Could A.I. Replace Healthcare Workers? - Gabe Dalporto, CEO of Udacity


Relax. Artificial Intelligence can't do EVERY job, but Gabe Dalporto believes it's a good idea to think about whether it could replace yours, even if you work in healthcare. As CEO of the popular tech education platform Udacity, Dalporto is a champion of regularly "upskilling" to be prepared for constant change and opportunities. He and host Shiv Gaglani talk about the role of AI in healthcare (see link in the transcript to a Udacity conference on the subject) and how the COVID pandemic is accelerating digital transformation.




SHIV GAGLANI: Hi. I'm Shiv Gaglani. Today on Raise the Line, I'm happy to be joined by Gabriel Dalporto, the CEO of Udacity, an online learning platform that focuses on making career advancement possible for learners through mastery of in-demand skills. Prior to joining Udacity, he held senior leadership roles at LendingTree and helped drive its growth into a multibillion-dollar financial services marketplace. 

Thanks, Gabriel, for being with us today.

GABRIEL DALPORTO: Pleasure to join!

SHIV GAGLANI: I'd love to start by learning more about your background and career path. How did you go from earning a Bachelor's and Master's degree in Nuclear Engineering to now running one of the most popular learning sites?

GABRIEL DALPORTO: I'm a little bit of a purple unicorn in terms of background. It's very unusual. I think the thing that runs through it is I just love learning. I am personally a lifelong learner, which I really think is, in a way, the skill of the future. I studied nuclear engineering because I thought it was fascinating.

I got out into industry and realized that I was working for a nuclear utility, which is the slowest moving organization on the planet by design, because last year the core didn't meltdown, and you just don’t do anything different.  So I shook things up a little bit, got a tour of duty in strategy consulting, cut my teeth on business, then moved into marketing because it was like the exact opposite of engineering, far more emotional and creative.

I worked my way up through the marketing ranks, got the nod as the CMO of LendingTree back in 2011, and helped LendingTree initially build a world-class marketing and growth engine to really ignite the growth of the company. Over time, I took over as president, shifted strategy a bit, diversified the business, launching different lines of businesses that worked out really well. 

One Friday night, I got a call from our CEO, Doug Lebda, who turned out to be a real mentor for me in my career. He said, “Hey, Gabriel, do you want to be the CFO?” I'm like, “What are you talking about?” He said, “You know the business inside out now, you know how to grow the business, and I need someone in finance who can drive the business, drive the P&L from a strategic perspective and build a world-class analytics organization for decision making.” I said, “That kind of makes sense. Let me give it a shot.”

It was a little uncomfortable for the first couple of weeks and months as I figured out the ropes of the finance role, but it turned out to be a fantastic opportunity for me to stretch myself and expand my skill set. Over the course of my tenure there from 2011 to 2018, we grew the company stock price from $5 to $251, the number three in return on the NASDAQ. I learned a ton. The company did great and went through a ton of growth. It was a great relationship. Which kind of brings me through today, which is after I left LendingTree and I joined the board, I spent a lot of my time thinking about my home state of West Virginia.

I grew up there in the 1970s and 1980s. At the time, the economy was pretty vibrant, but then I go back today, and I see towns boarded up, country clubs shut down, and people working in minimum wage service industries. It's a very acute pain. I see what happens when people don't re-skill when the world changes. That was very much on the top of mind when I met Sebastian, the founder of Udacity, and they were looking for a CEO.

He said, “Why don't you, instead of trying to figure out how to help West Virginia, do that too, but also help the world to figure out how to re-skill themselves, so they don't have to go through the pain that West Virginia went through.” I said, “You're right,” so that's what brought me to Udacity. It is absolutely my passion. The product is amazing. It changes people's lives for the better.

SHIV GAGLANI: That's quite a career and an amazing story about how you got involved. I actually didn't know about the West Virginia connection. That actually brings me to my next question, which is, we're in the heart of the COVID pandemic. We have seen record levels of unemployment.  Nothing has been worse since the Great Depression. There are some predictions that will get worse than the Great Depression. Have you seen, at Udacity, an increase in enrolments as a result of this? How has it affected the business at this point?

