IBM’s Watson Health is sold off in parts
461–470 of 711 posts
Re: IBM’s Watson Health is sold off in parts
#462As someone doing a CS degree now, I seem to be the only one who doesn't want to have anything on my resume to do with "AI", blockchain, ML, NFT, chatbots, etc... all I see is overhyped product after product, one-size-fits all solutions that frustrate customers and create problems for humans to clean up, hugely valued companies that have very little real improvement over conventional technology, etc. An "AI chatbot" i…
Re: IBM’s Watson Health is sold off in parts
#463Earlier quoted context omitted.
I generally dismiss these “equivalent outcome” studies. Any midlevel will (and should) bounce the more complicated cases to their supervising physicians. Outcomes at that point are meaningless. There’s definitely a trade off between resources devoted to education vs. acceptable risks from failed procedures, missed/delayed diagnoses, and increased utilization of imaging and referrals (and the physician radiologists an…
> I generally dismiss these “equivalent outcome” studies. Any midlevel will (and should) bounce the more complicated cases to their supervising physicians. Outcomes at that point are meaningless. If midlevels can successfully detect complicated cases to a supervising physician, and handle a whole lot of other care independently... and the net result is equivalent outcomes... this isn't a massive win? You've conserved…
#1 - Funky/misleading statistics - Generally they claim that these NPs with uncomplicated patients do as well as physicians with complicated patients. It's not claiming that of any randomly selected patient, regardless of who they see, the outcome is the same. Therefore, if uncomplicated patients saw physicians, outcomes for the physicians could improve. In primary care managing hypertension or diabetes, this isn't as pertinent. For something like anesthesiology, it's more so counting how many times shit hits the fan, and brain cells die when the anesthesiologist takes time to be summoned.
#2 - They're not conserving expensive resources. Imagine a patient comes in with a lump on their hand. An NP might see a weird lump, order an MRI which gets read by a radiologist, refer to an orthopedic surgeon who specializes in the hand, who removes tissue to send to a pathologist, who determines it's a common benign tumor of the fascia. That's three physicians who spent much more time here! The patient no longer has use of their interphalangeal joints. The physician would probably try to shine a light through it, note the patient's Scandinavian ancestry and family history of plantar fasciitis, and tell them to live with it and come back if it changes.
No resources were saved here, but the patient's DASH score (disability of the arm, shoulder, and hand) is still 0 so the outcomes are the same.
This happens all the time.
#3 - Bad incentives - Medicaid would not in a million years cover this, but the game of medical pinball where patients bounce around through in-network referrals can funnel those with decent insurance into procedures. Especially when most people have poor health literacy. A hospital executive probably just splooged in his pants seeing how much money their loss-leader of primary care is driving to radiology and the surgical specialties where they actually make money.
#4 - It's insincere. All of this can be viewed as possibly successful when the midlevels are part of the healthcare _team_ and know their limitations. But the NP groups are increasingly pushing for independent practice and prescribing rights in state legislatures across the country. CRNAs require a physician supervisor... in many places, that doesn't necessarily need to be an anesthesiologist, and the surgeon performing the procedure can suffice. The AANA recently changed its name to the "American Association of Nurse Anesthesiology"... It used to be "Anesthetists". The CEO and president (two different people) of the American Nurses Association both refer to themselves as "Doctor" in a healthcare setting even though one holds a DNP and the other a PhD. It's pervasive.
Re: IBM’s Watson Health is sold off in parts
#464They are a nightmare. I was part of a huge project to replace a large part of a telecom operators infrastructure. IBM global services ran the operators IT outsourced. The project failed after a year because of them. It was the 3rd such project to fail. The company in question couldn’t bring themselves to realise it had been their outsourced operators fault once again. Even though they had again lost the bid to do sai…
Listen up Accenture, I will drink expensive champagne when you go bust, or get bought.
