Investing in healthcare startups: Thanasi Tsiodras, Chair of the Board at Halo Health

WEBVTT

00:00:02.129 --> 00:00:05.389
Thanks so much for joining me today, Tanasi.

00:00:05.730 --> 00:00:09.150
Pleasure to be here. If you could give an introduction

00:00:09.150 --> 00:00:13.070
as to what you do today and how did you get here?

00:00:14.029 --> 00:00:20.309
Yeah, so I spend about 40 % of my time looking

00:00:20.309 --> 00:00:24.449
at investments with Halo Health and Genesis Ventures

00:00:24.449 --> 00:00:27.429
as a venture partner for Genesis Ventures, looking

00:00:27.429 --> 00:00:31.339
at all of their health tech portfolio. and Halo

00:00:31.339 --> 00:00:35.619
Health, I'm the chair of the board. Just as background,

00:00:35.840 --> 00:00:38.679
Genesis Ventures is a generalist fund out of

00:00:38.679 --> 00:00:41.320
Greece, investing globally in Greek founders

00:00:41.320 --> 00:00:43.380
mostly, or at least companies willing to have

00:00:43.380 --> 00:00:48.159
a presence in Greece. Recently listed by Sifted

00:00:48.159 --> 00:00:52.539
as a top five most active investors in Europe.

00:00:53.960 --> 00:00:57.619
And Halo Health is Canada's physician angel group.

00:00:57.899 --> 00:01:01.579
and we have about 400 members who are actively

00:01:01.579 --> 00:01:06.319
investing and we've invested in close to 50 companies

00:01:06.319 --> 00:01:10.719
so far with five exits. What are some similarities

00:01:10.719 --> 00:01:14.099
and differences you see in the European or the

00:01:14.099 --> 00:01:16.719
Greek way of investing in startups versus the

00:01:16.719 --> 00:01:21.140
Canadian way? The Canadian ecosystem is much

00:01:21.140 --> 00:01:23.319
more mature compared to the Greek ecosystem.

00:01:23.659 --> 00:01:27.180
About 10 years ago there was We're barely starting

00:01:27.180 --> 00:01:30.640
to see an ecosystem with the first fund being

00:01:30.640 --> 00:01:33.159
created at the time that the first private fund,

00:01:33.180 --> 00:01:36.700
there was maybe one public fund that was created

00:01:36.700 --> 00:01:39.620
in 2011, whereas venture capital has been in

00:01:39.620 --> 00:01:45.239
Canada for decades. So being less mature means

00:01:45.239 --> 00:01:49.620
smaller check sizes, smaller risk appetite, smaller

00:01:49.620 --> 00:01:53.379
exits, everything's just smaller, but growing

00:01:53.379 --> 00:01:59.359
fairly quickly. Fairly. We hit an upswing in

00:01:59.359 --> 00:02:03.459
2021 and then like a lot of things in venture,

00:02:04.060 --> 00:02:07.400
it's stalled a little bit since then. Was the

00:02:07.400 --> 00:02:10.400
plan to always transition to venture from being

00:02:10.400 --> 00:02:13.139
a physician? And if not, like, was there a moment

00:02:13.139 --> 00:02:16.520
or what made you realize that, you know, maybe

00:02:16.520 --> 00:02:18.780
I want to branch out and I want to branch out

00:02:18.780 --> 00:02:21.400
into investing in startups, not something else?

00:02:21.689 --> 00:02:24.669
So I still maintain a small clinical practice

00:02:24.669 --> 00:02:30.590
and it was not the plan. No. I practiced more

00:02:30.590 --> 00:02:35.909
than full -time for 11, 12 years. At the tail

00:02:35.909 --> 00:02:38.389
end of that, I started working for a startup

00:02:38.389 --> 00:02:43.069
called Dialogue, which ended up IPO -ing in 2001.

00:02:43.150 --> 00:02:48.069
So I joined in early 2018. They went in their

00:02:48.069 --> 00:02:51.259
seed stage and then the IPO three years later

00:02:51.259 --> 00:02:54.919
and the big health tech boom. So there I got

00:02:54.919 --> 00:02:57.840
to see a different way of doing things, much

00:02:57.840 --> 00:03:01.960
more agile, much more innovative, different attitudes.

00:03:02.539 --> 00:03:05.699
And that's how I got introduced to the world

00:03:05.699 --> 00:03:08.539
of startups. So investing was something that

00:03:08.539 --> 00:03:13.699
I was doing by myself in the public markets before

00:03:13.699 --> 00:03:18.340
that. But startups was through working at Dialogue.

00:03:19.079 --> 00:03:23.560
And as of 2020, I started looking into angel

00:03:23.560 --> 00:03:27.580
investing. I got to really educate myself through

00:03:27.580 --> 00:03:30.080
Halo Health. And then I started writing checks

00:03:30.080 --> 00:03:37.020
as of 2021. Talk to me a bit more about the mindset

00:03:37.020 --> 00:03:40.580
as a physician and as an investor in startups.

00:03:41.060 --> 00:03:43.400
From my experience, med school teaches you to

00:03:43.400 --> 00:03:49.110
be the anti entrepreneur, the anti VC. to minimize

00:03:49.110 --> 00:03:54.669
risk at all costs. And even the way we think,

00:03:55.530 --> 00:03:58.729
when a patient comes in, my thought process in

00:03:58.729 --> 00:04:01.770
urgent care and a walk -in is always, is this

00:04:01.770 --> 00:04:04.550
going to harm this patient? Is this going to

00:04:04.550 --> 00:04:06.949
lead to a loss of life or limb in the next 24

00:04:06.949 --> 00:04:09.949
hours? And then that timeline expands. Is this

00:04:09.949 --> 00:04:12.169
something here that's going to cause morbidity

00:04:12.169 --> 00:04:14.469
over the next six months? And then is this cancer?

00:04:14.810 --> 00:04:18.930
So it's a very I don't want to use the word pessimist,

00:04:18.970 --> 00:04:22.990
but it's a very risk mitigation approach to work

00:04:22.990 --> 00:04:26.550
and to practice. Whereas venture is often about,

00:04:26.730 --> 00:04:29.470
what is the upside here? How big can this be?

00:04:30.050 --> 00:04:33.569
And we understand risks are inherent in venture

00:04:33.569 --> 00:04:36.069
and we accept them. Whereas medicine, the type

00:04:36.069 --> 00:04:37.949
of thinking is very different. So talk to me

00:04:37.949 --> 00:04:41.750
about, how did you find, maybe through your work

00:04:41.750 --> 00:04:44.009
in dialogue and then Halo Health and Genesis.

00:04:44.490 --> 00:04:47.610
Like, what are your thoughts on the shift in

00:04:47.610 --> 00:04:52.449
mindset between the two practices you do? I think

00:04:52.449 --> 00:04:58.670
the closest parallel would be it's as if you're

00:04:58.670 --> 00:05:02.250
dealing with very undifferentiated patients all

00:05:02.250 --> 00:05:06.009
the time. So a startup, when you see a pitch

00:05:06.009 --> 00:05:09.430
deck, you're not sure what the outcome is going

00:05:09.430 --> 00:05:13.069
to be. There are plenty of risks and you just

00:05:13.069 --> 00:05:17.779
have to price in and then minimize as many of

00:05:17.779 --> 00:05:20.139
the risk as possible. The team, the market, the

00:05:20.139 --> 00:05:24.180
product, the technical due diligence, how well

00:05:24.180 --> 00:05:26.959
the team members know each other, not if they're

00:05:26.959 --> 00:05:35.899
just qualified to be in that market. And you

00:05:35.899 --> 00:05:41.019
just, you have to look for asymmetric bet. So

00:05:41.019 --> 00:05:44.240
we're going to write a check for this much, but

00:05:44.430 --> 00:05:49.170
if this goes well, how big could this be? Whereas

00:05:49.170 --> 00:05:51.709
in medicine, you're still mitigating risks, if

00:05:51.709 --> 00:05:57.389
we're talking about undifferentiated situations

00:05:57.389 --> 00:06:00.649
and not routine cases that you would see in a

00:06:00.649 --> 00:06:07.069
walk -in clinic or in any kind of settings. But

00:06:07.069 --> 00:06:10.310
in medicine, there's no upside. You're just...