GABRIEL DALPORTO: When I think about our core businesses, we have a consumer business, which is how everybody knows Udacity. We also have an enterprise business that is growing like crazy and will probably be our largest business this year. We are selling to Fortune 500 companies to retrain their workforces and digitally transform them. Then we have a government business that is helping governments retrain their populations and build the skills they need within their national borders. 

We expected a spike in the consumer business, which we saw. We've seen a 10x increase in daily sign-ups. We expected an increase in the government business, which we have seen. We've seen strength in that business. Enterprise business, we didn't really know because we thought enterprises might cut back their training spend and workforce development spend, but it turns out we have not seen that. We have actually seen a lot of strength in the enterprise business. The reason is we've really become attached to business transformation. 

We typically partner with CEOs and CTOs, understand the fact that these giant corporations need to go through digital transformation, they need to build AI and software development skills because those skills are hard to come by in the market, and we're helping them build those skills internally. Then those people help them transform their businesses. We're viewed as mission-critical in that respect.

Answering your question, we’ve seen an uptick across all of our business, we’ve seen an uptick, even though we weren’t totally expecting the enterprise piece, but we have certainly seen a big upswing in sign-ups on the consumer side.

SHIV GAGLANI: Udacity distinguished itself initially with the nano degrees and the fact that there was a heavy, heavy focus on cutting-edge AI training. That's where Sebastian came through leading the Google car development, for example. Is that still the secret sauce of Udacity? What else have you guys added that have made you distinguished from other education platforms?

GABRIEL DALPORTO: It is a really important question and a great question. If you think about just this massive flood of people that are now unemployed, if you try to retrain them on the university system, you will break the university system. There's just not enough capacity, and you would also probably bankrupt the country. It's very, very slow and expensive to educate someone through the university system. 

Udacity helped create a whole category called MOOCs (Massive Open Online Courses) which initially took university content, videotaped professors, put it online. Now, you could watch a University of Michigan or MIT professor deliver a lecture, and the MOOCs made it free or cheap, which is great. You could be in Argentina or Germany or wherever and watch an MIT professor. That is awesome. The problem with that is you don't learn karate by watching Bruce Lee movies.

The problem— Just taking a traditional university lecture and putting it online, it's interesting. You may get some information, but you don't learn like hands-on practitioner level skills, so Sebastian and my predecessors at Udacity came up with a better model in between those two, which is we start with project-based learning, so we start with an employable job resume.

If you want to be a machine learning engineer, if you want to be a data analyst, if you want to be a self-driving car engineer, your resume has to have these four bullet points like “I've done this, this, this and this.” I haven't just read about it or watched a video.  I've done it. Then we work backward, and we say, “How do we get them that hands-on experience?” “We create deep, immersive projects. You take software, you apply it to massive data sets, you apply machine learning algorithms, and you deliver it in a real-world, industry-specific context where you can say, “I delivered a credit model for a credit card company, or I delivered software for this security application.”

Then you go into industry, not into academia, and partner with people like Google, Amazon, Intel, and co-create the actual content. Think of our experience as a four to five-minute video lecture led by someone at Google or Amazon, followed by hands-on keys coding, followed by another lecture, coding, reinforcement learning -- end of the month you go deep into an immersive project, deliver a real-world tangible industry outcome, repeat that from four to five months, 10 hours a week, part-time, 2:00 a.m, 4:00 a.m, 6:00 p.m. It doesn't matter. Then you come out of this with a resume that you've done it, and you can do something. 

An example of this is our self-driving car engineer course. You get your project, upload your code into our self-driving car, then that drives around our parking lot with your code, and it stops at red and goes on green and doesn’t hit stuff. That's the core of what we do. It's a lot harder than taking a university professor and putting their content online, but it just works 10 times better. For 1/100 of the cost of a university degree, about a 1/10 of the cost of a boot camp, you come out of these programs with employable practitioner-level skills. 

