobs, and the fact that their lifespans are suffering correlates with ample research showing that earning less is tied to dying earlier.

http://blogs.wsj.com/economics/2016...white-americans-are-forcing-a-policy-rethink/

I'm not sure everyone has a higher quality of life.

 

[zaztong]

- said the med student from the ivory tower.

Pew’s America’s shrinking middle class shows that middle-class household income has declined throughout the population, while at the same time the gap between low- and upper-income households has grown, demonstrating a significant increase in income inequality across the US. A major contributor to economic decline and inequality has been the plunge in manufacturing jobs and wages.

http://www.globalresearch.ca/povert...dle-class-declining-manufacturing-pay/5525182

Being poor in the United States is so hazardous to your health, a new study shows, that the average life expectancy of the lowest-income classes in America is now equal to that in Sudan or Pakistan.

http://news.harvard.edu/gazette/story/2016/04/for-life-expectancy-money-matters/

These white Americans have been hard hit by the loss of manufacturing jobs, and the fact that their lifespans are suffering correlates with ample research showing that earning less is tied to dying earlier.

http://blogs.wsj.com/economics/2016...white-americans-are-forcing-a-policy-rethink/

I'm not sure everyone has a higher quality of life.

I urge you to look in the long-term, not the short-term, before disparagingly calling someone elitist and coming to conclusions without looking at the big picture.

Consider the Industrial Revolution and the Green Revolution. Productivity increased greatly, cost of goods dropped (relative to inflation), access to basics increased, we are still living longer than before, including the poor (even according to your article, the poorest are still living longer). Global economy overall increased. Back then lots of people lost jobs specifically in manual labor. People adapt, and there are still currently jobs, relatively low unemployment rates (5%). "A major contributor to economic decline" in what timeframe? Last time I checked the GDP of the USA adjusted for inflation has increased over the long-term (if we include the ups-and-downs) a lot.

Think about this - I've been to the ghetto, and I've seen supposedly poor people with smartphones. What does that tell you? It's all relative. I'm not saying we shouldn't dismiss the pain and suffering of the lower class, I'm just saying that relatively, we're doing better than the past. Goods and services are at lower cost to produce. We aren't starving in world famine as Thomas Malthus predicted.

Have you thought that income inequality isn't necessarily inherently bad? Every economy has income inequality. Forced equality means less opportunity to pursue what makes you individually great. In a free market economy, people become wealthy making what the rich enjoy today into something almost everybody can enjoy tomorrow. But if someone, say, a government bureaucrat, told you that your ambition had limits, that there was a ceiling above which you could not rise, there won't be much incentive to invent and gamble on new ideas. Let's say we attempt to redistribute wealth, rich leave, no more rich investors looking to get more rich, and our country would decline much, much more. What matters is we keep the doors open to opportunity and social mobility, while penalizing perverse incentives.

If we don't increase our productivity and stay competitive, another country will and develop technologies to beat us. Although I can't predict the future, long-term, we would likely be worse off not developing such technologies that increase our productivity.

 

[chipwhitley]

Just because there's relatively a lot of money doesn't guarantee much. There's a lot more money to beat the stock market than replacing doctors. You'd be a trillionaire or better.

In case you haven't noticed, they haven't beaten the stock market yet - not even close.

Stock market activity is largely dictated by the emotions of the traders. In other words, to beat the stock market, the machine would have to understand human psychology better than humans. That's far more difficult than understanding medicine, which, in 90%+ of cases, is basically advanced pattern recognition, and pattern recognition is one of the things computers are good at.

 

[zaztong]

Stock market activity is largely dictated by the emotions of the traders. In other words, to beat the stock market, the machine would have to understand human psychology better than humans. That's far more difficult than understanding medicine, which, in 90%+ of cases, is basically advanced pattern recognition, and pattern recognition is one of the things computers are good at.

Medicine includes psychiatry, and even generally speaking, interpreting symptoms requires empathy and even cultural knowledge. I'd say it's roughly the same or more difficult to solve medicine.

And there's something inherently problematic with solving 90% if we even could right now. Getting 90% is relatively easy. 99% or 99.9999% is terribly difficult on several magnitudes or more.

 

[chipwhitley]

Medicine includes psychiatry, and even generally speaking, interpreting symptoms requires empathy and even cultural knowledge. I'd say it's roughly the same or more difficult to solve medicine.