00:06:10.060 --> 00:06:12.959
capping your downside. And then that's a completely

00:06:12.959 --> 00:06:15.699
different way of doing things. You do not get

00:06:15.699 --> 00:06:18.779
rewarded for taking risks. You get slapped on

00:06:18.779 --> 00:06:22.959
the wrist for taking risks. In venture, if you

00:06:22.959 --> 00:06:28.439
mitigate risks well enough, you can get rewarded

00:06:28.439 --> 00:06:32.680
quite a bit. And it's also a numbers game. So

00:06:32.680 --> 00:06:37.600
you do your 10 investments, six go to zero, not

00:06:37.600 --> 00:06:40.470
a problem. you get your one or two, they give

00:06:40.470 --> 00:06:45.689
you a hundred X, you're in the top 25%. In medicine,

00:06:46.110 --> 00:06:51.670
six of your diagnoses are not right. You're looking

00:06:51.670 --> 00:06:55.310
at some complications. Well, that's a difference.

00:06:55.410 --> 00:06:58.629
So it's still making decisions on incomplete

00:06:58.629 --> 00:07:03.170
information. It's still mitigating risk, but

00:07:03.170 --> 00:07:05.750
the overall philosophy is very, very different.

00:07:06.829 --> 00:07:11.370
What advice do you have for non -clinical investors

00:07:11.370 --> 00:07:19.589
who want to invest in healthcare? So, if we're

00:07:19.589 --> 00:07:24.730
talking about digital healthcare or medical device,

00:07:26.509 --> 00:07:30.930
technology is good, and good technology is useful,

00:07:31.769 --> 00:07:35.920
but hospitals, physicians, they don't pay for

00:07:35.920 --> 00:07:38.160
good technology. They pay for a solution to the

00:07:38.160 --> 00:07:41.439
problem. And you need to offer that solution

00:07:41.439 --> 00:07:44.600
and you need to be able to integrate it into

00:07:44.600 --> 00:07:47.379
the daily workflow as seamlessly as possible.

00:07:48.980 --> 00:07:53.180
So unless you're willing to do that integration

00:07:53.180 --> 00:07:56.860
from A to Z and minimize friction as much as

00:07:56.860 --> 00:07:59.579
possible for everyone involved, whether that's

00:07:59.579 --> 00:08:03.490
the... the physician or the clinician who's using

00:08:03.490 --> 00:08:10.290
it, the patient, the administrative staff, your

00:08:10.290 --> 00:08:12.350
good and impressive technology is not going to

00:08:12.350 --> 00:08:16.209
get you anywhere. So this is the important thing

00:08:16.209 --> 00:08:20.430
that I would look at as an investor looking into

00:08:20.430 --> 00:08:26.009
digital health companies. You have two investment

00:08:26.009 --> 00:08:29.740
opportunities. One is an amazing team. But with

00:08:29.740 --> 00:08:32.500
a crappy product, you don't think fits the market.

00:08:32.620 --> 00:08:34.740
You don't think maybe we'll get adoption in healthcare,

00:08:35.240 --> 00:08:40.500
but the team is an amazing team. Team B is not

00:08:40.500 --> 00:08:44.259
such a great team, but their product is something

00:08:44.259 --> 00:08:47.000
that fits your thesis. It's an opportunity you've

00:08:47.000 --> 00:08:49.820
identified and it's a really good product and

00:08:49.820 --> 00:08:54.440
it solves the problem. Who do you pick? Can I

00:08:54.440 --> 00:08:57.120
just ask a qualifying question? What do you mean

00:08:57.120 --> 00:09:05.049
a good team? So a team that understands distribution

00:09:05.049 --> 00:09:08.669
and product iteration, and a team that maybe

00:09:08.669 --> 00:09:10.789
has some experience selling into healthcare,

00:09:11.429 --> 00:09:13.309
and a team that has experience maybe building

00:09:13.309 --> 00:09:16.309
software solutions adjacent to healthcare. All

00:09:16.309 --> 00:09:18.309
right, all right. So I think it's an easy answer

00:09:18.309 --> 00:09:21.289
for me, team, team all the way, team all the

00:09:21.289 --> 00:09:25.250
way. And I've had some recent examples, recent

00:09:25.250 --> 00:09:29.419
examples that remind me of that. So product is

00:09:29.419 --> 00:09:32.059
one thing and then timing and why now entering

00:09:32.059 --> 00:09:37.240
all of that is great. But if the team is not

00:09:37.240 --> 00:09:40.419
agile enough, is not resilient enough, you're

00:09:40.419 --> 00:09:42.500
not gonna get anywhere, even if you have the

00:09:42.500 --> 00:09:45.980
best expert or the best product. How do you test

00:09:45.980 --> 00:09:52.779
for that agility and resilience? It's an art,

00:09:52.899 --> 00:09:59.659
right? There's no silver bullet here. You have

00:09:59.659 --> 00:10:02.259
to, during your interactions, your multiple interactions

00:10:02.259 --> 00:10:03.820
with them, you have to ask them some challenging

00:10:03.820 --> 00:10:05.799
questions. You have to see how they respond.

00:10:05.899 --> 00:10:08.620
You have to ask them questions about their past,

00:10:08.659 --> 00:10:12.980
how they dealt with different challenges, especially

00:10:12.980 --> 00:10:15.259
work -related challenges when it comes to this.

00:10:15.419 --> 00:10:21.279
And then try to get a feel for how, if they're

00:10:21.279 --> 00:10:24.500
more of a missionary or a mercenary type of person

00:10:24.500 --> 00:10:26.779
where they're looking for. missionary being,

00:10:26.779 --> 00:10:28.600
you know, I believe in this company and I will

00:10:28.600 --> 00:10:32.279
do whatever it takes for just to see this materialized

00:10:32.279 --> 00:10:36.519
and to see my product being used by my patients

00:10:36.519 --> 00:10:40.179
or clinicians versus, oh, you know, the metrics

00:10:40.179 --> 00:10:42.419
are good and hopefully we're gonna exit in X

00:10:42.419 --> 00:10:49.220
amount of years for this valuation. So yeah,

00:10:49.299 --> 00:10:52.899
it's a bit of an arts resilience is important.

00:10:54.460 --> 00:11:02.519
having some kind of founder market fit and yeah,

00:11:02.919 --> 00:11:05.519
and high IQ, obviously in high EQ. So the way

00:11:05.519 --> 00:11:07.440
that they answer your questions when they get

00:11:07.440 --> 00:11:12.980
challenged, that's gonna be a big deal. And what

00:11:12.980 --> 00:11:16.500
are some common objective attributes you've seen

00:11:16.500 --> 00:11:19.639
in your successful founders? Are there a certain

00:11:19.639 --> 00:11:22.309
age? Do they come from a certain? area? Have

00:11:22.309 --> 00:11:26.429
they had certain life experiences? Maybe they

00:11:26.429 --> 00:11:28.309
used to play video games. I think that's talked

00:11:28.309 --> 00:11:33.230
about quite a bit in the BC ecosystem. Or they

00:11:33.230 --> 00:11:37.110
maybe failed a grade in school. Are there any

00:11:37.110 --> 00:11:41.470
commonalities you've found? A chip on your shoulder

00:11:41.470 --> 00:11:46.669
does help a little bit. Yeah, I have to say.

00:11:47.669 --> 00:11:50.830
So being smart, having a bit of a chip. I was

00:11:50.830 --> 00:11:53.769
wanting to stand out. I think one of the smartest

00:11:53.769 --> 00:11:59.250
founders that I backed is as a middle child in

00:11:59.250 --> 00:12:02.610
a family of doctors who kind of abandoned the

00:12:02.610 --> 00:12:06.370
pre -med path and then decided to go into banking,

00:12:07.029 --> 00:12:09.309
investment banking, which had to explain to her

00:12:09.309 --> 00:12:11.230
family why it's great. And then once they realized

00:12:11.230 --> 00:12:13.169
that it's great and she has a great salary, she's

00:12:13.169 --> 00:12:18.450
like, oh, well, I want to do a startup. And she's

00:12:18.450 --> 00:12:24.129
just raised an oversubscribed series A. So that

00:12:24.129 --> 00:12:29.389
definitely helps. But I wish it was that easy

00:12:29.389 --> 00:12:33.629
to tell you. Yeah, I think it's something I've

00:12:33.629 --> 00:12:36.629
kind of recognized. I've taken the bet on, you

00:12:36.629 --> 00:12:40.269
know, I have these theses in mind, and I don't

00:12:40.269 --> 00:12:42.429
know if it's something I should continue doing.

