February 20, 2023: David Ting, CTO at Tausight joins Bill for the news. There is a constant flow of new technology in health care. From Tausight’s new PHI awareness meter to ChatGPT, the growth and innovation does not stop. What are the repercussions of these modernizations? How can policy assist in the big issues such as cyber security or plagiarism? How can the healthcare industry become more resilient in the face of these obstacles?
Healthcare needs innovative ways to address staffing shortages from clinical to IT employees. Are you curious about how technology can help support your Healthcare staff? Join us on our March 9 webinar, “Leaders Series: The Changing Nature of Work,” to explore how Health IT can be used to supplement Healthcare professionals - https://thisweekhealth.com/leaders-series-the-changing-nature-of-work-remote-hybrid-onsite-shortages/
Alex’s Lemonade Stand: Foundation for Childhood Cancer Donate
This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.
Today on This Week Health.
The new technologies are changing how we need to think about technology evolution. It's gonna change our workflows. It's gonna change how we look at things. Computing is so pervasive and so cheap these days. When I was in graduate school doing this kind of work, it was impossible to get access to this much computing power.
Welcome to Newsday a this week Health Newsroom Show. My name is Bill Russell. I'm a former CIO for a 16 hospital system and creator of this week health, A set of channels dedicated to keeping health IT staff current and engaged. For five years, we've been making podcasts that amplify great thinking to propel healthcare forward. Special thanks to our Newsday show partners, Cedars-Sinai Accelerator, Clearsense, CrowdStrike, Digital Scientists, Optimum, Pure Storage, SureTest, Tausight and VMware for investing in our mission to develop the next generation of health leaders. Now onto our show.
Alright, it's Newsday and today we are joined by David Ting with Tausight and essentially a technologist. And the reason I give that background as a technologist, not only a founder of many companies, but a technologist is today we're gonna talk emerging tech.
We're gonna talk cybersecurity, and we're gonna talk just a lot about the technology and the things that are just around the corner. David, welcome back to the show.
Thank you, bill. Great to be back.
Well, I'm looking forward to it. I I was talking with somebody from Tausight today.
It sounds like you guys released some stuff and it's out in the wild and yes. And a bunch of people are picking it up. So a lot of things going on over there.
Yes. We're excited. We've introduced uh, what we call Phil, the p i locator product is for a lot of our pilots so that they can go out.
Point the tool to some file shares and databases and,
So, so I can get this tool for free and essentially point it at something and yes get information back,
give you the information on what files have phi, what was the creation date, what was the last use date, what's the owner, what the access rates.
All of a sudden you can start to create that map of where are all my phi. Lurking across my system. It's a single machine. You just point it and it starts to scrub the database or your file shares. I'm sorry
Yeah. When I was cio, we didn't have that kind of technology, and so people used to always ask the question, what keeps you up at night? I'm like, that's what keeps me up at night. I don't know where all my information is. And that's, it's crazy.
It's a hard problem, especially now that we digitized all the information and a lot of medical content tends to be in correspondence, which is unstructured.
So is you have to basically start to understand the nature of the document. The information that's trying to be conveyed. And so we're leveraging natural language processing, AI technology, deep learning technologies to say, how do you have a trained person sitting in that engine that has the intel and the ability to decipher this content regardless of what type it.
Well, I'm glad you're on the show. Clearly, we're gonna talk Google Bar and chat, g p t. . But we're also gonna talk there's a m I T article, which I like, and and you agreed to talk through and
I can't talk about all of them. Like crispr is not my background. Well, I know
We'll mention 'em, we'll dive into some of 'em are pretty, pretty interesting.
And I think I can't talk about all of 'em either. I mean, these. These are really cutting edge technologies, but let's start with chat, G B T and Google Bard. All right, just so people get context, we're recording on the ninth and yesterday Google Bard goes live. give some erroneous answers and Google loses a hundred.
Was it a hundred billion in market cap? It was
a lot. It was a lot for this. Yes.
But you know, so we have a lot going on right now. You have Bard was released as a competitor. Microsoft integrated into Bing. I went over to Bang I, I'll tell you, I think it's the first time I've been to Bing. A couple years, and right there on the right side is chat, G p T, right?