And there's something inherently problematic with solving 90% if we even could right now. Getting 90% is relatively easy. 99% or 99.9999% is terribly difficult on several magnitudes or more.

90% is more than enough to disrupt an industry in a huge way.

And we are moving quickly into an era where we are not reliant on people's recollection of symptoms to make a diagnosis. A variety of sensors built into everyday devices, genetic information, big data approaches using information from all the interactions people have in the world will give us more information we need to make a more accurate picture of an individual's health than we have ever had before. One day we will look back at how we used to ask people to describe their symptoms in order to come up with a diagnosis and laugh in the same way we laugh at early doctors for blood letting.

 

[zaztong]

90% is more than enough to disrupt an industry in a huge way.

And we are moving quickly into an era where we are not reliant on people's recollection of symptoms to make a diagnosis. A variety of sensors built into everyday devices, genetic information, big data approaches using information from all the interactions people have in the world will give us more information we need to make a more accurate picture of an individual's health than we have ever had before. One day we will look back at how we used to ask people to describe their symptoms in order to come up with a diagnosis and laugh in the same way we laugh at early doctors for blood letting.

"One day" is far away and optimistic. Big data is simply another term for big mess for anyone in the data analytics industry. Actually gaining insights from the data is rough when you have extreme noise. If a pilot on an airplane only landed 90% of the time correctly, the other 10% of the time, all the passengers may die.

 

[chipwhitley]

"One day" is far away and optimistic. Big data is simply another term for big mess for anyone in the data analytics industry. Actually gaining insights from the data is rough when you have extreme noise. If a pilot on an airplane only landed 90% of the time correctly, the other 10% of the time, all the passengers may die.

So basically you're hoping that medicine doesn't improve for a long time, because it might cut into your earning potential, although it would be beneficial for society.

 

[kbeitz]

Stock market activity is largely dictated by the emotions of the traders. In other words, to beat the stock market, the machine would have to understand human psychology better than humans. That's far more difficult than understanding medicine, which, in 90%+ of cases, is basically advanced pattern recognition, and pattern recognition is one of the things computers are good at.

True story, I used to work as a data analyst for a big data startup building an emotion recognition platform. I tracked sensitivity-specificity metrics, so my job literally involved telling the boss his program was a piece of ****.

So one day the boss walks in and says he has this great idea about using the platform to track the stock market. I turn around pulling up a Google search showing him that not only is his idea not unique, but those who try fail spectacularly. Boss walks out telling everyone else we're going to be filthy rich.

Having left this job only recently to come to medical school, I'm convinced of two things: there are simply places that computers cannot go because the parameters (e.g. emotion, care quality) are hard if not impossible to define ; and big tech CEOs are so full of **** they need daily disimpaction before coming to work.

 

[Cello]

When robots have all the jobs, who will buy all of the goods they produce?

 

[Newhorizons21]

When robots have all the jobs, who will buy all of the goods they produce?

The Industrial Revolution and its consequences have been a disaster for the human race. They have greatly increased the life-expectancy of those of us who live in “advanced” countries, but they have destabilized society, have made life unfulfilling, have subjected human beings to indignities, have led to widespread psychological suffering (in the Third World to physical suffering as well) and have inflicted severe damage on the natural world. The continued development of technology will worsen the situation. It will certainly subject human beings to greater indignities and inflict greater damage on the natural world, it will probably lead to greater social disruption and psychological suffering, and it may lead to increased physical suffering even in “advanced” countries.

(do not look at the source for the above quote)

 

[Cello]

The Industrial Revolution and its consequences have been a disaster for the human race. They have greatly increased the life-expectancy of those of us who live in “advanced” countries, but they have destabilized society, have made life unfulfilling, have subjected human beings to indignities, have led to widespread psychological suffering (in the Third World to physical suffering as well) and have inflicted severe damage on the natural world. The continued development of technology will worsen the situation. It will certainly subject human beings to greater indignities and inflict greater damage on the natural world, it will probably lead to greater social disruption and psychological suffering, and it may lead to increased physical suffering even in “advanced” countries.

(do not look at the source for the above quote)

Marx? Engels? Castro? Che?

 

[Newhorizons21]

Marx? Engels? Castro? Che?

Kaczynski (aka the Unabomber)

 

[Cello]

Kaczynski (aka the Unabomber)

Does this mean that we will have to bomb the robots in the future? I am struggling to see the relevance to my original question but feel like it is there, lurking just beyond my grasp!