00:12:42.710 --> 00:12:46.250
It's largely to paint a story for my LPs and

00:12:46.250 --> 00:12:50.809
my syndicate investors. I find if I tell them

00:12:50.809 --> 00:12:53.610
I'm just going to invest in the best team, right?

00:12:53.610 --> 00:12:55.690
I'm going to invest in the best founders building

00:12:55.690 --> 00:12:58.110
in health care. That is my whole thesis. That

00:12:58.110 --> 00:13:01.370
doesn't fly well, doesn't resonate. So I have

00:13:01.370 --> 00:13:04.450
to kind of say, OK, I'm looking for inbox management

00:13:04.450 --> 00:13:07.149
and AI diagnostics or, you know, the ambient

00:13:07.149 --> 00:13:09.549
scribe plus, which is what is the additional

00:13:09.549 --> 00:13:13.809
features here or the AI physician or, you know,

00:13:13.950 --> 00:13:17.690
scheduling or whatever it is, RCM. I need these

00:13:17.690 --> 00:13:23.169
theses or the the the the investors who invest

00:13:23.169 --> 00:13:28.490
in us. They don't buy a vision. Do you kind of

00:13:28.490 --> 00:13:30.570
find that being part of Halo Health and talking

00:13:30.570 --> 00:13:35.279
to physicians like Is thesis -driven investment

00:13:35.279 --> 00:13:40.500
the way? What are your thoughts on that? Yeah,

00:13:40.620 --> 00:13:43.080
that's a great question. So there is the prepared

00:13:43.080 --> 00:13:47.019
mind approach where you look into a market and

00:13:47.019 --> 00:13:49.159
you build your own thesis and then you look for

00:13:49.159 --> 00:13:51.899
companies specifically that solve the problems,

00:13:52.379 --> 00:13:57.120
which is great. It has worked for big funds.

00:13:57.279 --> 00:14:01.500
I think Accel was known for that. But at the

00:14:01.500 --> 00:14:04.820
same time, try doing that exercise going back

00:14:04.820 --> 00:14:07.759
in November of 2022 and saying, I'm looking for

00:14:07.759 --> 00:14:10.700
this. And then in January, well, open AI is here,

00:14:11.019 --> 00:14:14.279
right? So you would have to adjust quite a bit.

00:14:15.860 --> 00:14:20.879
So it is a bit of, you know, it is a mix of things,

00:14:21.080 --> 00:14:27.909
but I do like to have a basic a basic idea, a

00:14:27.909 --> 00:14:31.870
basic investment thesis, but then really team

00:14:31.870 --> 00:14:38.769
is the right team in a big -ish market. That's

00:14:38.769 --> 00:14:40.690
the biggest driver as far as I'm concerned. And

00:14:40.690 --> 00:14:44.250
then market is always, it's also hard to pinpoint

00:14:44.250 --> 00:14:48.929
because the market can increase, the right product

00:14:48.929 --> 00:14:52.529
can increase a market just by being so great

00:14:52.529 --> 00:14:56.659
and having more people use it. When we were speaking

00:14:56.659 --> 00:14:58.860
last year, we made a prediction that there will

00:14:58.860 --> 00:15:01.860
be at least one AI scribe unicorn. There's been

00:15:01.860 --> 00:15:06.240
multiple. And I got close into getting into one

00:15:06.240 --> 00:15:08.860
of them, but I didn't have access to capital,

00:15:08.980 --> 00:15:14.179
so I couldn't get in. But that would have been

00:15:14.179 --> 00:15:19.500
a good investment if I did. What are some, so

00:15:19.500 --> 00:15:21.919
to follow on that, what are some predictions

00:15:21.919 --> 00:15:26.059
you are kind of seeing? at this point and I'll

00:15:26.059 --> 00:15:30.779
kind of share mine after. So I think the AI scribe

00:15:30.779 --> 00:15:34.559
plus is here already. So we're going to have

00:15:34.559 --> 00:15:39.000
AI scribes that are just linked with AI agents.

00:15:40.000 --> 00:15:43.480
And basically everything that the physician does

00:15:43.480 --> 00:15:46.159
other than say this is the diagnosis and this

00:15:46.159 --> 00:15:50.460
is the plan is going to be automated. So you

00:15:50.460 --> 00:15:53.779
would see a physician would see a patient go

00:15:53.779 --> 00:15:57.500
through the interview, come up with a diagnosis

00:15:57.500 --> 00:16:00.279
and then plan and just say to the computer or

00:16:00.279 --> 00:16:04.460
to his cell phone, this is the diagnosis, this

00:16:04.460 --> 00:16:07.039
is the plan, organize a consultation with this

00:16:07.039 --> 00:16:09.100
person, organize a consultation with that person

00:16:09.100 --> 00:16:12.259
and I'll see you in three weeks. And then everything

00:16:12.259 --> 00:16:15.519
after that, it's gonna get done. So the appointment

00:16:15.519 --> 00:16:18.639
is gonna get automated, the medication is gonna

00:16:18.639 --> 00:16:22.370
be sent to the patient. going to be direct messages

00:16:22.370 --> 00:16:24.389
back to the physician if there is an adverse

00:16:24.389 --> 00:16:27.690
reaction. There's going to be an AI agent that

00:16:27.690 --> 00:16:29.909
calls you and tells you exactly where you need

00:16:29.909 --> 00:16:31.970
to go, what you need to do in plain language

00:16:31.970 --> 00:16:35.409
that you understand and gets a feel if you're

00:16:35.409 --> 00:16:39.450
anxious or you're not anxious and is able to

00:16:39.450 --> 00:16:43.309
troubleshoot some basic problems and then only

00:16:43.309 --> 00:16:47.309
escalate to the physician if and only the AI

00:16:47.309 --> 00:16:49.570
agent cannot do it. And then AI agents already

00:16:49.570 --> 00:16:52.830
are quite quite good when you talk to them. It

00:16:52.830 --> 00:16:55.929
is, they feel quite natural and then they're

00:16:55.929 --> 00:16:59.490
able to deal with even you asking them relevant

00:16:59.490 --> 00:17:01.809
questions or you interrupting the conversation.

00:17:01.889 --> 00:17:04.450
So I think all of that, all of that is going

00:17:04.450 --> 00:17:07.369
to get automated. So for, I think for the first

00:17:07.369 --> 00:17:10.410
time we're going to have software that makes

00:17:10.410 --> 00:17:12.690
your life easier and doesn't increase the amount

00:17:12.690 --> 00:17:15.069
of things that you need to do. Obviously it's

00:17:15.069 --> 00:17:16.690
not going to happen all at once and it's not

00:17:16.690 --> 00:17:18.029
going to happen. It's not going to be evenly

00:17:18.029 --> 00:17:21.440
distributed, but I think we're move forward there.

00:17:22.880 --> 00:17:25.759
I completely agree. And do you think this will

00:17:25.759 --> 00:17:30.180
be several point solutions like AI for prescription

00:17:30.180 --> 00:17:32.420
generation and anything that happens after that,

00:17:32.420 --> 00:17:35.640
AI for referrals, or do you kind of foresee one

00:17:35.640 --> 00:17:39.500
company taking on everything that happens during

00:17:39.500 --> 00:17:42.299
a physician visit and follow up? And then that

00:17:42.299 --> 00:17:45.799
ties into patient inbound as well and AI agents

00:17:45.799 --> 00:17:57.549
for, because this is a cycle. It's tough to know

00:17:57.549 --> 00:18:02.809
what the market's going to do, but probably two

00:18:02.809 --> 00:18:05.569
or three big players that end up consolidating

00:18:05.569 --> 00:18:08.009
the smaller players. I think that that's how

00:18:08.009 --> 00:18:12.470
I see it. I don't know if the big players are

00:18:12.470 --> 00:18:16.289
going to start off as scribes or agent and move

00:18:16.289 --> 00:18:20.759
into the other direction. Yeah. And does the

00:18:20.759 --> 00:18:23.579
defensibility here lies in the tech or the distribution?

