Latest / The Advancing Healthcare Innovation Show / E173 | Why Healthcare Leaders Still Feel Stuck Even When the Data Is Better

Transcript

E173 | Why Healthcare Leaders Still Feel Stuck Even When the Data Is Better · The Advancing Healthcare Innovation Show

0:10
Hello and welcome back to the Advancing
0:12
Healthcare Innovation Show. I am your
0:14
host, Michael Samatinos, and if you are
0:17
new to us, welcome. It's great to have
0:20
you here. We're really excited about a
0:22
few different things because we're
0:24
bringing some just incredible
0:26
conversations, some real leaders that
0:28
are painstakingly
0:29
reshaping the future of health. And
0:32
we're very focused and passionate about
0:34
a few different areas. We're passionate
0:36
about healthcare innovation, leadership,
0:38
about access, about adoption. And if
0:41
it's your first time tuning in, let me
0:43
just kind of peel back the curtains a
0:45
little bit and tell you a little bit
0:45
more about what we're trying to get done
0:47
here. So our focus is about bringing
0:49
real stories of innovation to life, real
0:51
stories of leadership to life, not just
0:53
things that you read in a book, but like
0:54
real people that are actually getting
0:56
stuff done. And the aim is to highlight
0:59
them and share their stories with our
1:01
community and ultimately with the world.
1:04
And if you haven't had the opportunity
1:06
to check out any of our historical
1:07
interviews, wherever you are, if you're
1:09
watching this, if you're listening to
1:11
this, please just hit the subscribe
1:14
button now so that any new interviews
1:17
come directly to you. So look,
1:22
healthcare's got no shortage of data.
1:24
Um, what it lacks in my opinion is
1:27
follow through. And our next guest has
1:31
spent his entire career, years working
1:34
in that gap. Dr. Khan Sadiki is the CEO,
1:39
a founder and chairman of Hopper. So
1:42
Khan is a physician by training who
1:44
built his career around really that hard
1:48
truth. When information doesn't move,
1:51
care doesn't either. And when that
1:53
happens, people get missed. And this
1:57
idea has shaped everything that he's
2:00
done. So before Hopper, he worked on
2:03
bringing preventative care into everyday
2:05
life. Think about, you know, grocery
2:07
stores,armacies, real places where scale
2:11
exposed what actually works and what
2:14
quietly breaks. So along the way he's
2:16
been invited into uh every one of the
2:18
rooms you can think of panels uh big
2:22
stages where hard conversations happens
2:25
sim you know not necessarily for him to
2:28
sell a vision but to wrestle with
2:30
reality what can we do when the data is
2:33
there but the trust isn't you know when
2:36
systems talk but people still can't act.
2:40
So, that's the tone of today's
2:42
conversation. That's not necessarily
2:44
going to be a highlight reel, although
2:46
I'm going to do my best to try to
2:48
extract as many stories as possible from
2:50
Khan because he's a phenomenal
2:51
storyteller. Um, I want to talk to him
2:54
about like just judgment, trade-offs, a
2:57
lot of different stuff. So, ideally, um,
3:00
we'll talk about what really matters.
3:02
And if you're looking for easy answers,
3:05
uh, this probably I don't know if this
3:07
is going to be the episode for you. it's
3:08
going to be probably derive more
3:10
thought, more questions. But if you care
3:11
about access, if you care about
3:13
population health and building things
3:15
that are durable, uh, that hold up in
3:18
real in the real world, then you're
3:20
going to want to stay with us. So Khan,
3:22
it's great to see you. Welcome to the
3:24
show.
3:25
>> Thanks, Michael. It's pleasure to be
3:27
here. Thanks for inviting me and doing
3:28
this.
3:29
>> Yeah, of course. Of course. So before
3:32
before we dig into um your just amazing
3:36
background about companies and
3:38
platforms, I kind of want to start
3:40
somewhere that's um that's closer to
3:42
home. Um you shared with me that you
3:45
know your mom's a physician, she still
3:47
practices today. She still runs a clinic
3:51
in Pakistan.
3:53
What was it like uh growing up and
3:57
watching her do medicine that way?
4:02
>> Yeah, thanks for bringing that up. You
4:06
know, my mom's both my parents were
4:08
physician and both have been my role
4:09
models, right? And uh I was having this
4:12
conversation recently with somebody.
4:13
She's 89, spends about six months with
4:16
me and six months in Pakistan and she's
4:18
there. She runs their clinic. she wishes
4:20
she could still operate
4:23
but I don't think she's uh you know and
4:26
then when she's here her entertainment
4:28
is keeping up with academic literature
4:31
in OBGYn and so she's up to speed on
4:33
what the latest things are happening and
4:35
and it's pretty fascinating that at her
4:37
age she's has her iPad on and um you
4:40
know um ju just showing what norm is I
4:45
think that's what happened for me both
4:48
my parents being physicians and having
4:50
this mentality of uh giving.
4:55
>> They both trained in the UK, right? So
4:57
they didn't have to go back to Pakistan
4:59
and serve communities. And um and I
5:02
always you know heard from her when I
5:04
was young that
5:06
>> you know be physician to help people and
5:09
care not to make money.
5:12
So if somebody showed her showed up at
5:14
her clinic or in her practice didn't
5:17
have money like I remember so many times
5:19
I'll get a you know she will ping me
5:21
like hey go get in the drawer their
5:23
