AI taking healthcare jobs, and enrolling patients in Medicaid at a laundromat

It has been a fun month on the pod! If you haven’t been tuning in because you’re time constrained or more of a reader than a listener — I took some time this week to recap a few of the discussions I’ve had lately, and share some thoughts of my own.

This month, we had two conversations on the Second Opinion podcast with guests from very distinct corners of healthcare. One, an entrepreneur, with a background in education, is building a company to enroll thousands of primarily lower-income women into Medicaid and other public services that they’re eligible for. That’s Courtney Bragg, the CEO of Fabric WHealth, a startup that is backed by investors like TMV and Alumni Ventures. I was fascinated to talk to Courtney as the health care sector is bracing itself for Medicaid cuts. Fabric, by the way, is a reference to literally fabric, because Bragg’s company interfaces with patients at the laundromat.

Eric Larsen, my other guest, has made a career of thinking about the future of healthcare, and spends a lot of time advising health system and health plans on how to incorporate AI into their day-to-day operations. As a person who is on a first-name basis with Vinod Khosla and Sam Altman and CEOs at health systems, he bridges the gap between the healthcare and technology sectors better than almost anyone I know, as he’s able to speak both languages fluently. Larsen is also a deep thinker, who rapidly flips between reciting quotes from philosophical and historical texts with intel from the front lines of AI. Dr. Ash Zenooz, a radiologist and health-technology executive, joined the discussion and provided a valuable clinical perspective.

These may seem like very different topics and guest profiles. We started from very different places in our conversations, and yet we somehow landed in the same subject matters at the end:

The future of healthcare will incorporate AI deeply, automating processes, administration, and diagnosis, but the importance of the face-to-face interactions. I walked away thinking that there’s more role than ever for marrying the scientists and technologists (bolstered by AI) with the artists and philosophers. There’s so much uncertainty right now in what the future will hold, particularly as we read headline after headline about the oncoming wave of job cuts and read articles about the computer science bubble bursting because of AI. It used to be that to ensure a job in an unstable market, smart students have gone into computer science or any other STEM degree that could get them to a more technical role. Now, that’s not even a guarantee, particularly as demand for technical roles already appears to be dropping (whether this is a temporary blip remains to be seen).

So what does that mean for the future of work, particularly in the healthcare industry?

Larsen didn’t sugarcoat any of the changes that he believes are coming in the near future, noting that all the talk of augmentation versus dislocation or job replacement is really “fairy tales” that industries are telling themselves. He also doesn’t believe that there’s one “lump of labor” where if you automate some things, then you’ll see an overall reduction of jobs. Instead, he thinks the nature of how we work will change, both in how it’s structured and the skills required to do the job effectively. Look at medicine, where we still admit people based heavily on quantitative skills. Should we actively seek out a very different type of medical student in the future? Perhaps it’ll be more of the individuals with deep empathy, who have the potential to have the best bedside manner, to work alongside those who are highly proficient scientists working on new treatments and cures. Yes, there’s certainly LLMs that exhibit elements of empathy, with some used in therapy settings. But imagine finding out you have cancer, or you have a relative struggling with addiction. There are truly powerful moments where a human being with lived experience, or a provider who’s treated patients with similar journeys, can truly empathize with what that person is going through.

I’ve also heard a compelling argument that we touched upon that because ChatGPT reflects back the sum of common knowledge today, do we need more people in roles pushing boundaries in research? For instance, doctors who are also scientific researchers and pioneers? In theory, that kind of work could be more important than those doing pattern matching related to common knowledge.

On a more optimistic note, Larsen spoke to long-term benefits where GDP rises and we create a post scarcity society. So perhaps we’ll all just end up working less because we’ll be so augmented by machines. Our future could be that we all work 15 to 20 hours a week, and spend the rest of our lives painting and reading philosophy (one can only hope). But all joking aside, the relationship between labor productivity and leisure time does not appear to be entirely linear. Work weeks got shorter when Henry Ford implemented 8 hour days on his manufacturing lines, since the 1980s in the UK and US, increased productivity does not appear to have results in increased leisure time. Something tells me that our constant need for GDP growth combined with a contracting labor market problems means that we’re not heading into some AI-enabled leisure utopia in the future.