00:18:25.559 --> 00:18:29.319
I would have to say distribution. You know, unless

00:18:29.319 --> 00:18:32.160
one product is significantly better than the

00:18:32.160 --> 00:18:34.000
other, but I think distribution is probably going

00:18:34.000 --> 00:18:37.460
to be going to be where it's at. And what are

00:18:37.460 --> 00:18:39.440
some hurdles you're seeing for something like

00:18:39.440 --> 00:18:42.640
this, say like an AI automated AI agent that

00:18:42.640 --> 00:18:45.480
does my referrals and prescriptions for me? Like,

00:18:45.619 --> 00:18:48.400
why doesn't that exist right now? Is it a technical

00:18:48.400 --> 00:18:50.799
problem? Is it a regulatory problem? Because

00:18:50.799 --> 00:18:53.539
I don't need it. I'll happily pay for this. I

00:18:53.539 --> 00:18:55.519
don't need to make money off this, right? Like

00:18:55.519 --> 00:18:58.119
AI scribes, I don't make any money off them.

00:18:59.000 --> 00:19:01.660
And that's, you know, in the past, I would have

00:19:01.660 --> 00:19:04.440
said, if you're not generating revenue for me,

00:19:04.859 --> 00:19:07.220
I'm not, I likely will not use your solution.

00:19:07.380 --> 00:19:10.400
I've changed my mind on that. Because the pain

00:19:10.400 --> 00:19:14.069
point with paperwork and referrals and referral

00:19:14.069 --> 00:19:17.029
rejection and you know I get this because I'm

00:19:17.029 --> 00:19:19.630
still in clinic once a day this referral was

00:19:19.630 --> 00:19:23.069
rejected please send it to someone else and that's

00:19:23.069 --> 00:19:26.450
literally 15 clicks and copy pasting five different

00:19:26.450 --> 00:19:28.690
things and attaching the documents it's like

00:19:28.690 --> 00:19:31.549
10 minutes of my time which should be one click

00:19:31.549 --> 00:19:36.410
um sorry can we go back to the to the question

00:19:36.410 --> 00:19:39.170
so the question is um what's what's the difficulty

00:19:39.829 --> 00:19:43.029
Yeah, is it a technical problem at this point?

00:19:43.269 --> 00:19:46.450
Because before you said that the defensibility

00:19:46.450 --> 00:19:48.730
lies in the distribution, not in the tech. So

00:19:48.730 --> 00:19:51.190
to follow on that, and I agree with that, is

00:19:51.190 --> 00:19:53.730
this a technical problem currently, or is this

00:19:53.730 --> 00:19:56.990
more of a regulatory problem about generating

00:19:56.990 --> 00:20:01.069
prescriptions and the agent needing a license

00:20:01.069 --> 00:20:03.609
maybe to prescribe certain medications? And what's

00:20:03.609 --> 00:20:08.880
the path to that? I don't know if it's a false

00:20:08.880 --> 00:20:13.079
dichotomy. I think it's a little bit of, so technically

00:20:13.079 --> 00:20:17.339
the technology is almost there. So I know a startup

00:20:17.339 --> 00:20:19.400
that does just integrate all of your software

00:20:19.400 --> 00:20:22.180
in one. So you don't have to do 15 clicks, but

00:20:22.180 --> 00:20:24.880
you're not able to talk to it, right? So technically

00:20:24.880 --> 00:20:26.880
the technology is almost there, but there's a

00:20:26.880 --> 00:20:29.599
lot of point solutions. So integrating all of

00:20:29.599 --> 00:20:33.440
that into, you know, customizing that to what

00:20:33.440 --> 00:20:39.190
you need, that's tough. So in terms of regulatory

00:20:39.190 --> 00:20:43.210
hurdles, they're fairly big. And mostly it has

00:20:43.210 --> 00:20:45.650
to do with different geographies and different

00:20:45.650 --> 00:20:54.250
regulatory rules. And then there is how to integrate

00:20:54.250 --> 00:20:58.630
workflow. So there's the clinical, that inertia,

00:20:59.009 --> 00:21:02.849
the status quo that becomes, that's still a problem.

00:21:04.319 --> 00:21:10.420
it's a combination of things, but I think you

00:21:10.420 --> 00:21:14.680
have to, I think if you're able to prove to clinicians

00:21:14.680 --> 00:21:19.180
that you're saving them time, you're providing

00:21:19.180 --> 00:21:22.640
value that's 10x what they're doing right now,

00:21:23.619 --> 00:21:27.880
it's not gonna be, that's gonna help a lot. But

00:21:27.880 --> 00:21:30.880
I think if you have your champion, that's the

00:21:30.880 --> 00:21:34.140
best way to go to market. You have one or two,

00:21:34.170 --> 00:21:37.990
champions per institution, they're going to be

00:21:37.990 --> 00:21:43.230
able to move things along. When do you think

00:21:43.230 --> 00:21:46.789
we will see the first AI agent getting the license

00:21:46.789 --> 00:21:51.049
to practice something in healthcare, be it counseling,

00:21:51.269 --> 00:21:53.890
be it physical therapy, be it like prescribing

00:21:53.890 --> 00:21:57.309
a medication or birth control? When do you think

00:21:57.309 --> 00:22:00.430
that will happen? And you know, what's the path

00:22:00.430 --> 00:22:03.440
to that? If I'm, if I'm a founder, and I'm like,

00:22:03.440 --> 00:22:07.319
you know, I speak to Chad GPT or Gemini or an

00:22:07.319 --> 00:22:09.980
AI agent. I'm like, this is really helpful. I

00:22:09.980 --> 00:22:12.559
would love for this to be my counselor and be

00:22:12.559 --> 00:22:15.660
reimbursed by insurance. And the barrier being

00:22:15.660 --> 00:22:19.839
that it needs to be a registered healthcare AI

00:22:19.839 --> 00:22:23.599
agent professional. Like, what is the path to

00:22:23.599 --> 00:22:25.240
that? When do we see that? And am I thinking

00:22:25.240 --> 00:22:27.700
of this the right way? Because my thinking is

00:22:27.700 --> 00:22:32.769
you need that license to unlock revenue. I agree,

00:22:32.809 --> 00:22:37.130
I agree. I don't know if it's easy to make that

00:22:37.130 --> 00:22:43.289
prediction because there's the issue of liability

00:22:43.289 --> 00:22:48.069
and a lot of healthcare delivery is personal

00:22:48.069 --> 00:22:52.970
and people like the human touch. So I could see

00:22:52.970 --> 00:22:55.690
it, let's say over the next five years being

00:22:55.690 --> 00:22:59.380
authorized for re -prescriptions. or in a really

00:22:59.380 --> 00:23:03.319
menial task. So I, or, you know, re -prescription

00:23:03.319 --> 00:23:08.500
of an OCP or contraceptive, I could see it, maybe

00:23:08.500 --> 00:23:12.099
treating something really, really simple, like

00:23:12.099 --> 00:23:15.140
a benign, like a recurrent rash, for example,

00:23:16.000 --> 00:23:20.059
like another eczema flare up. But I think for

00:23:20.059 --> 00:23:24.019
the most parts, the lowest hanging fruit is administrative

00:23:24.019 --> 00:23:26.680
tasks and administrative costs that have increased

00:23:26.680 --> 00:23:29.430
quite a bit. in North America, and I'm guessing

00:23:29.430 --> 00:23:33.109
other markets as well. So that one, I think it's

00:23:33.109 --> 00:23:37.309
a safer bet. Now, yeah, getting a full license,

00:23:38.609 --> 00:23:41.329
but that's a harder one. And there was also a

00:23:41.329 --> 00:23:43.250
recent publication, I think in the New England,

00:23:45.130 --> 00:23:52.430
where chatting with the therapist chatbot was

00:23:52.430 --> 00:23:56.539
shown to be non -inferior, I think, too. compared

00:23:56.539 --> 00:23:59.400
to chatting with a therapist. Even with that,

00:23:59.539 --> 00:24:02.019
I don't think that's enough to change the status

00:24:02.019 --> 00:24:04.460
quo in the near future. Like, and by near, I

00:24:04.460 --> 00:24:08.279
mean in the 2020s. Yeah, I mean, I've been talking

00:24:08.279 --> 00:24:11.700
to Gemini quite a bit. It always tells me I'm

00:24:11.700 --> 00:24:16.400
smart and ask very interesting questions. Yeah.

