Mainstream media - A.I. is better than humans to dx. images in Pathology

[coroner]

Daphne Koller - Stanford PhD in Computer science; former chief computing officer at Google’s A.I. health initiative; and, founder of online learning platforms Corsera and Insitro, was on CNN’s ‘GPS’ with Fareed Zakaria discussing how A.I. can impact the future of healthcare.

Excerpts from the interview pertinent to our field are below.

Link to the podcast: Bill Gates & Tax Rates Fareed Zakaria GPS podcast


Zakaria: What’s the most obvious place machine learning can impact healthcare?
Koller: The most obvious application for machine learning in healthcare is on the diagnostic side. When it (A.I.) looks at a scan, etc. and predicts what the pt. has (dx.). This includes multiple sources of data including an X-ray scan, a pathology slide, or a liquid biopsy and what fragments of DNA are found in the blood.

Z: Why would the computer be better than the great doctor who’s seen a bunch of these same images and tell what the patterns look like?
K: What’s surprising is on the imaging side where we think human beings have been looking at these for over a century, why would a computer be able to do better? Computers are able to discern subtle patterns in these images that are very hard for people to see, partly because they can look at so many different samples and extract commonalities people just can’t hold in their head.
You look at a cancer pathology which has thousands of cells and some of them are aberrant that way and others are aberrant this way, how do you put all of that together into something that corresponds to the diagnosis? But it becomes more relevant when looking at liquid biopsies when you’re looking at very subtle changes in the composition of DNA, people have no idea how to look at that data.

Zakaria concluded by saying, “I think everybody understands the computer can look at images with much greater detail and accuracy than human beings…”

I always said that it’s not a question of “If..”, but “When…” AP in its current way of being practiced will become obsolete. I guesstimated we’ve got about 50 years left. Now, I’m not so sure it’s even that much…

---

Members don't see this ad.

 

[Unty]

Daphne Koller - Stanford PhD in Computer science; former chief computing officer at Google’s A.I. health initiative; and, founder of online learning platforms Corsera and Insitro, was on CNN’s ‘GPS’ with Fareed Zakaria discussing how A.I. can impact the future of healthcare.

Excerpts from the interview pertinent to our field are below.

Link to the podcast: Bill Gates & Tax Rates Fareed Zakaria GPS podcast


Zakaria: What’s the most obvious place machine learning can impact healthcare?
Koller: The most obvious application for machine learning in healthcare is on the diagnostic side. When it (A.I.) looks at a scan, etc. and predicts what the pt. has (dx.). This includes multiple sources of data including an X-ray scan, a pathology slide, or a liquid biopsy and what fragments of DNA are found in the blood.

Z: Why would the computer be better than the great doctor who’s seen a bunch of these same images and tell what the patterns look like?
K: What’s surprising is on the imaging side where we think human beings have been looking at these for over a century, why would a computer be able to do better? Computers are able to discern subtle patterns in these images that are very hard for people to see, partly because they can look at so many different samples and extract commonalities people just can’t hold in their head.
You look at a cancer pathology which has thousands of cells and some of them are aberrant that way and others are aberrant this way, how do you put all of that together into something that corresponds to the diagnosis? But it becomes more relevant when looking at liquid biopsies when you’re looking at very subtle changes in the composition of DNA, people have no idea how to look at that data.

Zakaria concluded by saying, “I think everybody understands the computer can look at images with much greater detail and accuracy than human beings…”

I always said that it’s not a question of “If..”, but “When…” AP in its current way of being practiced will become obsolete. I guesstimated we’ve got about 50 years left. Now, I’m not so sure it’s even that much…
View attachment 251034

Doubt it will happen within my lifetime or career (next 20-30 years), so who cares? Definitely will have an impact but how much? Who knows.

No one should be worrying about AI. If AI was so good it would replace all doctors. Won't be happening anytime soon.

 

[WEBB PINKERTON]

Get as many sources of income going as you can once you are out practicing. Rental properties, businesses etc.

No one thought cytopath would die off as fast as it did either. AI (automated screeners) and molecular have reduced it to rubble.

There is a lot of disruptive tech on the way. Whether it is AI or Guardant360 blood testing or whatever. The days of the skilled morphologist are numbered.

 

[Unty]

Get as many sources of income going as you can once you are out practicing. Rental properties, businesses etc.

No one thought cytopath would die off as fast as it did either. AI (automated screeners) and molecular have reduced it to rubble.

