IBM’s Watson Health is sold off in parts • Hacker News Board

Re: IBM’s Watson Health is sold off in parts

#411

Post by saltminer » Sun, Jan 23, 2022, 7:22 PM UTC

HiJon89 wrote: post #160

mnd999 wrote: post #81

Earlier quoted context omitted.

Self driving cars are just that, marketing. Tesla for all their hype have limited driver assist only and even Google who seem to have the most advanced offering only works in limited areas in good conditions and the dataset that drives it requires lots of maintenance. The general problem is too hard, and general practice has some of the same problems. Probably less adversarial data, but there’s still litigation to be…

Limited driver assist? Are all the videos of people using Tesla full self-driving on YouTube fake?

It's unpredictable enough that it should be considered just "limited driver assist". Only a fool would look at https://www.youtube.com/watch?v=2ub2F-UnXIU and go "yeah, that's full self-driving, all right." Slapping a "beta" label on it means nothing - you could be talking about how Gmail was in beta, or you could be talking about Fallout 76.

Tesla's marketing is extremely disingenuous, IMO. The name of your product creates expectations in peoples' minds, and sure, you can absolve yourself of liability by putting in the fine print "this isn't anywhere close to true FSD and your car might not be powerful enough to support it by the time we get there, so you gotta keep your hands on the wheel," but that doesn't make it right.

If they called it "advanced driver assist" or something similar, I'd be fine with it (it is more advanced than traditional driver assistance tools like cruise control and lane departure warnings, after all). But I doubt they could get people to pay $10k or whatever the current price is if they were more honest. Instead, they would prefer to earn more money by slapping that FSD label on it and letting people immediately turn their brains off.

Re: IBM’s Watson Health is sold off in parts

#412

Post by james-redwood » Sun, Jan 23, 2022, 7:22 PM UTC

cube00 wrote: post #30

saxonww wrote: post #14

I really expected that we'd see a change in my lifetime, that GPs in particular would be replaced by a lower-cost Watson descendant, with there being some other role for patient interaction, wet work, and data entry (perhaps just nurses). My mom worked for a GP for about 20 years, and it seemed to me that most of what made that guy a doctor was bedside manner + being able to remember a lot of things. But GPs often ma…

> But GPs often make astounding amounts of money while leaning heavily on their staff to actually handle patients and keep the business running. The private ones might the ones in the public system at least in my country need to see a patient every 10 minutes to keep their heads above water.

This is more or less the case in many European countries. Less so here in the UK, but approaching there rather quickly due to astoundingly poor management by various successive governments. It’s tough back in South Africa as well, and they pay their doctors extraordinarily well.

Re: IBM’s Watson Health is sold off in parts

#413

Post by baobabKoodaa » Sun, Jan 23, 2022, 7:22 PM UTC

colinmhayes wrote: post #225

donkarma wrote: post #101

Earlier quoted context omitted.

you think YouTube is profitable?

Yes, I think youtube is making billions of dollars a year. Youtube's revenue was $25 billion last year. It's a mature product. You don't think it's making money?

Revenue is not profit.

Re: IBM’s Watson Health is sold off in parts

#414

Post by brendoelfrendo » Sun, Jan 23, 2022, 7:23 PM UTC

mardifoufs wrote: post #311

colinmhayes wrote: post #225

Earlier quoted context omitted.

Yes, I think youtube is making billions of dollars a year. Youtube's revenue was $25 billion last year. It's a mature product. You don't think it's making money?

Google still doesn't release the profit figures for youtube, only revenue. Even the revenue figure was a secret until 2--3 years ago, so it was probably not that high. If youtube was very profitable, I'm not sure why Google would hide it in its earnings since they almost always try to show how they aren't exclusively dependent on their search ad business. I'm not saying youtube does not make any profit at this point,…

I mean, the opposite is also true, isn't it? If Google's search ad revenue was going down while YouTube ad revenue was going up, I feel like Google would want to keep that a secret so that people don't realize that search ads are shrinking in relevance.

Basically, I don't see that Google has any incentive to break out its profits by line of business unless someone forces them to. They're better off if you just look at a black box of ad revenue and say "yeah it's all profitable, so ads are as strong as ever."

Re: IBM’s Watson Health is sold off in parts

#415

Post by spookthesunset » Sun, Jan 23, 2022, 7:23 PM UTC

ed25519FUUU wrote: post #254

sofixa wrote: post #145

Earlier quoted context omitted.

The biggest airline booking system, Amadeus, can actually run on Kubernetes and some airlines are migrating to it. Fun fact: a few years back Amadeus employees were in the top 15 of k8s contributors by company.

Is airline booking software very complex? I don’t see why that software would necessitate the updating of k8s.

Every single airline they ingest probably has a phone book sized list of special cases and edge conditions that are distinct and unique. Multiply that by every airline on the platform and every consumer of the dataset and yeah… it’s incredibly complex.

Not to mention the uptime requirements and other SLAs in place (which probably are all different for each contract they sign)… yup. It’s probably a monster.