00:24:16.579 --> 00:24:20.660
And so I think that the motivational interviewing

00:24:20.660 --> 00:24:23.700
techniques are built in in these chatbots, which

00:24:23.700 --> 00:24:26.430
is very interesting. Um, and I had a conversation

00:24:26.430 --> 00:24:28.869
with an AI agent, which was a networking thing

00:24:28.869 --> 00:24:33.289
and it was very, it made me feel good. Um, I

00:24:33.289 --> 00:24:35.329
don't know if it's good for me to be validated

00:24:35.329 --> 00:24:39.609
that much, but it made me feel good because you

00:24:39.609 --> 00:24:42.930
know, the constant validation also will reduce

00:24:42.930 --> 00:24:46.650
your, your stress tolerance. Um, or when you

00:24:46.650 --> 00:24:49.829
get a negative feedback, you won't be used to

00:24:49.829 --> 00:24:52.029
it because your counselor or your best friend.

00:24:52.359 --> 00:24:55.420
or this is something we're seeing, your partner

00:24:55.420 --> 00:24:59.220
who is an AI agent, in some cases, is not providing

00:24:59.220 --> 00:25:04.059
any constructive feedback. Is that something

00:25:04.059 --> 00:25:08.660
you think about? And with the rise of AI agents,

00:25:08.740 --> 00:25:14.539
how do you see the future of mental health? So

00:25:14.539 --> 00:25:16.539
that's an interesting question, because we had

00:25:16.539 --> 00:25:19.519
that discussion with the portfolio company at

00:25:19.519 --> 00:25:25.160
Genesis. My honest opinion is that I don't see

00:25:25.160 --> 00:25:30.740
them replacing a human therapist completely over,

00:25:30.740 --> 00:25:34.559
at least over the next five years. I'm willing

00:25:34.559 --> 00:25:37.680
to bet some money on that, let's say. But it

00:25:37.680 --> 00:25:41.940
is - Why not? You still need that resistance.

00:25:42.599 --> 00:25:45.740
You still need someone to go against, to push

00:25:45.740 --> 00:25:49.900
against your assumptions, your schemas. your

00:25:49.900 --> 00:25:52.819
defense, your ego defense mechanisms, you still

00:25:52.819 --> 00:25:56.720
need that. And do you foresee like payers saying

00:25:56.720 --> 00:26:00.039
you need that to get better? Like, do you foresee

00:26:00.039 --> 00:26:02.900
like data coming out that patients love it? And

00:26:02.900 --> 00:26:05.900
like, I love talking to Gemini, but it's not

00:26:05.900 --> 00:26:08.279
helping my mental health, even though I love

00:26:08.279 --> 00:26:10.500
it. Like, is that kind of you see that data coming

00:26:10.500 --> 00:26:13.660
out at some point? You would need a study that

00:26:13.660 --> 00:26:16.960
shows that outcomes over a long period of time

00:26:16.960 --> 00:26:23.859
are comparable or better. And maybe AI is going

00:26:23.859 --> 00:26:26.240
to go quickly, but there's a lot of nonverbal

00:26:26.240 --> 00:26:29.019
communication that you cannot input into JAD

00:26:29.019 --> 00:26:32.839
GPT, the latest version, or DeepSeq, or Perplexity,

00:26:32.980 --> 00:26:36.880
or whatever you're using. So I don't see it over

00:26:36.880 --> 00:26:40.500
the next five years. I do see a lot of some follow

00:26:40.500 --> 00:26:42.720
-up communication that you may want to have with

00:26:42.720 --> 00:26:45.400
your psychologist. That could be automated if

00:26:45.400 --> 00:26:51.559
you have your CBT, a lot of your journaling could

00:26:51.559 --> 00:26:55.960
be, you can get nudges to do your journaling

00:26:55.960 --> 00:26:58.420
or do the exercises that you agreed upon during

00:26:58.420 --> 00:27:01.140
your CBT session. I could see that being very

00:27:01.140 --> 00:27:11.319
helpful, but not completely. I cannot see a full

00:27:11.319 --> 00:27:15.140
AI agent doing psychotherapy. With that being

00:27:15.140 --> 00:27:19.779
said, at least in developed countries. But then

00:27:19.779 --> 00:27:21.119
if you're talking about the developing world,

00:27:21.220 --> 00:27:24.859
I'm not sure how open the developing world would

00:27:24.859 --> 00:27:29.940
be to having an AI agent as their psychotherapist.

00:27:31.720 --> 00:27:33.960
Talk me through your diligence process. What

00:27:33.960 --> 00:27:36.859
do you look for in startups that kind of come

00:27:36.859 --> 00:27:41.779
in your inbox or are introduced to you? Nothing

00:27:41.779 --> 00:27:48.829
new here. team for sure. Are they complementary?

00:27:49.289 --> 00:27:54.569
Do they have enough experience or enough experience

00:27:54.569 --> 00:27:57.690
in the specific market or in another market,

00:27:58.029 --> 00:28:00.470
but they bring relevant skills to this market?

00:28:01.430 --> 00:28:03.210
Have they known each other for a while? Have

00:28:03.210 --> 00:28:06.930
they worked together because partner, founder

00:28:06.930 --> 00:28:10.450
breakups are a very common cause of failure.

00:28:12.710 --> 00:28:17.480
Then the market needs to be big enough, the timing,

00:28:17.740 --> 00:28:24.319
there needs to be a good answer to why now. And

00:28:24.319 --> 00:28:26.539
some traction or some kind of validation. If

00:28:26.539 --> 00:28:27.759
we're talking about healthcare, there needs to

00:28:27.759 --> 00:28:31.720
be some kind of validation, whether that's some

00:28:31.720 --> 00:28:37.119
IP for a medical device or some beta testing

00:28:37.119 --> 00:28:42.779
or some... at a minimum interviews with the stakeholders

00:28:42.779 --> 00:28:45.279
where they got consistent feedback and now they're

00:28:45.279 --> 00:28:54.700
ready to build. That's about it. I think that's

00:28:54.700 --> 00:28:57.460
the initial screening. And then you have to go

00:28:57.460 --> 00:29:01.640
into due diligence and really look at, are these,

00:29:01.759 --> 00:29:04.299
is this team supporting their claims? What does

00:29:04.299 --> 00:29:06.740
the competition look like? How are they differentiated

00:29:06.740 --> 00:29:13.529
from the competition? and go from there. What

00:29:13.529 --> 00:29:16.730
are you excited about in longevity medicine?

00:29:17.509 --> 00:29:20.869
From what I see, there's this inherent belief

00:29:20.869 --> 00:29:23.690
that we all have, including myself, that more

00:29:23.690 --> 00:29:30.950
data is better, which goes against Western medicine

00:29:30.950 --> 00:29:33.490
in a lot of ways, because we really focus on

00:29:33.490 --> 00:29:37.450
pretest probability. Is this data useful? or

00:29:37.450 --> 00:29:39.150
will it create more uncertainty? And you can

00:29:39.150 --> 00:29:40.970
kind of see this in startups. There's a point

00:29:40.970 --> 00:29:43.009
where you're doing diligence. You don't need

00:29:43.009 --> 00:29:45.829
more data. You will not get more certainty because

00:29:45.829 --> 00:29:48.069
the asset class is uncertain. It's a high -risk

00:29:48.069 --> 00:29:51.029
asset class. Unfortunately, medicine is uncertain.

00:29:51.670 --> 00:29:53.450
Healthcare is uncertain. And I don't think this

00:29:53.450 --> 00:29:56.789
is something patients know or realize that we

00:29:56.789 --> 00:30:00.009
can collect as much data as we want. Prognosis

00:30:00.009 --> 00:30:03.569
and outcome prediction is inherently uncertain.

00:30:03.869 --> 00:30:06.529
That's a probability. It's not deterministic.

00:30:07.220 --> 00:30:11.539
We can never predict your future risk of something

00:30:11.539 --> 00:30:16.779
as 100 % or 0%. And at some point, and I think

00:30:16.779 --> 00:30:19.000
from a business perspective, it's great because

00:30:19.000 --> 00:30:21.900
it creates a data flywheel. You order a test,

00:30:22.180 --> 00:30:24.519
creates uncertainty. You need more testing, more

00:30:24.519 --> 00:30:27.019
uncertainty, more testing. And this is this loop

00:30:27.019 --> 00:30:29.359
where it just keeps going. And so it makes sense.

00:30:29.539 --> 00:30:31.319
And I think it's a massive market opportunity.