And so now you get not only the listing of sites that you would normally get at a Google search or a ping search, but you also get, Hey, I noticed you asked this kind of question. Would you like me to answer it for you? So thanks Gary . Well, it is a little bit, and that's sort of what I want to ask you, what makes.
Chatgpt smarter than Google? Bard or Google Bard smarter than Chatgpt What's gonna be the differentiator here?
I think the differentiation's going to be the type of information they've trained it on and how much time they have had to refine it.
I mean, here's an example. You we tested it just to say, can we write FAQs with the thing? We ask it a bunch of questions about the stuff that we're working on. It comes back with reasonably good answers. I said, gee, these FAQ responses are as good as the ones we would hand write and handcraft. Yeah.
Then out of the blue we said, let's ask you to write a program in c plus. And my lead engineer and I are watching this thing come back with a fully annotated c plus program that ran the first time. And I said, okay. Had we had more time and more access to ChatGPT which is actually pretty hard because there's so many people are pounding on it, right?
You have these very limited windows to of use, but there was an article I read that said, here's five ways the software developers can use it to improve their code, one of which is understand spaghetti code. It literally went through and documented all the code. It went and improved on it, and I said, and one of my software engineers basically said, The sky's falling.
They're not gonna need me in the future. .
Well, that's one of the stories that was just released it's interesting, Google put chat, g p t through it's hiring process and essentially came back and hired it at a level three that has a starting salary of like $189,000 a year.
Now granted it's not a human, it's not. Generative per se. One of the things I think that makes it distinct and that I've heard that makes it distinct is they trained it on code. Yes. It's like it, because they trained it on code, code itself is inherently logical.
And so they almost can't explain it, but the algorithms behind it now are logical. It thinks more logically because it was trained on code. And the code examples are so fascinating to me. I
incredibly fascinating because the challenge has always been one of your developers leaves and you have this big spaghetti ball of code that's been maintained, modified by many people and nobody's left to maintain it.
What do you do with it? And uh, there are programs out there that'll help you entangle it, but it takes a lot. Human intelligence to navigate the lines. Well,
programming might be different. I was postulating on this yesterday with somebody. I said, programming might be not me sitting in front of my keyboard, but me talking.
To chat g p t and say, yeah, no, not those variables. And yeah, let's pass it in this way. Or, and it just, because if you do refine, I don't know if you did this, but it gives you the code and then you refine it and then you say, ah, I want that in C Sharp. I don't want it in, it'll then translate it and put it all
in C So you, we end up not trying to figure out the.
Mechanics of programming, but figuring out what the intent of that program should be. We end up defining the intent and software gets written to meet that intent, that des desire, that specification. If you all, I think that's where it has to go,
is does that make a whole new class of programmer a different programmer than we've had in the past?
I think so, yeah. That's, Engineering used to be translating design into blueprints, into specifications, building prototypes, trying it, evaluating where it breaks, and going back and iterating it. We've done that with cad cam to improve on it so that it's more streamlined. I think software has been the last of these engineering professions that we still hand chisel, code line by line, character by character.
Tools like these. AI technologies will definitely change how we write software.
Well as a cto. Are you thinking about, okay, here's how we're going to bring this into our company?
Yes. That's why we were testing it.
We were testing it to say, how do I help with the standard stuff, manually filling out and writing 40 FAQs. But it's that iterative refinement model, which is, I pose a question. It comes back with the response. You ask it to, you get what you ask for, right? You keep changing the parameters and it keeps iterating through.
One of the things in chat, G p T and most AI machines is there's a degree of randomness that they add in. So if you ask the same question twice, it'll slightly give you variance because like most neuronets , there's a random component built in. Right. So it will improve and it will also, like you say, based on historical evidence, changed the parameters because it's starting to understand what you're asking for.
Yeah. It's it's really fascinating. One of the things we will do is we'll take this transcript of you and I talking. We will drop it in and essentially say essentially write the show notes and it'll say, here's a summary of the conversation that, that they just had. And we can say things like, give us five bullet points, give us whatever.
And it, it does that very effectively. And I don't want to alarm people, but we used to have a full-time person doing. All that work for us and we don't want anymore. And that's sort of the fear. But the, I do say that to say I don't think that's what people are thinking it's going to be used for, but I think everybody has to start thinking of their role in terms of upskilling and the next level of, okay.