There is a lot of disruptive tech on the way. Whether it is AI or Guardant360 blood testing or whatever. The days of the skilled morphologist are numbered.

If AI disrupts pathology, it'll disrupt every other field in medicine. Don't think it'll happen in my lifetime at least.

 

[Dro133]

Interesting perspective on this topic from Andrew Beck, a Stanford-trained pathologist who left his assistant professor position at Harvard to start up a company called PathAI. In summary, it sounds like he thinks some of the diagnostic tasks of pathology could be done by AI in the future, but that pathologists will still be in demand for the value they can provide to patient care in other ways.

His part starts at ~23:15

 

[WEBB PINKERTON]

I gave the stock tip for Nanostring Technologies on this site like a month ago or so. If you invested in it, you are making money and getting closer to retirement. At least make some money on the companies that are threatening to make us obsolete.

 

[coroner]

Doubt it will happen within my lifetime or career (next 20-30 years), so who cares?

Don't think it'll happen in my lifetime at least.

I agree and have stated this in the past as well. My point was to illustrate current events relevant to our field instead of the usual job market or 'Which residency programs are better?' threads...

I agree, technology will render us obsolete. But it won't happen in your lifetime or mine...

As far as the second, I think the statement towards the end of the article sums it up:
'For now, much of the talk is theoretical — the company is years away from selling a finished product, with a staff of fewer than 20 employees.'
This goes back to what I've said in the past. Is it inevitable: Yes. Is it going to make our skillset irrelevant in this lifetime: No. Keep calm, carry on...

Get as many sources of income going as you can once you are out practicing. Rental properties, businesses etc.

No one thought cytopath would die off as fast as it did either. AI (automated screeners) and molecular have reduced it to rubble.

There is a lot of disruptive tech on the way. Whether it is AI or Guardant360 blood testing or whatever. The days of the skilled morphologist are numbered.

^^This^^ Pathologists will be still needed in the future, but their defined roles and in what capacity they serve will be certainly be different. Medical staff directorship and attending meetings will certainly be there. But, even for things that once needed "boots on the ground" i.e. frozen sections in rural community hospitals may no longer be the case. Also, reimbursement could significantly change. If the machine is doing the interpretation, that could kill off PC for the pathologist as we know it...

Webb laid it out pretty simply: Diversify investments and don't depend on this field for sole means of income. Pre-meds and wannabe pathologists take note...maybe a field like Psych has better job security where there's still human interaction. Then again, if you've seen the movie 'her'...

 

[WEBB PINKERTON]

AI will free pathologists up to do other things....like look for a new job.

I remember when the AI pap screening machines came out in the mid 2000s, they were constantly saying that it would free cytoscreeners up to do other job responsibilities. Well it freed them up all right as there were many layoffs and consolidation.

 

[path24]

I give it under 20 years. At first they will need us and then a lot of us will go away. New job time. Grossing will be the last to go so maybe all day at the bench with a rough pay cut.

They say the trucking industry has <10 years.

AI is going to hurt a ton of people. When it will hurt your industry/job is the big question.

 

[WEBB PINKERTON]

The trucking industry will not change (in my opinion). I don't see a future of autonomous big rigs on our interstates. I don't even think consumers really want autonomous cars. Driver aids will continue to grow but full autonomy seems unlikely.

I recently had my hair cut at a fully automated barber shop. I sat in the chair and said the style I wanted and the machine did it. No barber, no small talk. It was an amazing experience.

 

[WEBB PINKERTON]

I agree and have stated this in the past as well. My point was to illustrate current events relevant to our field instead of the usual job market or 'Which residency programs are better?' threads...










^^This^^ Pathologists will be still needed in the future, but their defined roles and in what capacity they serve will be certainly be different. Medical staff directorship and attending meetings will certainly be there. But, even for things that once needed "boots on the ground" i.e. frozen sections in rural community hospitals may no longer be the case. Also, reimbursement could significantly change. If the machine is doing the interpretation, that could kill off PC for the pathologist as we know it...

Webb laid it out pretty simply: Diversify investments and don't depend on this field for sole means of income. Pre-meds and wannabe pathologists take note...maybe a field like Psych has better job security where there's still human interaction. Then again, if you've seen the movie 'her'...