 

[zaztong]

So basically you're hoping that medicine doesn't improve for a long time, because it might cut into your earning potential, although it would be beneficial for society.

I'm not hoping, just a guess. If I had the opportunity to jump on a promising project, I would improve medicine. It's much more self-fulfilling than just money. I've got more than enough of that, and I enjoy a frugal life. I'm saving to self-fund my projects.

When robots have all the jobs, who will buy all of the goods they produce?

Economic forces will prevent this from happening. If there are no earners, there will be no buyers, assuming humans are earners and buyers initially.

There will always be demand for things humans want that aren't necessarily reproducible, such as art. Humans can focus on more creative pursuits and sell those instead. All needs that require manual labor could become extremely cheap.

 

[zaztong]

The Industrial Revolution and its consequences have been a disaster for the human race. They have greatly increased the life-expectancy of those of us who live in “advanced” countries, but they have destabilized society, have made life unfulfilling, have subjected human beings to indignities, have led to widespread psychological suffering (in the Third World to physical suffering as well) and have inflicted severe damage on the natural world. The continued development of technology will worsen the situation. It will certainly subject human beings to greater indignities and inflict greater damage on the natural world, it will probably lead to greater social disruption and psychological suffering, and it may lead to increased physical suffering even in “advanced” countries.

(do not look at the source for the above quote)

War and suffering has plagued humanity before scientific revolution.

Look at all metrics of Human Development Index. It has increased over time over all countries. It's easy to say everyone is suffering psychologically more because of technology without proper quantitative evidence, if you are basing this of anecdotal evidence with your own subjective experience.

 

[Cello]

There will always be demand for things humans want that aren't necessarily reproducible, such as art. Humans can focus on more creative pursuits and sell those instead. All needs that require manual labor could become extremely cheap.

Let's say that machines become adept at composing music which people find satisfying. They could produce incredible amounts of music instantaneously and flood the market. Worse still, it may be impossible for anyone to tell the difference between human and machine. Google is already making progress on songwriting. What is stopping machines from creating 'art' which competes with human creations? Google is producing art works which are selling for thousands of dollars. Machines could produce at such a scale that we mere mortals could never hope to compete with. Anyway, you can't run an entire economy simply by trading art.

I think that AI has a long way to go before usurping us entirely. Perhaps someday, but (hopefully) not in our lifetimes.

 

[zaztong]

Let's say that machines become adept at composing music which people find satisfying. They could produce incredible amounts of music instantaneously and flood the market. Worse still, it may be impossible for anyone to tell the difference between human and machine. Google is already making progress on songwriting. What is stopping machines from creating 'art' which competes with human creations? Google is producing art works which are selling for thousands of dollars. Machines could produce at such a scale that we mere mortals could never hope to compete with. Anyway, you can't run an entire economy simply by trading art.

I think that AI has a long way to go before usurping us entirely. Perhaps someday, but (hopefully) not in our lifetimes.

There are some things you can do to include a unique signature on the artwork itself that cannot be replicated by machines using cryptography, such as public-private keys (a beefed up version that would mathematically take longer than the life of the universe to solve). Only the unique one would have value as a collectible, as the fake ones without the signature would not have the same value. Thus, no competition from robots and obvious human demand. And that is only one example. There are many other things that can run the economy based on human supply and demand. But that is likely future mini-me's problem, not going to happen in our lifetime. And if robots can make anything extremely cheaply (close to zero prices), then the common person would be living closer to kings and tending towards more creative work over manual work, as long as we are careful with how it is developed. Think about in terms of legalized slaveowners-slavery, but with much more steps. The slaveowners obviously lived much better back then relatively speaking, before human slavery was abolished

 

[Cello]

Why would a consumer care if the work was human in origin? If it looks good, I hang it on my wall. If it sounds good, I listen to it.

 

[zaztong]

Why would a consumer care if the work was human in origin? If it looks good, I hang it on my wall. If it sounds good, I listen to it.

Same reason why rich people won't buy a digital replica of human art (laser printing has made huge strides). Pride in owning the original or "real", whatever that means to the individual. If someone found out they owned a replica, it's frowned upon and that person takes a blow to their reputation. Just because you would be fine with a copy, doesn't every single person would be.

Ask yourself this - why does gold or diamonds cost so much? Or why "imitation" gold or diamonds markets exist? It's because enough people believe it's worth the price, even when utility is far less than it's price. "Artificial prices" are everywhere, some more subtle than others.

I've even seen people buy