I'm not the chisler, if you will. I'm now able to see the, the item being chiseled and what part am I gonna play in that
Exactly. We're not out there hammering characters into stones anymore . That evolution has got to go to the path where you start out with better ideas and you know how to refine it.
There are two camps in the schools that are afraid of chat. G P t one will basically say, no, you can't use it at all. And there are schools that are saying, how do we teach our students to be better at using these newer tools? And I think that's the same. Analogy that we're gonna have in many knowledge based industries.
Yeah. And how I explain it to my team from a journalism standpoint, a reporting standpoint, they're like, will that replace, whatever. And I'm like, look you're, you still need investigative journalists. You need people that can. Do footwork and ask people and interview people and ask questions that the technology is nowhere near that.
Right? So there's still aspects of journalism that we will do because it is way beyond where AI is today. Now, AI might get there in 20 years or 30 years. We'll have to I think so. My kids will worry about that. I'm not gonna worry about it
too much. No I completely agree. In my past I helped build editorial systems for newspaper systems.
A lot of the technology that were built in for proofing, editing, improving the cycle of revisions and their approval process and just compacting stuff that's didn't replace the re reporter. The reporter still had come back with the same. Lines of text in the story, but it was the mechanics reducing the typos, the fat fingers, the forgotten.It is:
You can get more information about them at alex's lemonade.org. How can you help? For the month of February, we will be holding a download drive. We're doing a bunch of different drives this year. , and our hope is to raise $50,000 for Alex's Lemonade stand and for February the download drive for every download over 20,000.
And just so you know, our average is roughly about 20,000 every month for this week. Health over the various channels. So what we're gonna do is for every download over 20,000 in the month, February, we're gonna donate $1 to Alex's lemonade stand. So if we get to 25,000, we'll donate 5,000. If we get to 30,000, we'll donate 10,000 to Alex's lemonade stand.
A download is counted as. Anytime someone listens to the episode of this Week, health on either of our channels on the conference channel or on the Newsroom channel let your staff know your peers, whoever you think might benefit from listening to our interviews and this content, and support the work of Alex's Lemonade Stand.. Already raised in:
Go ahead and give a donation. Leave a little note. We'd love to thank you for participating in that and look, it's really easy. Shoot a note to somebody who you think would benefit from listening to this content. And for every download above 20,000 this month, we are gonna give $1 on your behalf.
So we want to thank you for all your sport and help as we try to give back this year. 📍 📍
All right, so let's look at some of the other m i t technology review, m i t technology review. Every year, they pick 10 technologies that will matter most right now. and they have 10 of them. CRISPR for high cholesterol. I'm not gonna go into it. People who know what CRISPR is. Essentially we are modifying it's genetic modification essentially is what we're looking at.
And four, high cholesterol, essentially genetic modification of addressing high cholesterol. That's interesting. AI that makes images.
High TPTs derivative, techno generative technology.
Yeah. Have you seen some of this artwork?
I have. It's crazy. It is. My brother has played with it quite a bit and he's quite impressed.
He said, look at these pictures that I'm generating completely from a verbal
Yeah. It is really fascinating and this is where I think the educational institutions that embrace it will produce a better person who's going out into the workforce because this is a tool that they can.
And utilize I think there is still some gray area in the legal aspect of this. It's consumed all this art. Or all this code, and now it's generatively putting out new art and new code, but it's based on all this stuff that it's seen. If we were to do that as a human, would that be considered plagiarism at all?
I don't know.
there's no plagiarism on ideas yet. We copyright and we patent ideas. Sorry, implementations Ideas what's the legal boundary for that? As these things become the chiselers and the hands on the ideas become important. I don't think our laws are quite there yet to deal with that.
Yeah. Well, and the law has to catch up to a lot of this stuff. Chip designs that change everything. So you have risk five, which is spelled R i s c dash v. And this is a international standard in that they are putting together, that essentially puts puts a set of standards on top of the how chips communicate and how registers change and whatnot.
Why do you think this is a transformative technology?
So, chip design used to be the venue that very large r and d organizations can afford to implement. Risk, the reduced instruction set that people have come up with we're in generation five. I think the sun spark was one of the first generation.