Re: IBM’s Watson Health is sold off in parts
#465As someone doing a CS degree now, I seem to be the only one who doesn't want to have anything on my resume to do with "AI", blockchain, ML, NFT, chatbots, etc... all I see is overhyped product after product, one-size-fits all solutions that frustrate customers and create problems for humans to clean up, hugely valued companies that have very little real improvement over conventional technology, etc. An "AI chatbot" i…
Re: IBM’s Watson Health is sold off in parts
#466I'm excited by this. I recently left Red Hat for greener pastures. From where I sat, IBM was slowly turning toward wisdom again, having been run aground by its previous few CEOs. I was skeptical when IBM bought Red Hat, but after several years of not screwing it up, I'm pretty hopeful. Now, Krishna is working on streamlining the business and making the rest of IBM more like Red Hat. Splitting off the low performing K…
Curious, what does RedHat actually do ?
Lots of well-hated projects come from Red Hat: systemd, wayland, ... but they have also contributed well to some other projects which are much less controversial.
Re: IBM’s Watson Health is sold off in parts
#467Earlier quoted context omitted.
What data was it using again? The blood panel your doctor orders every one to two years? Medical diagnosis is a trivial problem for machine learning to beat humans. Humans are terrible at this. The problem is there is not even the concept of high frequency medical data. Imagine machine learning in quant stock trading with samples once a year. Of course it isn't going to work. The problem is all in the externalities.…
I agree with this so strongly. The medical gatekeeping is absolutely absurd. I can't remember the last time I learned something from a doctor that I hadn't already learned myself with a quick online search. For every visit in the past decade or longer they either told me what I already knew or we were both stumped. If drives me crazy that I can't make my own decisions about my own health, even for trivial cases. I've…
If I had the ability to order my own tests, blood work, and adjust the dosages of my meds I could do it a hell of a lot better than the doctor I see every 6-12 months. Antibiotics are the some of the worst with this. The doc literally doesn’t even look at me, I just describe my symptoms and they go “yep sounds like an infection” — like Gods almighty you could be a web form. And it would would actually cost my insurance less which is even more infuriating.
Specialists I have found are actually useful and so I can’t really bring myself to hate that Tier 1 MDs act as a screen for people whose time is valuable.
Re: IBM’s Watson Health is sold off in parts
#468Earlier quoted context omitted.
USA has an extremely inefficient system of training for doctors which wastes literally years of human potential. First you go to university from 18-22 but not even 2/3rds of the year. The rest of the year is some level of waste for most students. You often purposefully dumb down and study less rigorous subjects than an engineer because you need all top grades to be get into medical school and can't risk it. Then medi…
Dutch system is quite similar: 3 years bachelor, 3 years masters, including 18 months coschappen (residency), and then 3 years of specialization to GP (other specializations take longer). ( https://universitaire.bachelors.nl/faq/ik-wil-dokter-worden/ ) Many specializations come with very low job opportunities, being a GP is very stressfull: low paid assembly line type of work, a GP is expected to see 6 patient per ho…
Re: IBM’s Watson Health is sold off in parts
#469As someone doing a CS degree now, I seem to be the only one who doesn't want to have anything on my resume to do with "AI", blockchain, ML, NFT, chatbots, etc... all I see is overhyped product after product, one-size-fits all solutions that frustrate customers and create problems for humans to clean up, hugely valued companies that have very little real improvement over conventional technology, etc. An "AI chatbot" i…
+ People care about what other people are talking about. They like to fit in, like they're part of the cutting-edge.
+ Less experienced people have less…experience with the downsides of what they're reading about.
+ CS is no longer mostly people who care about computer science, in the same way that economics isn't only for people who want the understand economics. Tech salaries — especially engineers' — are super high, like investment bankers. So people study the respective fields as a means to an end.
+ Twitter is driven by VCs, tech press, and people marketing themselves. They're work themselves into circular frenzies all the time. Little of it matters. Almost none of them have any record of predicting what's next and a long, long record of being wrong. This is true of most people! But these are the spaces many people look to to see what is "wanted".
You seem to have good instincts. Don't be distracted by peers who work at "hot" startups or big named companies. Find something you believe should actually exist in the world and work on that. It will give you an intrinsic reward that money can't buy and status can't fill.
Re: IBM’s Watson Health is sold off in parts
#470What's the chances that we'll have robotic domestic servants before 2032? With AlphaZero, AlphaFold and sort-of OK machine translation, I think it's 50:50, no?