Jul 31, 2025 · 28m · a16z
Health Tech Founders: The Future of Care Is Personalized, Proactive—and AI-Powered
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the a16z Podcast, host Bryan Kim talks with Function Health CEO Jonathan Trollin and Slingshot AI CEO Daniel Reid Cahn about transforming healthcare through direct-to-consumer AI platforms. Together, they explore how personalized biological data and AI therapy can scale proactive wellness, alleviate practitioner burnout, and augment human care.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The host as informed peer 2.0 Guest teaching 2.8 Guest disagreement 1.7 The host pushing back 0.8
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 15:20 Daniel attacks addict-by-design tech models and ChatGPT sycophancy
Daniel forcefully rejects the standard tech playbook of trying to addict users for retention, while calling out ChatGPT's sycophantic validation as useless for real therapy.
Hardest push from the host ▶ 8:20 Bryan challenges whether an AI tool should be framed as a doctor or therapistBryan interrupts and directly pushes back on the guests' framing, questioning whether an AI conversational model can truly be conceptualized as a doctor or therapist.
Biggest teaching moment ▶ 12:21 Jonathan reframes consumer app engagement from digital DAU to biological contextIn response to Bryan's inquiry about traditional app retention metrics, Jonathan educates the host on why health AI must measure hourly active biological context rather than screen time.
The host holds their own ▶ 16:37 Bryan outlines thesis on AI data leverage and biomarker translationBryan demonstrates strong investor domain expertise by explaining how AI becomes exponentially more useful when translating complex health biomarkers into actionable consumer insights.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| a16z Podcast Legal Disclaimer and Animated Title Sequence | 1 | 2 | 1 | 0 | Host Bryan Kim opens with a friendly welcome and invites the founders to share their origin stories. The guests provide background on systemic healthcare gaps, including Daniel citing mental health access statistics, while the host maintains a purely receptive role. | |
| Reimagining User Experience and Framing AI in Health | 2 | 3 | 2 | 2 | The guests explore how AI reframes UX, comparing current paradigms to motorized horses. Host Bryan interjects with mild pushback, questioning whether calling AI a doctor or therapist is accurate before Daniel explains why consumer demand drives the terminology. | |
| Building Trust, Execution Speed, and Momentum as a Moat | 2 | 2 | 1 | 0 | Bryan prompts a discussion on momentum as a moat in the AI era. Both guests collaboratively outline how rapid execution, waitlist management, and continuous daily model training build user trust without host resistance. | |
| Biological Context, Hourly Active Use, and Non-Addictive Design | 2 | 4 | 3 | 1 | Bryan asks standard consumer questions about engagement and retention metrics. Jonathan and Daniel reject traditional tech playbooks around daily active use and addictive design, with Daniel criticizing ChatGPT's sycophantic validation. | |
| Personal Data Integration, Psychological Safety, and Systemic Burnout | 3 | 3 | 2 | 1 | Bryan presents his thesis on data leverage, sharing personal difficulty understanding raw biomarkers. The guests elaborate on how AI creates psychological safety for honest user disclosure while criticizing systemic physician burnout. | |
| Augmenting Human Care and Supporting Healthcare Practitioners | 2 | 3 | 1 | 1 | Bryan asks what unique roles human practitioners retain that AI cannot replace. Daniel and Jonathan explain how AI augments clinical workflows by turning noise into structured signal rather than replacing human practitioners. |
Statements from this episode (17)
Cahn: 54% of individuals with mental illness receive no care
“54% of those with a mental illness, quote unquote, don't have any care whatsoever.”
Trollin: Replicating Function Health's MVP previously cost tens of thousands of dollars
“I had to spend tens of thousands of dollars. And tens of hours or over a hundred hours to get on top of my health, to get to equivalent and get near parity with where function was at an MVP level.”
Cahn: There are now more than 10,000 people for every practicing therapist
“There are now more than 10,000 people for every therapist.”
Cahn: AI products shouldn't just replicate human therapist or doctor workflows
“We can't just take what's existed with humans and then just pretend that an AI can do exactly the same thing.”
Trollin: Calling applications 'AI doctors' is like saying 'motorized horses'
“To look backwards, even call an AI doctor or an AI therapist. It's like saying, like, it's the motorized horse.”
Trollin: Function Health projects reaching nearly 1 billion lab tests by 2026
“We're at a place where this time next year, we'll probably, we'll have, if we do what we're supposed to do, we'll have done nearly a billion lab tests.”
Cahn: Slingshot AI trains new models roughly every two days
“We train new models, like almost every day. Right now, it's like every two days, but there's this constant sense of improvement that's really important.”
Trollin: Health AI platforms will be measured by hourly active use
“We think that it's not going to be about daily active use. It's going to measure an hourly active use.”
Trollin: Personal health data is the prerequisite context for daily AI
“We think that most AI experiences are being built around a personal context that's limited to digital activity. But we believe that one's personal health is actually the prerequisite content for AI that people will use every day and all day.”
Cahn: Ash AI therapy users average engagement at least thrice weekly
“I would say that our users who come in, users tend to come in when they really need help. And as a result, I would say like our normal usage is something like at least three times a week that people are talking and it varies.”
Cahn: AI therapy must challenge users rather than validate them
“People don't like that. That's not what's helpful. That's not what builds a therapeutic alliance. What builds an alliance is being challenged. It's finding a new way to look at your problem.”
Cahn: Users open up faster with AI therapy than human therapists
“Compared to a human it might take 10 conversations with your therapist before you tell them what you're actually there for, because you're like judging them. You're trying to figure things out. With Ash, people move way faster. That's been really fascinating. …”
Cahn: High mental health demand lets platforms profit with lower-qualified therapists
“And those are honestly amazing businesses because the demand is so absurdly high that no matter what, when you get more supply of therapists, even if it means that you're just paying therapists less and you're getting more therapists and you're getting less qu…”
Cahn: Ash AI users increase real-world human connections unlike ChatGPT users
“In a space of chat, like chatbots, like people who talk to ChatGPT about their mental health largely over time lose connections with other humans. Like they actually connect with people less than replace it. With Ash, we've shown now Over time, on average, peo…”
Cahn: Therapists actively recommend Ash AI and use it themselves
“We have a huge number of therapists who refer their patients. I mean, we get so many people who use the app who say, I found this because my therapist recommended it. We have so many therapists that tell us they use it because they don't have someone good enou…”
Cahn: Slingshot AI is building a foundation model for psychology
“We call our technology a foundation model for psychology, and at the core, when we think about what we build as a company, Ash is our product, our foundation model for psychology is what we build”
Cahn: AI therapy must preserve patient autonomy over short-term problem solving
“It is often better. And this is really controversial for a therapist. You know, you go to your therapist and you're like, is my boyfriend toxic? And the therapist might actually have to say, or to realize internally, it is better that this person remain in thi…”