At the surface, Courtney and Eric should have polar opposite views of where healthcare is going. One is working at the grassroots of healthcare, while the other is involved in private equity deal making. But they didn’t. We concluded in both conversations that there will be a version of healthcare in the future where the most “physical” jobs are the most protected. Larsen believes it’s the jobs that require human presence will be valued by society in the decades to come. Embodied AI may be in our future, but it’s going to be some time before that happens. And in the meantime, who is going to help an elderly, frail person with mobility issues out of the bathtub? Or hand an overwhelmed mother at the laundromat a laundry cart and ask them how their day is going?

If you have a long car ride ahead of you, tune in to both episodes ideally in the following order:

The Eric Larsen interview

The Courtney Bragg interview

Larsen is one of the people I speak with most in my network about the incoming wave of AI tools and what it means for healthcare. He sees a potential future where vast swathes of our workforce are replaced or will experience major changes to their roles. That could happen to knowledge workers sooner than we all might think. Most at risk in his view: Knowledge worker roles that are essentially at the desk, and performed by people with multiple college degrees.

“All technologies to this time have been an amplification of human brain or low level cognition,” he said on the podcast, referring to prior technological revolutions like the printing press. “This one may actually replace and obsolete much of what we do in knowledge work.” Axios is calling it a “white collar bloodbath.”

Overall, he believes that AI will be positive for humanity and lead us in the direction of having more leisure time (he calls it a potential age of “abundance”) but he noted there are very likely going to be a lot of uncertainty in the immediate future as humans may not be wired well for that.

So what does this mean for those working in healthcare? In the USA, healthcare is a significant employer with over 20 million people across lots of smaller sub-sectors. That isn’t just the doctors, nurses and hospital or health plan administrators, but also countless people who make their living because of their knowledge and network in the industry (finance, law, journalism, consulting & more).

Some of those individuals may find their skills are needed more than ever. I’m personally bullish on the need for nurses because of their close relationship with patients. There’s also still going to be a need for people who can check blood pressure, administer medication, take blood, and so on. I also believe there will be more care advocates and coaches to help people make positive changes and bring them into healthcare settings when it’s indicated. I recently read a super interesting study showing that tracking changes in grocery orders, and building AI models to analyze that, might be a signal of ovarian cancer pre-diagnosis. I believe we’ll find many more connections like this as we parse more and more data, but someone will have to approach these patients, talk to them about something as terrifying as a potential cancer diagnosis, and direct them to get them the treatment they need.

Sometimes we fixate on the data, and forget what it takes to connect them to support. And no, it’s rarely an app-based “nudge.”

For their part, physicians will increasingly leverage AI as a clinical decision support tool and also in back office/administration. That is starting to feel inevitable, even as you continue to see some detractors. What will be fascinating to both watch and study will be how they react to this new information coming in. I’ve seen some interesting studies here also, indicating that physicians seem to want to gravitate to the AI generating “options.” Larsen thinks that human judgement may become an input, versus the be all, end all.

Whether or not AI holds the same bias as we humans do is an open question, particularly as it’s humans that have built and trained the models. But in theory, we are still in early innings when it comes to AI. One of my favorite quotes I’ve heard on it is that we’re in the “squealing baby” stage — over time, these models will only get more powerful and more accurate.

It’s worth noting that Larsen is one voice of many these days with a view on where AI is headed. Go on X these days, and you’ll find a few hundred venture capitalists looking to opine on the topic. I’ve also spoken to experts I respect who believe it’s overhyped, and have shared massive concerns about safety, efficacy and the potential for fabrication. I can’t think of a technology in my lifetime that garnered responses ranging from “it’s total smoke and mirrors” to “it’s a new speciation event.” But that’s the world we live in.

Either way, I believe the best thing we can do right now is to get prepared. I speak everyday with founders who are shedding engineering because of how AI has impacted the speed and efficiency of development. Coding projects are getting done in a single weekend, where it used to take months. Venture Capitalists are already discussing a time when we’re going to see “single person unicorns” There are changes already underway, so I don’t believe there’s value in not thinking or talking about how to adapt to them.

Some of these changes will require new policies. To offset some of the worst of the job losses, many tech execs including OpenAI’s Sam Altman are proponents of universal basic income or UBI, which guarantees some degree of wealth so that people can afford to buy the basics like housing and food. There are experiments going on as we speak to prove the viability and examine the effects.