It's a simpler instruction set similar to the arm processors that are used in your mobile phone. It's a simpler design to build The design still has to be generated in order to fabricate. So this standard makes it available for a lot of people to create their own
So that's the power, right? So if it's based on a set of standards, now you've just opened it up to a whole new,
have it to any, it's like generating CADCA tapes to say, I want the thing fabricated by somebody else. Once I've designed it with my home design tool, I can have it fabbed today, I can have a cat can.
Design a cat design, turned it into a machining design, sent out to a fab shop and have something sent back to me, completely machined and done. This is gonna do the same thing. It's gonna open up the avenues for a lot more customized designs, customized processors that will do more than just your general purpose computer and be more efficient at it.
Yeah. Wow. I'm gonna skip a couple of these just cuz we're not gonna go through all 10 organs on demand. So I saw pigs, .
What's that? I saw all those pigs in that backdrop of that image As they're harvesting.
They are, I mean it's 3D printed lungs and other things, but it's also. They are genetically modifying engineering pigs whose organs could be transplanted into humans, essentially.
Makes sense. Yeah. It's it's wild. I, again, not probably in our space of healthcare and technology, but there's two of 'em that have to do with battery technology, and I think this is gonna be one of the biggest things over the. Five years for us. You have the inevitable ev and there's been a, just a massive, huge amount of companies heading in this direction.
But then you also have battery recycling, huge problem. Huge problem. But because there's limited materials, we've now, replaced oil, which exists in a handful of places with a battery. that has some, scarce resources in very hostile places that we have to get to.
So we, we haven't really solved the geopolitical problem of oil. We've just traded it with It's shifted. Shifted, yes. But battery recycling is interesting to me because I've often thought that and I have a Tesla and it's 10 years old. Wow. And I'm getting few fewer miles on it.
And I'm wondering, it's like what happens to these batteries when I'm done? And this is this has to be the direction because we're gonna have a ton of EVs on the road. We're gonna deplete those batteries. Those batteries, those, that material has to be reused somehow. And
that material's environmentally, Challenging both to recycle and to refine.
And so we have to recycle. It takes way less energy, way less impact on the environment to recycle that. Yeah,
if you have yet to hear, we are doing webinars differently. We got your feedback. You wanted us to focus on community generated topics, topics that were relevant to you in your role. We have gone out and gotten the best contributors that we possibly can. They are not product focused. They are only available live.
And we try to have them at a consistent time, the first Thursday of every month with some exceptions. And the next March happens to be that exception. March 2nd, I'm on vacation. So March 9th is going to be our next webinar, March 9th at one o'clock Eastern Time, and we're gonna do a leadership series on the changing nature of work.
We're gonna talk about a couple things. One is the remote distribution of health IT staff and what we have to do from a management standpoint in that regard. We're also gonna talk. The lack of staff specifically in the clinical areas and technicians and whatnot, and what the role healthcare and technology in particular is gonna play.
With regard to that. Love to have you sign up. Our first two webinars for done this year have been fantastic. Over 200 people signing up for each one of them, and we expect just spending for this one. This is a great conversation. Great panelists. We have Tricia Julian Baptist Health System out of Kentucky.
Will Weeder Peace Health and Andy Crowder with Atrium Health are going to join us for this discussion? And I've talked to each of them about this topic and I love their insights and look forward to sharing 'em with you. If you wanna sign up, hit our website, top right hand corner. We always have the next webinar listed.
Just go ahead and sign up, put your question in there and we'll incorporate it into the discussion. Look forward to seeing you then. 📍
📍 📍 All right. Well, I'd be remiss if we didn't talk cybersecurity because you were actually on the task force, so
Correct 4 0 5. Five years ago.
Five years ago.
You still remember all the findings, right? What was that
seven years ago now? ?
Yeah. 20, yeah,
been 20 15, 20 16. Yeah. Almost seven.
Yeah. Yeah. If people wanna refresh her on that. But essentially here's the story. We're looking at how Senator War Warner aims to mitigate healthcare cybersecurity risks through legislation. So Senator Warner spoke with Health IT Security about, and that's where this article is, by the way, health IT Security.