00:30:31.420 --> 00:30:36.829
But we'd love to get your thoughts. Yeah, longevity

00:30:36.829 --> 00:30:39.029
is very trendy these days and longevity means

00:30:39.029 --> 00:30:42.329
different things to different people. So anything

00:30:42.329 --> 00:30:45.650
from I'm just getting blood tests or extra tests

00:30:45.650 --> 00:30:50.269
more often to, well, here's why I'm using metformin

00:30:50.269 --> 00:30:55.269
and rapamycin and then this new, I don't know,

00:30:55.390 --> 00:30:59.910
NAD injection and stem cell injection. So it

00:30:59.910 --> 00:31:01.769
means different things to different people at

00:31:01.769 --> 00:31:05.940
this point. I'm very happy there's interest in

00:31:05.940 --> 00:31:11.000
this segment. To answer your question, yes, so

00:31:11.000 --> 00:31:13.319
traditional Western medicine and evidence -based

00:31:13.319 --> 00:31:15.859
medicine stresses the fact that there is uncertainty

00:31:15.859 --> 00:31:20.779
and that's great, but a couple of things here.

00:31:20.799 --> 00:31:26.019
So if you have better data, then you can change

00:31:26.019 --> 00:31:28.359
your pre -test and post -test probability. What

00:31:28.359 --> 00:31:31.779
do I mean by better data? So full body MRIs may

00:31:31.779 --> 00:31:36.329
give you some false positives, but they will

00:31:36.329 --> 00:31:39.569
cover some edge cases. So they will cover the

00:31:39.569 --> 00:31:42.609
pancreatic cancer. They will cover the glioblastoma

00:31:42.609 --> 00:31:47.549
multiforme, like a rare brain tumor. And if you

00:31:47.549 --> 00:31:53.410
have, you know, a full body MRI at age 50 and

00:31:53.410 --> 00:31:56.569
it's completely clean, you have an idea that

00:31:56.569 --> 00:31:59.910
your presence probability of having, if you go

00:31:59.910 --> 00:32:03.109
do another test, another, let's say, abdominal

00:32:03.109 --> 00:32:07.559
ultrasound in two years, the chances of having

00:32:07.559 --> 00:32:09.819
pancreatic cancer have gone down quite a bit,

00:32:10.299 --> 00:32:13.279
just based on the full body MRI. So I think with

00:32:13.279 --> 00:32:16.380
technology advancing and technology becoming

00:32:16.380 --> 00:32:22.140
more affordable, a lot of preventable deaths

00:32:22.140 --> 00:32:26.019
will go down. And evidence -based medicine so

00:32:26.019 --> 00:32:29.599
far has been very useful because we went from...

00:32:29.640 --> 00:32:34.019
this is just what my opinion is, or this is what

00:32:34.019 --> 00:32:36.259
this KOL does, or this is how we've always done

00:32:36.259 --> 00:32:38.859
it, to okay, this is what the research shows,

00:32:38.880 --> 00:32:44.740
and this is how we use best practices and the

00:32:44.740 --> 00:32:49.059
best research to harmonize treatment and diagnosis,

00:32:49.319 --> 00:32:52.059
which was absolutely necessary in the 90s and

00:32:52.059 --> 00:32:56.069
up to now. But now with technology going... evolving

00:32:56.069 --> 00:32:59.450
really fast and hopefully costs coming down,

00:33:00.049 --> 00:33:05.109
we can have more useful data and cover edge cases

00:33:05.109 --> 00:33:09.109
more easily. So the rare cancers, but also in

00:33:09.109 --> 00:33:12.329
terms of prevention, we can catch cardiovascular

00:33:12.329 --> 00:33:16.349
disease a little bit early or we can catch cancer.

00:33:17.069 --> 00:33:21.609
We can catch hopefully some autoimmune, we can

00:33:21.609 --> 00:33:24.650
prevent some autoimmune diseases from developing.

00:33:26.399 --> 00:33:31.440
So I think that just having cheaper access to

00:33:31.440 --> 00:33:34.359
diagnostics and wider access to diagnostics is

00:33:34.359 --> 00:33:41.740
gonna help just avoid preventable deaths. Now,

00:33:41.740 --> 00:33:43.920
is that gonna happen right away? No, is that

00:33:43.920 --> 00:33:46.980
gonna be evenly distributed right away? No, but

00:33:46.980 --> 00:33:49.599
that's where I think we are going because just

00:33:49.599 --> 00:33:52.420
looking at the latest meta -analysis and just

00:33:52.420 --> 00:33:56.680
trying to... base our clinical guidelines on

00:33:56.680 --> 00:34:00.720
that, already you've lost, there's a lag, okay?

00:34:01.960 --> 00:34:03.880
And your statistics are what they are, that they're

00:34:03.880 --> 00:34:05.519
gonna cover most patients, but they're still

00:34:05.519 --> 00:34:07.420
gonna be edge cases. So we're gonna have fewer

00:34:07.420 --> 00:34:14.099
edge cases just by having more useful data. And

00:34:14.099 --> 00:34:16.739
there are some examples already of a technology

00:34:16.739 --> 00:34:21.219
really changing. the way diseases at least are

00:34:21.219 --> 00:34:23.360
managed and hopefully in the future are prevented.

00:34:23.579 --> 00:34:26.539
So CGM is already helping with a lot with diabetes

00:34:26.539 --> 00:34:29.920
and you have a lot more data, but it's useful

00:34:29.920 --> 00:34:32.320
data and people are able to titrate their medications

00:34:32.320 --> 00:34:35.219
better and then to have more useful discussions

00:34:35.219 --> 00:34:39.099
with their endocrinologist. So wearables are

00:34:39.099 --> 00:34:42.280
going to give us increasingly more useful information.

00:34:42.820 --> 00:34:46.219
So the Apple watch was just approved for diagnosed

00:34:46.219 --> 00:34:49.750
sleep apnea. it's already been approved to diagnose

00:34:49.750 --> 00:34:53.030
atrial fibrillation. And I've had two of my patients

00:34:53.030 --> 00:34:56.889
get diagnosed themselves with their Apple watch.

00:34:57.650 --> 00:35:02.630
And now I'm excited because sleep apnea affects

00:35:02.630 --> 00:35:05.010
about 10 % of the population. And you go through

00:35:05.010 --> 00:35:12.219
this usual questionnaire that's not very helpful

00:35:12.219 --> 00:35:15.719
to try to figure out whether they have sleep

00:35:15.719 --> 00:35:17.880
apnea or not. And the current standard of care

00:35:17.880 --> 00:35:19.480
is that they just have to go to clinic, pick

00:35:19.480 --> 00:35:22.460
up a machine, take their home, and then send

00:35:22.460 --> 00:35:25.619
the machine back, then see a specialist and then

00:35:25.619 --> 00:35:29.119
have treatment delivered to, or have access to

00:35:29.119 --> 00:35:32.599
treatment. Now they can come directly to me and

00:35:32.599 --> 00:35:35.780
then I can send to see the specialist. If they

00:35:35.780 --> 00:35:38.599
wanna do some more testing, fine, but at least

00:35:38.599 --> 00:35:44.119
it's picked up on time. If you had infinite funding

00:35:44.119 --> 00:35:47.119
to build a startup in this space, how would you

00:35:47.119 --> 00:35:54.019
begin? In the preventative medicine space, you

00:35:54.019 --> 00:35:56.019
know, you could have molecular solutions where,

00:35:56.079 --> 00:35:58.639
you know, just go directly at the mitochondrion

00:35:58.639 --> 00:36:02.360
and then, you know, just reverse aging or stop

00:36:02.360 --> 00:36:05.880
disease at the cellular level. I don't have any

00:36:05.880 --> 00:36:08.019
insight. I don't have a lot of insight into that.

00:36:08.039 --> 00:36:10.619
And I'm glad that people are looking into that.

00:36:11.000 --> 00:36:12.900
But if we're talking about, you know, with what

00:36:12.900 --> 00:36:19.380
we have now, it would be... Would you do like

00:36:19.380 --> 00:36:22.320
a NECO style from Sweden, like which is like

00:36:22.320 --> 00:36:24.559
a clinic or would you do like a function health

00:36:24.559 --> 00:36:28.960
or something that's more, you know, telehealth?

00:36:30.000 --> 00:36:32.000
They're converging. The two are converging because

00:36:32.000 --> 00:36:33.619
function health, I think, just bought a full

00:36:33.619 --> 00:36:38.349
body MRI a startup. So I think we are going in

00:36:38.349 --> 00:36:43.329
that direction. And NECO health is using a lot

00:36:43.329 --> 00:36:47.329
of money. So I think that's where we're going.