When AI came to cytopathology in the mid 2000s with automated screeners it brought with it a different CPT code that paid more BUT guess who got the extra money? The referring physicians of course. You found yourself with the added cost of this AI equipment that included special AI stains and you could barely (if at all) raise your client price due to all the competition. AI equaled even higher productivity and lower margins for cytopathology.

 

[Alteran]

I’m not worried for one reason only. When was the last time anyone trusted software, and software alone, for high risk jobs? Is any company willing to be on the hook for a billion dollar plus class action suit if this AI algorithm starts to miss diagnoses? Can you imagine the kind of insurance policy that this company would have to take out if pathologists are cut out of the equation?

The American public at large will never accept their health care decisions to be left soley up to a computer. And certainly no one wants to be the test subject either.

 

[coroner]

I’m not worried for one reason only. When was the last time anyone trusted software, and software alone, for high risk jobs? Is any company willing to be on the hook for a billion dollar plus class action suit if this AI algorithm starts to miss diagnoses? Can you imagine the kind of insurance policy that this company would have to take out if pathologists are cut out of the equation?

The American public at large will never accept their health care decisions to be left soley up to a computer. And certainly no one wants to be the test subject either.

People also said a computer will never be able to beat Bobby Fischer (former world's greatest chess player) one-on-one...until Deep Blue did. Now, supercomputers don't even play the grandmasters anymore because it's accepted they're better. I understand we're dealing with patient's lives and not board games, but A.I. will eventually be more advanced (if it isn't already in some aspects) than even the best morphologists at some point. Given time, there will be plenty of articles and double-blinded randomized control trials to prove that and the literature will back it up. Once the evidence is there, the question becomes: Does society and health care professionals accept a pathologic dx. with 99.9% accuracy from a machine vs 99% from a human? I agree that pathologists won't be completely cut out of the equation, but our current roles, what capacity we serve in, and reimbursement will probably change significantly. Going back to the chess example; and, knowing how advanced the evolution of supercomputers & A.I. have become, consider this: In a best of 100 game matchup to win the most # of games, are you gonna pick the supercomputer or the #1 ranked chess player in the world? Just like in 20 years when (not if) A.I. gets to that point are you (and more importantly the pt.) gonna pick the computer to nail the dx. or the "world's greatest" pathologist?

 

[cmz]

Didn't Deep Blue beat Gary Kasparov? I am pretty sure Bobby Fischer was huddled somewhere in Iceland writing about his love for the Jewish people.

Anyway, regarding AI, I don't think we will see AI become autonomous until the game truly changes and some new diagnostic method comes out that shatters what we've spent years learning and even longer practicing. In the meantime, AI will be pathologist's new best friend helping them kick cases out faster and with more accuracy.

How else are we going to cope with the looming pathologist shortage in the years to come? Actually, in my neck of the woods, I can safely say that there is indeed a shortage of pathologists (with quite a few that need to be replaced). I think if you live anywhere not named Dallas, San Diego, etc you'll probably agree with me.

 

[deleted314957]

It’s gonna happen. Just over the course of my 67 years the world has gone from
primitive computers and pre-commercial jetliner travel. I mean, transistor radios
were the hot thing when I was a teen. Technology will drastically change our field in ways we cannot even imagine today.

 

[clinicallycorrelated]

It’s gonna happen. Just over the course of my 67 years the world has gone from
primitive computers and pre-commercial jetliner travel. I mean, transistor radios
were the hot thing when I was a teen. Technology will drastically change our field in ways we cannot even imagine today.

It won’t happen. I wouldn’t expect boomers to understand technology so I don’t fault you, but no.

 

[coroner]

It won’t happen. I wouldn’t expect boomers to understand technology so I don’t fault you, but no.

By saying, "it won't happen", you're saying technology won't drastically change our field in response to mike's statement? Care to elaborate since you seem to have such a profound understanding of technology that baby boomers and the rest of us don't?

 

[path24]

How does this not happen? Timing is very debatable. Really hoping it takes 20 years. GI biopsies seem like really low hanging fruit.

Trucking industry. Flyover country is already wide open for automated trucks to run on the interstates. Based on my very limited knowledge, silicon valley is 99% there.

 

[WEBB PINKERTON]

By saying, "it won't happen", you're saying technology won't drastically change our field in response to mike's statement? Care to elaborate since you seem to have such a profound understanding of technology that baby boomers and the rest of us don't?