So Is AI a force for good or evil? Both Zenooz and Larsen agree it’s neutral - it could be shaped for either motivation. In AI maximalist circles, there’s questions circling on humans can better understand how to understand what Larsen calls “these creatures” — so that it doesn’t ultimately control us. “AI will shape us far more than we will shape it,” added Zenooz.

I’d love to hear your views on the topic, and how you think societies need to adapt. One of my last questions to Larsen was whether he would advise friends to have more kids, given all the destabilization that AI might bring. He actually doubled down on that, arguing that there will be beautiful aspects to the years ahead of us - not just uncertainty and change.

Millions of Americans spend countless hours at laundromats on a weekly basis, the majority of them women. So how could we intervene in that moment as clothing is spinning to make a positive difference in people’s lives? Well, Fabric Health’s Courtney Bragg is fascinated by that very question, even in an era where so much of our focus is on technology and tools, including AI. Because I remain passionate that human-to-human interaction still has power, I had her on the podcast to talk about it.

She believes there’s an “overlooked and underutilized” moment to meet people at laundromats and build a relationship by having conversations with them. “It’s two hours of dwell time” she said, and a rare opportunity to meet mostly women who are making care-related decisions for their family. Her team doesn’t just walk up to people and ask if they have insurance. Instead they spend hours of quality time with them, answering questions, offering to help and generally building trust. The questions about insurance come later. The fact is that even in this technology saturated world, where practically everyone has a smartphone, some of the most valuable ways for interacting with underserved populations still require in-person interactions.

“It’s about offering mom a laundry cart, and in that time doing a health risk assessment, a social determinants of health (SDoH) screener.. you can be doing all those things as long as you’re human first,” she explained.

While services is at the heart of the interaction, there’s technology built in to make things more efficient — and yes, potentially a role for AI in the background. One example: Courtney’s company integrates with the laundromat point-of-sale systems so they know once the laundry cycle has started. At that point, they can send a text right as the person walks in so they’re prepared for the encounter. There’s also a fascinating meta point here that people who go to laundromats are being rented productivity improvements via a machine that has been around for almost 100 years. What does this say about a future where AI is going to be similarly “rented” to people, also for the sake of increased productivity?

Per Bragg, Fabric Health is working with 27,000 families at this moment in time - at a lot of laundromats - to enroll them in Medicaid and support their health and social needs. And that’s at a time when Senate is proposing billions more in Medicaid cuts. Like many in the health policy world, she is deeply concerned about the cuts, and wonders about the strategy given that so many Medicaid recipients voted for Trump (very good question!).

What struck me as so refreshing about our conversation is that Fabric Health is building a real business focused on personal interactions and conversations with people, because it builds revenue when patients access health care. When most companies I interact with speak about the patient on the other side of the interaction, they use vague jargon (“empower,” anyone) and rarely describe what the experience feels like. Courtney talks about having a conversation, whether it’s by text or in-person.

“There is a human component to healthcare,” said Bragg. “Anytime you get on the phone, the first thing anyone does is press zero - and that is to get to a human and to get an answer.”

So what do these two guests have in common?

As Bragg often says: “It’s got to be muffins, before mammograms.” I haven’t shared that quote with Larsen to ask what he thinks, but I believe he’d like it. What she’s essentially saying is that it’s important to build the relationship first, then talk about the intervention. You’re going to be a lot more successful that way.

But let’s return to the topic of AI. The muffin might break the ice, which opens up the opportunity to talk to a patient about a potential health intervention that could save their life. But consider what AI can do to speed all that up. AI could let Courtney’s team know when the patient has entered the laundromat, it could help her access a record that incorporates claims, clinical information, social determinants to help make recommendation. It could interpret the results of the mammogram, alongside a human radiologist, and help the person get scheduled in for any treatments they need. Who knows. Someday it might even take care of the laundry on the individual’s behalf. And perhaps, we’ll find very different forums to meet and interact.

Needless to say, I’m both excited and terrified by the future that we have ahead of us.

The Second Opinion podcast is off to the races. If you’d like to propose a guest, I’d love to hear from you! My only requirement is it must be a perspective we don’t hear everyday.