Dot com. There you go. Cybersecurity challenges discussed in his recent policy options paper and how he plans to address them. You've read through this, you're looking at what Senator Warner is talking about while I go back and look up who we interviewed what are your thoughts on Senator Warner's direction on this?go, but now fast forwarded to:
The cost of cyber insurance all are the things that seven years ago in the 4 0 5 committee we discussed, there's still that huge shortage of staff for cybersecurity professionals that we see in this country. The difficulty in having that kind of skill set in healthcare, because again, healthcare doesn't put enough investment into.
Similar to other industries that have had a longer period to mature in their digital journey, healthcare got pushed into it through an incentive plan incentive program. And so now we're starting to see the gee, before you can really leverage the power of digital healthcare information, we have to get better at securing it.
Senator Warner's policy and all the statements that he's making. Really reflect the concern that I and a lot of other professionals have in this space, that we're not gonna be able to extensively use this information if we're constantly being attacked both at the infrastructure level, the. To gain access to the patient information, the lack of privacy of your digital records.
These are the great points that he's pointed out. So if you're getting it, I'm a really big fan of what he's saying. ?
Yeah. You like where he is going? So is Theresa Meadows, cook Children's yourself and Bradford.
Oh yeah, Ray Marsh,
major first, first health advisory. So yeah, we were talking about the the healthcare industry cybersecurity task force, and they identified six key imperatives.
So this, you're saying this was seven years ago, that these imperatives were identified that the imperatives were to find and streamline leadership, governance and expectations. The health industry. Cybersecurity increased the security and resilience of medical devices. And Health. It developed the healthcare workforce capacity.
Necessary to prioritize and ensure cybersecurity awareness and technical capabilities. Number four, increase healthcare industry readiness through improved cybersecurity awareness and education. Number five, identify mechanisms to protect research and development efforts and intellectual property from attacks or exposure.
And number six, improve information sharing of industry threats, risks, and mitigation. I'm actually impressed with the progress on number six. The CISOs I think are one of the most connected groups of people within healthcare. They have mechanisms for communicating when they're under attack and that information seems to flow.
Was that not the case? Seven years ago? That was,
that was still the case, but it's gotten to be more formalized through the information sharing of cyber threats and cyber concern. And cyber incidents. So,
so what can policy do here? So I'm the, by the way I don't disagree with you, but policy sort of got us into this mess too.
So there's always unintended consequences. We have to be careful when we talk about policy. What can policy do well here and where should we where can policy maybe over overstep? So
the 4 0 5 D when it came out and the 4 0 5 actually had talked about. They didn't want it to be a tight, strict regulations with penalties associated with it.
It was basically, how do we make a carrot be the incentive for healthcare to improve their cyber hygiene, improve their posture? So the carrot was a soft one. You get a safe harbor. For any kind of penalties if you have demonstrated for 12 consecutive months that you followed all the best practices as outlined in the 4 0 5 D recommendations.
That's a nice to have. Everybody goes, well, if I don't smoke in my house is probably not gonna burn down. It's. Likely as somebody who's smoking in the house. Those are very soft. What you really need to do is to say, we have set a bar that says you will get incentives to improve your posture if you can demonstrate for sure that you're improving your cyber hygiene, and taking the appropriate steps to improve it actively.
Similar to how we did, created the movement to digitize and move away from paper records. We said, if you can demonstrate you're moving off paper records into an electronic media, we will pay you for it. I think there has to be payment stimulus almost to improve. That's a tough one. I'm not a politician.
I'm. Saying the policies or how the policies should be done, but I think as an industry we're gonna need some sort of an accelerator to get the thing done.
Yeah. And we are under attack. I mean, I was joking with somebody, I'm like, every day I could do a cybersecurity article. Like I do a today show where I just do a 10 minute thing.
I'm like, every day I could do some cybersecurity. , some cybersecurity warning, some, every day. . And if that's the situation we're in, that's essentially we're in a wartime situation on the cybersecurity front. It's hard to get strategic. I mean, you have to continually defend.
Right. And what you're saying is the policy could help us to actually get strategic and to move beyond just defending every day.
I think Senator Warner pointed out that it is basically our critical infrastructure that's under attack. I think you've brought that up as well. It's no different than a national emergency.
We can declare national emergencies and put more attention on this or the largest industry in the US is going to be. In a bad state,
Not to put you on the spot, but I do polls every Monday on LinkedIn. If I were to do a poll and say, which one are you more concerned about?