Canada is far behind the rest of the world in innovation due to government interference. That country will be many decades behind the US unless they get rid of the leftists.

 

[WEBB PINKERTON]

How does this not happen? Timing is very debatable. Really hoping it takes 20 years. GI biopsies seem like really low hanging fruit.

Trucking industry. Flyover country is already wide open for automated trucks to run on the interstates. Based on my very limited knowledge, silicon valley is 99% there.

They won't go to a driverless big rig if that is what you meant by the trucking industry having less than 10 years left. They will require someone to ride along. Job will become more like a pilot.

AI shouldn't be number 1 on your list of worries about the future of pathology. If it is like cytopathology AI, it will still require human eyes to make decisions. But it will lead to higher productivity demands and smaller margins that are made up for by all that productivity. At least that has how it has played out in cytopath.

 

[Guest8]

The trucking industry will not change (in my opinion). I don't see a future of autonomous big rigs on our interstates. I don't even think consumers really want autonomous cars. Driver aids will continue to grow but full autonomy seems unlikely.

I recently had my hair cut at a fully automated barber shop. I sat in the chair and said the style I wanted and the machine did it. No barber, no small talk. It was an amazing experience.

Disagree on a couple points. Long-haul trucking will most definitely go driverless. The last mile will still require drivers, but the interstate part will not. It will allow much longer drives (no stopping for rest) and it's mindless and tedious already, and highway driving is by far the easiest part for autonomous vehicles. The autonomous trucks will drive between major warehouses next to or on highways, and then drivers will carry the load the last short distance. Absolutely will happen.

And certain consumers definitely want autonomous cars. Not to buy (there's no point in owning an autonomous vehicle), but to follow the Uber/Lyft model. They'll start in cities and slowly spread outward. Will people in rural states with long drives use them? No. Will urban dwellers who already take Uber/Lyft everywhere use them? Absolutely 100%.

 

[ScubaV]

I for one, can't wait for automated cars to take over and get all those crappy drivers off the road. Of course, I am an excellent driver, but the rest must be purged.

 

[y2k_free_radical]

Didn't Deep Blue beat Gary Kasparov? I am pretty sure Bobby Fischer was huddled somewhere in Iceland writing about his love for the Jewish people.

Anyway, regarding AI, I don't think we will see AI become autonomous until the game truly changes and some new diagnostic method comes out that shatters what we've spent years learning and even longer practicing. In the meantime, AI will be pathologist's new best friend helping them kick cases out faster and with more accuracy.

How else are we going to cope with the looming pathologist shortage in the years to come? Actually, in my neck of the woods, I can safely say that there is indeed a shortage of pathologists (with quite a few that need to be replaced). I think if you live anywhere not named Dallas, San Diego, etc you'll probably agree with me.

Where are these locales that fulfill the long prophesied promise of looming shortages

 

[WEBB PINKERTON]

Disagree on a couple points. Long-haul trucking will most definitely go driverless. The last mile will still require drivers, but the interstate part will not. It will allow much longer drives (no stopping for rest) and it's mindless and tedious already, and highway driving is by far the easiest part for autonomous vehicles. The autonomous trucks will drive between major warehouses next to or on highways, and then drivers will carry the load the last short distance. Absolutely will happen.

And certain consumers definitely want autonomous cars. Not to buy (there's no point in owning an autonomous vehicle), but to follow the Uber/Lyft model. They'll start in cities and slowly spread outward. Will people in rural states with long drives use them? No. Will urban dwellers who already take Uber/Lyft everywhere use them? Absolutely 100%.

They are using autonomous trucks in some places BUT they are requiring a driver be present. There will be laws passed that won't allow driverless big rigs on the interstate. Driverless vehicles seem like an easy target for criminals as well. They will be losing load after load.

I just checked out the student trucker network website and there is a thread on there titled "flee trucking now" that discusses all of this.

 

[clinicallycorrelated]

Canada is far behind the rest of the world in innovation due to government interference. That country will be many decades behind the US unless they get rid of the leftists.

Ehh, don’t know where you got that from. You guys have a high volume which is why we take your fellowship positions, but nobody who goes down south thinks medicine is more advanced, let alone other services. Medicine does have more money to waste, but at the detriment of the people.