SHIV GAGLANI: I have to ask the follow-up question. You mentioned that people upload their software to their own self-driving cars. Have you ever been very positively or very negatively surprised by the outcome of that?

GABRIEL DALPORTO: The beauty of project-based learning is you upload that code, which either works or doesn't. Sometimes it works, sometimes it doesn't. If it works, you pass. If it doesn't work, then we have what's called project reviewers, and whatever project, whether you pass or fail, you get line by line code reviews back. 

If you upload it, and it runs into a wall or something, they're going to review it and give you feedback, and you're going to submit it again. The beauty of the model is that your code actually has to work to pass the class.

SHIV GAGLANI: I love that model a lot. In medical education, we have a lot of problem-based learning, and obviously, you have to go see the patients and take care of the patients before you can actually get your degrees. 

You mentioned earlier that you have seen a lot of demand because of COVID. We call this podcast Raise the Line because our focus at Osmosis is essentially 'how do we increase healthcare capacity through workforce training'. I believe that you have launched an AI in healthcare course. I'd love to learn more about that and what Udacity is doing in healthcare.

GABRIEL DALPORTO: Absolutely. Healthcare is obviously, like right now, an absolutely crucial field, but always a super important field. We have seen a lot of innovation in healthcare. As you mentioned, we just launched an AI for a healthcare nano degree. Think about this as how you use AI across a bunch of different applications in healthcare. One of our projects is looking at x-rays and applying AI to diagnose x-rays. 

There are tons of applications that you can apply to COVID like tracking patients, like predicting the spread of a pandemic, and they can really inform governments on how to respond and how quickly they respond. I think AI in healthcare has never been more critical or more relevant with COVID, but also beyond just your typical COVID applications in terms of creating smart hospitals and augmented decision making where you may have a physician, but you also have an AI sitting there looking at the same data and coming at it from a different angle, forcing each other to think about it in different ways. 

Also, mobile medicine is a very interesting market that's projected to be a $23 billion market in the next couple of years, not only delivering medicine online or video streaming, but also diagnosis and features like tracking your heart rate to predict different health issues. There are just tremendous applications of AI and machine learning to healthcare. We're just scratching the surface. 

SHIV GAGLANI: AI and healthcare have always been interesting to me because right before I started Osmosis, I wrote a paper with a guy named Eric Topol, who I'm sure that you know. He wrote the book on AI in healthcare. One of the articles we published was on mobile medicine and how it can be used to improve health education. 

It's very nice to see that Udacity is leading the charge in that. About the people who enroll in AI in healthcare and medicine, do you see a lot of clinicians doing that, or are there more software engineers who are interested in healthcare? How do you see that map out?

GABRIEL DALPORTO: It's a great question. I think it is probably a bit of both as there's a big industry around research, so it's going to be the research centers or anyone who's interested in how to apply AI to their situation and make it more efficient. Hospitals, obviously, can benefit here.  

You mentioned Dr. Topol. We are in the middle— today and tomorrow we will have an AI for healthcare conference that we just put on, a virtual conference in the time of COVID. He is the keynote tomorrow, Wednesday, May 13, and that will all be made online for your viewers in the next few days, so if you want to see Dr. Topol talk about AI in healthcare and applications, you can come to our site and watch him speak about it. https://www.udacity.com/ai_in_healthcare/virtual_conference

For sure, it's a broad mix, which we see across our programs. We find people who are wanting just to learn. We find people who are wanting to upskill themselves and do their jobs better. We find people who want to move into a new area, and our program is designed to give you those hands-on practitioner level skills, so no matter what angle you're taking, you're coming out of that with real, applicable skills. You can apply it to machine learning or AI or to a dataset and see the outcomes as you go. 