00:36:47.630 --> 00:36:52.630
So which means regular blood tests and full body

00:36:52.630 --> 00:36:54.929
scans. I don't know if I would do them every

00:36:54.929 --> 00:36:57.510
year, but you could do maybe full body scans

00:36:57.510 --> 00:36:59.869
every few years and then in between some liquid

00:36:59.869 --> 00:37:02.670
biopsies that work better than what we have right

00:37:02.670 --> 00:37:06.079
now because nothing works really well. from the

00:37:06.079 --> 00:37:09.500
blood test that existed in the market. So I think

00:37:09.500 --> 00:37:12.480
that's, yeah, that's very feasible in the next

00:37:12.480 --> 00:37:14.699
few years until you have the next revolution

00:37:14.699 --> 00:37:21.420
at the molecular level. Let's say I'm a physician,

00:37:21.679 --> 00:37:24.599
you know, I graduated from high school, went

00:37:24.599 --> 00:37:26.980
to undergrad for pre -med, did med school, did

00:37:26.980 --> 00:37:28.960
residency, did fellowship, and I'm attending.

00:37:29.579 --> 00:37:32.139
The only thing I know is medicine. I know nothing

00:37:32.139 --> 00:37:37.110
else. And I have a problem in my workspace and

00:37:37.110 --> 00:37:39.570
I'm thinking, you know, this is a solution that

00:37:39.570 --> 00:37:43.710
could work. And then I come across A16Z and Sequoia

00:37:43.710 --> 00:37:45.869
and I'm excited and I'm like, OK, this looks

00:37:45.869 --> 00:37:49.590
promising. I want to build a startup. How should

00:37:49.590 --> 00:37:53.230
I go about what are some mistakes I'm going to

00:37:53.230 --> 00:37:56.969
make that I can avoid? And what are some mistakes

00:37:56.969 --> 00:37:59.570
I'm just going to make anyways that are unavoidable?

00:38:00.090 --> 00:38:04.269
What advice do you have for this physician? So

00:38:04.269 --> 00:38:08.949
a lot of the physicians that I see are exactly

00:38:08.949 --> 00:38:10.889
that and they tend to be academic physicians

00:38:10.889 --> 00:38:14.710
and they want to build something based on their

00:38:14.710 --> 00:38:18.550
research. And the most common mistake is just

00:38:18.550 --> 00:38:20.690
trying to do everything themselves and trying

00:38:20.690 --> 00:38:25.070
to recruit among either their students or some

00:38:25.070 --> 00:38:28.510
undergrad or someone who doesn't have enough

00:38:28.510 --> 00:38:31.780
expertise. there's a lot of not knowing what

00:38:31.780 --> 00:38:36.179
you do not know. This is the most common problem

00:38:36.179 --> 00:38:38.559
that I see and trying to keep all of the equity

00:38:38.559 --> 00:38:42.539
to yourself and not understanding that this is

00:38:42.539 --> 00:38:46.159
a completely different game. And your skills

00:38:46.159 --> 00:38:50.039
have made you an expert in what you do, but building

00:38:50.039 --> 00:38:53.219
a company going from, so I have good technology

00:38:53.219 --> 00:38:56.500
or good research or maybe even good IP, going

00:38:56.500 --> 00:39:00.550
from that to I have a successful company that

00:39:00.550 --> 00:39:04.670
is a completely different game. And there's ego

00:39:04.670 --> 00:39:08.829
that gets involved here, or there's just ignorance

00:39:08.829 --> 00:39:11.630
of what we don't know, which is normal. We all

00:39:11.630 --> 00:39:14.969
have blind spots. But the number one thing would

00:39:14.969 --> 00:39:17.829
be to build a team around you, understand that

00:39:17.829 --> 00:39:19.809
you don't have commercialization experience.

00:39:20.449 --> 00:39:22.789
You will need someone to take care of technology,

00:39:22.949 --> 00:39:26.489
even if you are the chief scientific officer.

00:39:27.849 --> 00:39:30.849
You need people to build a commercially minded

00:39:30.849 --> 00:39:37.170
team around you. So let's say that you get over

00:39:37.170 --> 00:39:39.389
that. The other thing to understand is that someone

00:39:39.389 --> 00:39:42.809
who worked at a big pharma company and has 20

00:39:42.809 --> 00:39:45.070
years of experience is not necessarily the best

00:39:45.070 --> 00:39:47.409
co -founder for you. You need someone who has

00:39:47.409 --> 00:39:49.449
some technical skills or some commercialization

00:39:49.449 --> 00:39:52.849
experience, but you need someone who understands

00:39:52.849 --> 00:39:55.269
startups and the uncertainty of startups and

00:39:55.269 --> 00:39:58.420
the resilience that's required. and to be able

00:39:58.420 --> 00:40:03.539
to tolerate rejection, whether that's from investors

00:40:03.539 --> 00:40:06.440
or from the initial customers where you're going

00:40:06.440 --> 00:40:13.239
to have setbacks and you're starting on a new

00:40:13.239 --> 00:40:17.519
adventure essentially and then you need the best

00:40:17.519 --> 00:40:20.960
team and you need to realize that your previous

00:40:20.960 --> 00:40:29.190
career has not prepared you for this. If your

00:40:29.190 --> 00:40:33.949
kids came to you and said, dad, I want to be

00:40:33.949 --> 00:40:36.969
an emergent medicine doctor, what would you tell

00:40:36.969 --> 00:40:45.250
them? My kids too. So that would be in 16 years.

00:40:50.530 --> 00:40:55.070
I would say just get a good idea of what it is.

00:40:55.869 --> 00:41:01.110
just go shadow a few emergency physicians and

00:41:01.110 --> 00:41:04.030
then, you know, talk to people who are early

00:41:04.030 --> 00:41:06.110
in their career, talk to people who are late

00:41:06.110 --> 00:41:11.170
in their career, get a feel for, you know, the

00:41:11.170 --> 00:41:18.289
profession and then get a feel for, you know,

00:41:18.829 --> 00:41:24.309
and see if you like it. But I would start the

00:41:24.309 --> 00:41:27.860
real Question here is, is why? Okay. And I'm

00:41:27.860 --> 00:41:29.760
trying to really get to the, what do you like

00:41:29.760 --> 00:41:33.460
about it? What is it that drives you to do this?

00:41:33.760 --> 00:41:36.460
What are your strengths? What are your weaknesses?

00:41:37.019 --> 00:41:40.239
And then is this the only thing that will make

00:41:40.239 --> 00:41:44.739
you happy? I think starting with knowing yourself

00:41:44.739 --> 00:41:54.670
and strengths, weaknesses, ambitions. Interest,

00:41:55.329 --> 00:41:58.170
once you get to know that, I think then you can

00:41:58.170 --> 00:42:01.510
see that there's a lot of ways to be happy or

00:42:01.510 --> 00:42:04.750
to have a successful career or whatever your

00:42:04.750 --> 00:42:07.730
ambition is to make a lot of money or have a

00:42:07.730 --> 00:42:13.389
big impact in society. And once you know yourself,

00:42:14.130 --> 00:42:16.369
you have to identify at least a few solutions.

00:42:16.530 --> 00:42:18.630
It cannot just be, I will be an emergency physician

00:42:18.630 --> 00:42:24.480
or die. identify a few solutions, go try them

00:42:24.480 --> 00:42:27.280
out. If you think emergency medicine is for you,

00:42:27.539 --> 00:42:31.079
go for it. And just because you did it for 10

00:42:31.079 --> 00:42:32.840
years or 12 years doesn't mean that you're married

00:42:32.840 --> 00:42:37.920
to it for life. So that's how I would address

00:42:37.920 --> 00:42:48.539
that. I would say no. That's been my answer for

00:42:48.539 --> 00:42:51.199
a few. for a few years, but my wife tells me,

00:42:51.440 --> 00:42:53.920
no, no, no, she has to walk me back from that.