Can’t wait till my socialist millennial generation takes power in the US. Y’all sound like you’re straight from the_Donald

 

[Adrian Cocot]

Reading Life 3.0 right now. It's a must if you want to get a better idea of the academic view of AI. A general A.I. that can read biopsies from multiple organs, and that clears the red tape for prospective use, is at least 20 years away. And it'll only take one lawsuit for a machine-based misdiagnoses to upend things (who's responsible? The developer? What if it's a recursive self-improving AI? Should machines have malpractice insurance?). As always, radiology will be our canary in the coal mine. Hedge your bets by buying LEAP options on the new AI startups.

But I digress. The big money in AI isn't for diagnostics, but for algorithms that will minimize -- or entirely replace -- invasive procedures. To that end, here's a contrarian opinion: with trends going towards progressively smaller tissue samples, combining dx and tx into a single modality, and preferring tissue for molecular testing, cytopathology will actually become an integral part of the job, especially in the inpatient setting. First time I heard someone say "cyto is a dead field in 10 years" it was some hotshot at a CAP meeting back in 2007. That dude's gone now, but the field is still around.

 

[Guest8]

They are using autonomous trucks in some places BUT they are requiring a driver be present. There will be laws passed that won't allow driverless big rigs on the interstate. Driverless vehicles seem like an easy target for criminals as well. They will be losing load after load.

I just checked out the student trucker network website and there is a thread on there titled "flee trucking now" that discusses all of this.

Didn't say it is happening now, but give it 10 years or so. Interstate long-haul trucking is ripe for autonomous vehicles - it's near-mindless, and the driver is one of the biggest costs of the shipping involved. Easier to find ways to make the shipments harder to steal than to keep paying >$80K/yr plus benefits to truckers.

 

[WEBB PINKERTON]

Didn't say it is happening now, but give it 10 years or so. Interstate long-haul trucking is ripe for autonomous vehicles - it's near-mindless, and the driver is one of the biggest costs of the shipping involved. Easier to find ways to make the shipments harder to steal than to keep paying >$80K/yr plus benefits to truckers.

It could making being a trucker a much healthier job. Activate the co-pilot for long stretches while you sleep or workout or whatever. Will states adopt the same laws? You could find yourself needing a driver for stretches in states that are more stringent. There is so much to work out with this.

If there was no one on board, I could see us reverting back to the wild west with trucks being robbed of their cargo. Driving down the interstate seeing a group of dudes breaking into the trailer could be a common sight.

Definitely will need to hire a lot of actuaries.

 

[ScubaV]

I imagine there will eventually be AI to handle much of the highway driving with humans able to step in via remote control as needed for construction/traffic/local roads similar to how military drones are operated remotely.

As for robbery, I can't imagine this is a significant reason for keeping human drivers. How much do humans prevent theft right now? Thieves probably either hit the truck when it's unoccupied or pack enough manpower/firepower that one or two drivers, even armed, are not going to find it worth trying to protect someone else's (likely insured if it's worth stealing) cargo with their own lives. I'd guess most thefts anyway are inside jobs, as who else is going to know what truck has valuables and what's filled with mattresses?

Hell, getting rid of human drivers probably cuts down on theft as the truck never has to stop except to refuel or recharge batteries (this is the future) and even that can take place at secure non-public facilities if the transport companies want to set them up. Unless, of course, you're imagining Gone in 60 Seconds or The Fast and the Furious happening every day on the highway, in which case you counter with terminators in the cab instead of flimsy flesh and blood humans.

Damn, I should flee pathology now and start up an automated trucking business. Who wants to give me some starter capital?

 

[y2k_free_radical]

I gave the stock tip for Nanostring Technologies on this site like a month ago or so. If you invested in it, you are making money and getting closer to retirement. At least make some money on the companies that are threatening to make us obsolete.

I bought some as well as GH at the same time-thanks for the recommendation

 

[pathstudent]

100% AI will be better at pattern recognition than a human being. At first we will just confirm its findings but eventually it will work autonomously. AI doesn’t have a mind that can drift and miss the metastatic carcinoma in the lymph node. It doesn’t get tired. It can work all night long. It doesn’t have kids or spouses distracting them. It doesn’t come in hung over. It doesn’t need vacation or paid cme. And surgeons will adjust to the change in relationship. It will be just like when they send for a CBC.

Radiology is going down too. Maybe even interventional rads.

Daphne Koller - Stanford PhD in Computer science; former chief computing officer at Google’s A.I. health initiative; and, founder of online learning platforms Corsera and Insitro, was on CNN’s ‘GPS’ with Fareed Zakaria discussing how A.I. can impact the future of healthcare.