Re: IBM’s Watson Health is sold off in parts

#416

Post by erosenbe0 » Sun, Jan 23, 2022, 7:24 PM UTC

ako wrote: post #407

erosenbe0 wrote: post #234

Earlier quoted context omitted.

Mediocre GP makes 200k cash in Omaha or any average city and has choice among many employers and systems. Stanford CS represents top talent and moreover does not make 200k cash in Omaha, and can only make the big bucks at a fixed number of entities in fixed locations with fixed hierarchies and systems.

A freelance GP (waarnemend huisarts) in the Netherlands makes on average 65 euros per hour... It pays better to be a freelance software engineer. Less education required, less responsibilities, impact of mistakes is also usually less.

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 medical school is an excessively long four years which has further periods of time waste and a lack of integral training until late in the process. Then there is a highly unfocused three year residency with what are widely understood as illegal discriminatory labor requirements holding on to their anachronistic 1910 white-male origins by a proverbial thread. So at age 29 you can finally be a GP after many unnecessary years of wasteful, repetitive, unfocused study, vacations and debt. This exclusivity and opportunity hoarding results in a reasonably high skill at a very very unnecessary level of cost and personnel shortages.

Re: IBM’s Watson Health is sold off in parts

#417

Post by ako » Sun, Jan 23, 2022, 7:25 PM UTC

kumarvvr wrote: post #381

cantrememberpw8 wrote: post #185

I'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 ?

Services/products of a company usually bring one of the following to customers: 1) improve revenue, 2) lower costs or 3) manage risks. RedHat is probably mostly about the last one, manage the risks of running linux.

Re: IBM’s Watson Health is sold off in parts

#418

Post by staz » Sun, Jan 23, 2022, 7:25 PM UTC

leokennis wrote: post #397

kumarvvr wrote: post #381

Earlier quoted context omitted.

Curious, what does RedHat actually do ?

They make sure enterprises can run Linux that doesn’t “suddenly” (read: with less than 2-3 years notice) break their critical workflows because some component loses support or some dependency reaches EOL - they do this by extended maintenance, backporting (security) patches to old versions, providing tailored support etc. This is very valuable to enterprises and so they pay a lot for it. For example, you can still ru…

To save you a web search : Red Hat 6 was relased 6 November 2010. Roughly 1 year after Window 7, which ended support 2 weeks ago.

Re: IBM’s Watson Health is sold off in parts

#419

Post by throwawayboise » Sun, Jan 23, 2022, 7:25 PM UTC

saxonww wrote: post #14

I really expected that we'd see a change in my lifetime, that GPs in particular would be replaced by a lower-cost Watson descendant, with there being some other role for patient interaction, wet work, and data entry (perhaps just nurses). My mom worked for a GP for about 20 years, and it seemed to me that most of what made that guy a doctor was bedside manner + being able to remember a lot of things. But GPs often ma…

I mean, I won't even use the self-service kiosk at McDonald's. I sure as hell am not going to interact with a computer for my medical care.

Re: IBM’s Watson Health is sold off in parts

#420

Post by kilburn » Sun, Jan 23, 2022, 7:28 PM UTC

wiz21c wrote: post #24

saxonww wrote: post #14

I really expected that we'd see a change in my lifetime, that GPs in particular would be replaced by a lower-cost Watson descendant, with there being some other role for patient interaction, wet work, and data entry (perhaps just nurses). My mom worked for a GP for about 20 years, and it seemed to me that most of what made that guy a doctor was bedside manner + being able to remember a lot of things. But GPs often ma…

> bedside manner + being able to remember a lot of things My impression is that accompanying a patient is super important, it helps to understand illness, to have a plan in case of more complex treatments, etc. Then my doctor has the ability to know me and gauge my health. She's also very good at probabilities and detecting when something really goes wrong. I'm sure that being able to do that require a lot more than…

What many people don't realize is that medicine as a whole is already some sort of expert system (i.e.: a flavor of AI).

There are researchers that conduct experiments to produce meaningful data and extract conclusions from that data. Then there are expert panels that produce guidelines from the results of that research. Most diagnostics and treatments are prescribed following decision diagrams that doctors themselves call... algorithms!

There are several limitations that prevent us from applying other AI techniques to the problem. Off the top of my head:

- We do not have the technology for machines to capture the contextual and communication nuances that doctors pick up on. There can be a world of difference between the exact same statement given by two different patients or even the same patient in two different situations. Likewise, the effect of a doctors' statement can be quite literally the opposite depending on who the patient is and their state of mind. One of the most important aspects of the GP's job is to handle these differences to achieve the best possible outcomes for their patients.

- Society at large is not ready to trust machines to make such intimately relevant decisions. It is not uncommon for patients to hide relevant information from their doctors, and to blatantly ignore the recommendations from them. This would be many times worse if the doctor part wasn't human.

- We cannot apply modern inference techniques (e.g.: deep learning) to the global problem because we have strict rules that prevent medical data collection and analysis without a clear purpose. Furthermore, these techniques tend to produce unexplainable results -which is unacceptable in this field-. As a result, there's not enough political capital to relax those rules.