00:42:54.400 --> 00:42:59.460
So. Yeah, I think my 12 year old, she has an

00:42:59.460 --> 00:43:02.739
interest in science and I can see, and even the

00:43:02.739 --> 00:43:05.320
five year old that they're starting to kind of

00:43:05.320 --> 00:43:09.440
pick on things and starting to go the medical

00:43:09.440 --> 00:43:11.659
route. And I'm trying to kind of sway them away,

00:43:11.900 --> 00:43:16.579
but we'll see if I win that battle. What brings

00:43:16.579 --> 00:43:22.230
you more satisfaction? and joy investing in a

00:43:22.230 --> 00:43:25.409
team that needs coaching or investing in a founding

00:43:25.409 --> 00:43:28.809
team that you just your hands off? Do you think

00:43:28.809 --> 00:43:34.190
the job is to help founders specifically coaching

00:43:34.190 --> 00:43:36.690
with, you know, scaling and team building and

00:43:36.690 --> 00:43:40.150
management of that sort, or is the job to just

00:43:40.150 --> 00:43:41.989
pick the best founders and get out of the way

00:43:41.989 --> 00:43:43.489
and maybe help them a little bit here and there,

00:43:43.489 --> 00:43:51.539
but you're not involved that deeply? No, picking

00:43:51.539 --> 00:43:53.639
the best founders and getting out of their way.

00:43:54.019 --> 00:43:57.440
Like there's not, I don't care to maximize my

00:43:57.440 --> 00:44:00.500
impact on within that team, picking the best

00:44:00.500 --> 00:44:05.139
founders. It's a difficult decision anyway, just

00:44:05.139 --> 00:44:07.920
sometimes just even finding the right team is

00:44:07.920 --> 00:44:10.820
hard. And then let alone doing due diligence

00:44:10.820 --> 00:44:15.719
and then negotiating and everything else that

00:44:15.719 --> 00:44:18.280
goes with the deal. So no, picking the right

00:44:18.280 --> 00:44:21.719
team. getting out of their way, that's absolutely

00:44:21.719 --> 00:44:24.980
fine. Now, if I can help in any way, that's fine,

00:44:24.980 --> 00:44:28.039
but I do not want to be involved in the day -to

00:44:28.039 --> 00:44:30.920
-day. And it doesn't, you know, if you have a

00:44:30.920 --> 00:44:33.579
lot of big portfolio, it's not really realistic

00:44:33.579 --> 00:44:35.340
to say, oh, well, I'm going to help you with

00:44:35.340 --> 00:44:39.059
this and that, or I'm going to be doing part

00:44:39.059 --> 00:44:41.320
of the work. Mind you, over the last six months,

00:44:41.320 --> 00:44:44.760
I have been helping a company, but, so maybe

00:44:44.760 --> 00:44:50.389
that's colored by that, but. No, just it's so

00:44:50.389 --> 00:44:53.269
delightful just getting those updates and things

00:44:53.269 --> 00:44:55.869
are going up into the rights and then, you know,

00:44:55.929 --> 00:45:00.530
this new opportunity that gets unlocked or there's

00:45:00.530 --> 00:45:05.489
interest by this acquirer. That's venture is

00:45:05.489 --> 00:45:10.429
hard enough as it is. So I prefer to have picked

00:45:10.429 --> 00:45:16.429
a good team. If you had to pick out of investing

00:45:16.429 --> 00:45:19.309
in medical device, versus digital health, what

00:45:19.309 --> 00:45:24.469
would you pick? Oh, medical device is hard. So

00:45:24.469 --> 00:45:28.650
definitely digital health. Medical device, you

00:45:28.650 --> 00:45:33.730
have all the technical risk of building something,

00:45:34.030 --> 00:45:35.989
going to regulatory, and then you have to build

00:45:35.989 --> 00:45:39.489
distribution, which is very hard. Which do you

00:45:39.489 --> 00:45:42.210
think is harder, FDA approval or distribution

00:45:42.210 --> 00:45:45.949
and market adoption? I think distribution is

00:45:45.949 --> 00:45:48.960
harder. I think distribution is harder. Interesting,

00:45:49.380 --> 00:45:53.099
because we celebrate FDA approval so much, you

00:45:53.099 --> 00:45:56.519
know, from it almost feels like the startup has

00:45:56.519 --> 00:45:59.920
succeeded post -approval. They haven't, but I

00:45:59.920 --> 00:46:02.119
think that that's what the, and maybe it's because

00:46:02.119 --> 00:46:04.260
both things are so hard, we need to celebrate

00:46:04.260 --> 00:46:07.239
our successes. The only thing that I'm going

00:46:07.239 --> 00:46:12.219
to say here is, oh, you could just get acquired

00:46:12.219 --> 00:46:16.769
as soon as you get your FDA approval, and That's,

00:46:16.789 --> 00:46:21.170
I think that's a great path to exit. There are

00:46:21.170 --> 00:46:23.369
not a lot of players that will acquire you. Okay,

00:46:23.510 --> 00:46:26.349
that's been consolidated to four or five players

00:46:26.349 --> 00:46:30.710
that do acquire a medical device. So you need

00:46:30.710 --> 00:46:32.829
to probably have built some relationship with

00:46:32.829 --> 00:46:35.449
one of them or a few of them from before you

00:46:35.449 --> 00:46:40.110
get your FDA approval. But, you know, selling

00:46:40.110 --> 00:46:43.690
after that, It's hard. Building your Salesforce,

00:46:43.829 --> 00:46:47.869
going after KOLs, that takes a lot of capital,

00:46:48.110 --> 00:46:52.150
a lot of energy, and there's a lot of challenges

00:46:52.150 --> 00:46:55.650
there. And it takes a bit of different skillset

00:46:55.650 --> 00:46:58.409
from the team, because your CTO is not as active,

00:46:58.630 --> 00:47:01.769
the technical team is maybe improving the device

00:47:01.769 --> 00:47:05.849
a little bit, but you need to hire a second team.

00:47:06.710 --> 00:47:11.250
Whereas digital health, faster go -to -markets,

00:47:11.989 --> 00:47:17.210
no defensibility, yes. You may or may not be

00:47:17.210 --> 00:47:22.210
able to be as sticky as a medical device. But

00:47:22.210 --> 00:47:29.150
generally, with AI, with a lot of the inefficiencies

00:47:29.150 --> 00:47:32.849
in healthcare, I find it's a better investment.

00:47:34.030 --> 00:47:36.510
And it definitely fits my background better.

00:47:38.510 --> 00:47:42.550
Is it easier to teach entrepreneurship and business

00:47:42.550 --> 00:47:46.150
to a healthcare expert or physician or to teach

00:47:46.150 --> 00:47:50.570
healthcare to a business or entrepreneurship

00:47:50.570 --> 00:47:55.789
expert? I think it's easier to have help on board,

00:47:56.909 --> 00:48:00.889
to have healthcare expert or healthcare expertise

00:48:00.889 --> 00:48:03.429
to have access to that when you're an entrepreneur.

00:48:04.610 --> 00:48:06.989
Becoming an entrepreneur, there's a lot of...

00:48:06.920 --> 00:48:11.119
personality traits that go with it that you medicine,

00:48:11.380 --> 00:48:15.000
clinical medicine selects out for it and that

00:48:15.000 --> 00:48:18.340
are actively discouraged. So I would prefer an

00:48:18.340 --> 00:48:23.159
entrepreneur who has a great advisor and then

00:48:23.159 --> 00:48:26.139
a few clinicians on his team, leading a team

00:48:26.139 --> 00:48:29.860
as opposed to a physician by himself or I don't

00:48:29.860 --> 00:48:33.739
know, maybe a physician who's well surrounded

00:48:33.739 --> 00:48:36.300
could do it, but generally I would I would invest

00:48:36.300 --> 00:48:40.679
more in an entrepreneur. OK. Last question, Tennessee.

00:48:41.119 --> 00:48:43.820
If you could go back in time and talk to your

00:48:43.820 --> 00:48:46.519
20 -year -old self, what would you tell him?

00:48:49.199 --> 00:48:51.300
One thing is certain things are going to change

00:48:51.300 --> 00:48:53.599
a lot. And you're not going to be able to predict

00:48:53.599 --> 00:49:00.059
how they're going to change. So just stay flexible.

00:49:00.340 --> 00:49:05.019
Just stay flexible. And success is just, you

00:49:05.019 --> 00:49:08.860
know, or happiness going from failure to failure

00:49:08.860 --> 00:49:11.400
without a loss of morale. Just stay flexible,

00:49:11.599 --> 00:49:16.659
stay resilient, and you're gonna end up ahead

00:49:16.659 --> 00:49:19.179
in the end. Thank you.