Excerpts from the interview pertinent to our field are below.

Link to the podcast: Bill Gates & Tax Rates Fareed Zakaria GPS podcast


Zakaria: What’s the most obvious place machine learning can impact healthcare?
Koller: The most obvious application for machine learning in healthcare is on the diagnostic side. When it (A.I.) looks at a scan, etc. and predicts what the pt. has (dx.). This includes multiple sources of data including an X-ray scan, a pathology slide, or a liquid biopsy and what fragments of DNA are found in the blood.

Z: Why would the computer be better than the great doctor who’s seen a bunch of these same images and tell what the patterns look like?
K: What’s surprising is on the imaging side where we think human beings have been looking at these for over a century, why would a computer be able to do better? Computers are able to discern subtle patterns in these images that are very hard for people to see, partly because they can look at so many different samples and extract commonalities people just can’t hold in their head.
You look at a cancer pathology which has thousands of cells and some of them are aberrant that way and others are aberrant this way, how do you put all of that together into something that corresponds to the diagnosis? But it becomes more relevant when looking at liquid biopsies when you’re looking at very subtle changes in the composition of DNA, people have no idea how to look at that data.

Zakaria concluded by saying, “I think everybody understands the computer can look at images with much greater detail and accuracy than human beings…”

I always said that it’s not a question of “If..”, but “When…” AP in its current way of being practiced will become obsolete. I guesstimated we’ve got about 50 years left. Now, I’m not so sure it’s even that much…
View attachment 251034

 

[WEBB PINKERTON]

Looking at slides aint the future of this field anyways. The work we do now seems archaic and I question the utility of many things we routinely do. The job of the pathologist is going to be eroded by companies like Guardant Health, Foundation One, Nano-string etc. We are seeing it now.

 

[Nilf]

No wonder any respected pathologist laughs at this forum. All you talk about is how “everything is going down” sheesh. Not even going to address how asinine your comments are.

Students, not worth your time on these depressed trolls and baby boomers. They’re out of touch with reality.

Judging by your comments you were in middle school at the most when I first started reading this forum 15 years ago. Over this time I was able to glean a lot of precious insights here, which had impact on long-term choices in my career planning. I came to know and respect opinions of many posters with far more experience than me, the veil of anonimity notwithstanding.

Concerning the threat of AI: yes it's real, but its is far far far less worrisome than many other threats which are facing the field. For example it might make it even easier for clinicians to skim easy cases and leave us with the more complicated ones. I do envision a scenario in which slides are scanned, AI screens them, sorts them out, then a clinician or even tech signs out the tubular adenomas, prostate cancers, basals and sebs, and leaves the rest for the pathologists.

 

[deleted314957]

Wow Nilf, the tragedy is that you are right. An MD/DO can legally do whatever they want. Now, the neurosurgeons don’t have to worry very much but I guaran-damn-tee
you the derms and GI’s would be on us like a bum on a ham sandwich. If I were advising their profession i’d tell them to push more in this direction and less in the lab per-se.

 

[Torsed]

Could the AI have caught that ureter in the colon resection I grossed last week? No PA would have found 33 nodes in a divertulitis resection, guess who found a bonus.

Heck no. Gimmie my medal now. Or just my crappy 88307 charge.

Get outta here AI.

 

[clinicallycorrelated]

Could the AI have caught that ureter in the colon resection I grossed last week? No PA would have found 33 nodes in a divertulitis resection, guess who found a bonus.

Heck no. Gimmie my medal now. Or just my crappy 88307 charge.

Get outta here AI.

I didn’t do path to gross specimens. I do think a PA could do all that and more it’s not rocket science

 

[deleted314957]

Could the AI have caught that ureter in the colon resection I grossed last week? No PA would have found 33 nodes in a divertulitis resection, guess who found a bonus.

Heck no. Gimmie my medal now. Or just my crappy 88307 charge.

Get outta here AI.

Wow! Ain’t seen a ureter with colon ina long time. No request for urohelp w stent?
Resident case, belly full of tumor, blood feces? What a dumpster fire

 

[dukeresident]

AI in medicine...

How IBM Watson Overpromised and Underdelivered on AI Health Care

After its triumph on Jeopardy!, IBM’s AI seemed poised to revolutionize medicine. Doctors are still waiting

spectrum.ieee.org