SHIV GAGLANI: That's wonderful. It reminds me of another famous person whom I heard speak a couple of years ago, Vinod Khosla, who got in a little controversy at one point because he said that 80% of what doctors do will be replaced by machines in the next decade.  It was controversial because I think he was misinterpreted as believing that 80% of doctors wouldn't be needed, but it will change the role of doctors. 

Since so many of our audience members are current or future health professionals and you all are doing so much around AI and healthcare, what do you see as the role of a clinician being over the next decade, based on all the AI that's coming down the pipeline?

GABRIEL DALPORTO: I think AI does really well with, for example, taking huge data sets and then predicting likely outcomes. AI and software, in general, do really well at taking anything that could be repeated and automating that. What AI will not do well, at least not for the next several years, is creative thought. 

There's going to be a lot of situations where what was thought of as the core value of a doctor, or a clinician is automatable through AI, but there's going to be a lot of other areas where it is not, so I think you should constantly be looking at your skills. You should not assume that what you're doing now is going to be what you will be doing in five years.

You should be thinking about how do I upscale myself in those areas that AI is not good at and emphasize there, or how do I become the person who is creating the technology that is automating, and you'll probably make a lot of money doing that. One of those two is fine. I think that assuming that your job is not going to be touched is not a good strategy.

SHIV GAGLANI: That makes sense. We hear a lot of talk about how retail workers or transportation workers will be automated. Their jobs will be automated away, so they have to reskill, but health professionals need to remember that this may help them focus less on memorizing things and regurgitating facts and more on the care aspect of being a caregiver or a clinician. 

GABRIEL DALPORTO: Empathy is not something AI is particularly good at, so understanding the specific situations, being creative and empathetic and things like that, it's something that's quite hard for an AI to do.

SHIV GAGLANI: Totally, 100%. Do you have any final thoughts or comments that you'd like to share with our audience?

GABRIEL DALPORTO: Look - there are more people out of work, as you mentioned, than any time since the Great Depression. Who knows where it goes from here? My guess is that we are going to face a really hard year or two. Personally, I think this is the time for individuals, governments, and enterprises all to step up and take upskilling seriously.

Many of those jobs that have been eliminated are just not coming back. You better believe that COVID is going to accelerate digital transformation and automation. Many of the jobs that haven't been eliminated will be eliminated. If we don't re-skill, if we don't take this seriously, a lot of us are going to be sitting in menial service industries, making minimum wage, and that's not what anybody wants.

I think, number one, the government has an important role to put forward a specific plan. We should think about our whole notion of spending all of our federal dollars against universities, as this is a terrible idea. Universities play a really important role, but we should think about it. If we took 10% of the budget spent on universities and spent that on upskilling programs, you could re-skill 3.5 million people. That's a big number. 

If we gave tax credits, even small tax credits to corporations for retraining, those corporations would match those tax credits and then invest in reskilling their workers instead of laying them off. Then individuals need to take responsibility for upscaling themselves. Don't assume that your employer is going to do it because some of them will, but most of them won't.

You need to think about your career going forward, the skills of the future, and make sure you're on the leading edge. I started marketing back, as I mentioned, when most marketing was TV and direct mail. My first job was in direct mail, and that entire industry is mostly digital now. The people who made the transition got better jobs and made more money. People that didn't were left behind in a significant way. I think the same will be true of every industry.

I would just encourage everybody to think about that, think about what the future looks like, think about how to continue to add to your skillset. It doesn't have to be Udacity. It doesn't have to be training in AI or software development, but figure out what the jobs of the future are, go for those, and continually evolve yourself.

SHIV GAGLANI: That's a piece of great advice and a super fascinating conversation. I appreciate how specific some of your proposals have been, and again, all the work you and your team at Udacity are doing with it. I'd like to thank you for taking the time to be with us today, Gabriel.

GABRIEL DALPORTO: Awesome. Thank you so much. I appreciate being on your show, and it was great speaking with you. Thank you.

SHIV GAGLANI: With that. I'm Shiv Gaglani. Thank you for checking out today's show. Remember to do your part to flatten the curve and raise the line: we're all in this together.