Our electrical grid, our air transportation system, our healthcare or our financial system, I. They're all pretty high up there, aren't they?
So, having worked in the power system, I could tell you that power grids are a little bit more resilient than we think they are because they've had a longer time to worry about interconnectedness and the ability to isolate and essentially firewall or circuit break, one affected system from another.
Financial systems have been resilient for a long.
So you're saying from a resiliency standpoint, well, I guess we've seen that healthcare isn't overly resilient at this point. We've seen systems go down for, the better part of 30 days. Yes, that's, it is interesting. So that's the goal is to get more resiliency built in across the board.
David, that's 20 minutes. Can you believe that's 20 minutes?
I can keep going on this, on a lot of this stuff. I think the new technologies are changing how we need to think about technology evolution and how we can, it's gonna change our workflows. It's gonna change how we look at things. Computing is so pervasive and so cheap these days.
Problems that we can solve today. When I was in, when I was in graduate school doing this kind of work, it was impossible to get access to this much computing power.
I remember when I first got my Commodore 64, and I started doing my papers on the word processor, whatever.
I go college, and the college only had like four computers in the lab, or six computers in the. and my friends were still typing it out on the IBM Electric. , and I'd show 'em the word processor and it, that response of, oh my gosh, that changes everything. That's, I feel like we're having those like every other week now.
Like yes. There's so much going
on. I think so. And I think that rate's only going to accelerate. Cat G P T is just the beginning of, a whole new transition and paradigm. I was blown away by the answers that came back with the grammar the idea, the, it was just fascinating.
It's fascinating. I can't wait to have you on the next show, but we'll put the chat g p t stuff at the end. Although it's on the hype cycle, doesn't it feel like if, like it's redefining the hype cycle, like it's gotta be a higher mountain at this point?
It's , certainly setting a very high.
But I can see just so many ways we could use it to improve even simple things. Like for my engineers to stop thinking about documenting, let chat g p t document it for you so I could actually read and understand your code.
Yeah. Will it be consistent and it's documentation though, will it be consistent?
I think it is. It has a pretty good it seems to have a pretty good tone, if you will, a tender to the way it document. So when I read the documentation that it inserted in the code, it was very explanatory in the sense that, here's what it's trying to do. Here's what your code is trying to do.
And I said, that's what I want documentation to do, help the next engineer who has to come along and read the code, walk through the flow of that.
And the promise in medicine is, think of all the places we have unstructured data and just crazy Tie it together. Yeah. Go. Now we say, okay, go ahead and read the note on this patient.
And summarize the key things that I need to know from this. And it creates a summary document that pulls out the most relevant. Now it's not gonna do, I don't hear me saying it's gonna do that today. It can only do what it's been train. To do. Correct.
But it's still pretty impressive for an algorithm.
Yeah. And I think that's what people are wondering is like, when can it be trained on medic? Look, we've been waiting for this since Watson and Watson and maybe Watson was the hype cycle and it was so early. I think
It was too early and it was, it overshot its promise.
Yeah. And the promise is overshot well chat, g p t, everything you log in says, Hey, experiment, this is new.
That kind of stuff. So your expectations are lowered before you go in and then you start playing with it and all of a sudden your expectations go, whoa.
Did it really just do that to us? Yeah. I was pleasantly surprised.
Yeah. Me too. Well, David, thanks again for coming on the show. Always great to catch up with
Yeah, same here.
And that is the news. If I were a CIO today, I would have every one of my team members listening to a show just like this one and trying to have conversations with them after the show about what they've learned and what we can apply at our health system. If you wanna sport this week, health, one of the ways you can do that is tell someone about our channels. We have two channels this week, health. And this week, health Newsroom, check them out today. You can find them wherever you listen to podcasts, and we are just about everywhere. You can find us on YouTube, apple, Google, overcast. You get the picture. And please subscribe on our website this week, health.com as well. We wanna thank our Newsday partners, Cedars-Sinai Accelerator, Clearsense CrowdStrike, Digital Scientists, Optimum, Pure Storage. SureTest ,Tausight and VMware, who 📍 invested in our mission to develop the next generation of health leaders. Thanks for listening. That's all for now.