money sitting here this patient doesn't
5:25
have
5:26
>> money to go back home
5:28
>> and so not only not charge her but
5:30
actually pay them to so that they can go
5:32
back home like a lot of these folks you
5:34
know unfortunately show up uh in those
5:37
part of the world when you know no other
5:39
option is available to them and uh and
5:43
then she always has the mentality like
5:45
if I can't do anything about it and why
5:46
I shouldn't be charging
5:49
them for giving them bad news. So, so
5:53
for me what had happened was
5:56
>> that kind of mentality became the norm
5:58
as I was growing up right so like my
6:00
expectation like that that became the
6:02
baseline you know you think of when you
6:04
were working on your healthcare like
6:05
your health you know it talks about your
6:07
base right what's your base cardio is
6:10
>> like that became my base cardio of
6:12
mentality of thinking and how do you
6:14
create impact so
6:15
>> for me
6:17
>> uh and then then both my parents made an
6:19
effort to make sure that we are engaged,
6:22
we are they're present.
6:25
>> Um, you know, so I never never felt
6:28
like, you know, yeah, we saw them work
6:29
really hard, but that doesn't mean they
6:32
weren't available to us, right? They
6:33
attended every single, you know,
6:36
activity I was doing in school. Uh me
6:38
and my both my younger sisters
6:41
uh you know like going you know with my
6:46
father swimming and doing all the race
6:49
activities in the evenings and weekends.
6:51
>> Um it was great but you know this idea
6:54
of or the work ethic
6:57
>> was very important
6:59
>> and kind of created that foundation
7:01
where I kind of am. Um and so
7:06
so I think that's all it is, right? It
7:08
just normalizes
7:10
um has normalized that for me and uh
7:13
hopefully I've done the same thing with
7:14
my kids. Um I think a lot of times when
7:18
we talk about work life balance
7:22
uh when I joined Microsoft uh we'll talk
7:25
about that more. I struggle with work
7:27
life balance type stuff, right? Like
7:29
yeah, you know, I was having a lot of
7:31
issues. I have two young kids at the
7:33
time and then one of my mentors at
7:34
Microsoft said like I think you're
7:35
taking this wrong.
7:37
Uh if you're trying to separate this out
7:40
then neither does your work know what's
7:41
happening in your family life and
7:43
neither the family knows what happening
7:44
in your work life and if they don't
7:46
understand various
7:48
>> you know stress and work is not constant
7:51
right there ups and downs.
7:52
>> Same thing in the family life right?
7:55
>> Yeah. So his guidance was
7:57
>> his guidance was you need to connect
7:59
them together
8:01
>> and that's what I ended up doing right
8:03
so like
8:04
>> our dinner starts like how was the day
8:08
what happened and we talk pretty
8:09
detailed deeply like my wife talks about
8:11
she's a radiologist too like what she
8:13
did what things were happening what was
8:15
frustrating what was good what was the
8:17
highlights the kids talk about the
8:19
school stuff I talk about my work stuff
8:21
and then same thing at work we always
8:23
have with the leadership team talk
8:25
about, you know, what's happening in our
8:28
at our homes. So then
8:31
we know when an important event is
8:32
coming up in one of the leaderships, so
8:35
we know that we don't want to bug them.
8:37
>> Like my co-founder Robert had to drop
8:39
his son off to New York. He's, you know,
8:41
he's graduated. He's moving there.
8:43
>> We all knew he's doing move on Monday,
8:45
Tuesday.
8:46
>> I'm not going to bug him because he
8:48
needs to pay attention to that. And we
8:50
moved things around to make that work.
8:52
He didn't have to say anything because
8:53
we already knew that has happened.
8:54
Right.
8:55
>> So that's kind of Yeah.
8:57
>> You know what's cool about that is it's
8:59
uh you know we talk a lot in in medicine
9:01
we talk we talk about whole person care
9:04
>> and the the proactive care and the fact
9:09
that you're able to to kind of have this
9:11
work life harmony is uh it's sort of a
9:16
responsibility. It's it's as if you're
9:18
taking the training that your mom and
9:20
father, your mom and dad demonstrated to
9:23
you about how to care for people and
9:26
you're now just applying that into your
9:28
businesses and your work life.
9:30
Obviously, sometimes it's a little it's
9:32
harder for the lessons to kind of cement
9:34
themselves into and there's a lot of
9:36
trial and error,
9:38
>> but that tells me that there's a there's
9:40
a there's a weight and a responsibility
9:42
that you carry and you can't just
9:45
separate the two, you know, home life
9:47
and work life, they're comingled and
9:51
like you got to pay attention to that.
9:52
If if something's going on at home, like
9:54
you're obviously not going to bring your
9:56
best effort on the work front. So, it's
9:59
it's cool that there's a level of
10:01
transparency and like honesty
10:03
about the ups and downs, highs and lows
10:05
because that is life. You're not always
10:07
winning all the time.
10:10
>> You talked about my mom stuff, right?
10:12
So, I think a lot of things that
10:13
happened involuntarily,
10:16
the fact that we named our company at,
10:19
you know, um at one of the famous women
10:21
in STEM, Grace Hopper,
10:24
I wasn't thinking mom, but guess what?
10:26
That probably was in my back of my mind.
10:30
>> Yeah.
10:31
>> That why
10:33
some constantly that happens, right? I
10:34
mean like I told you her story, right?
10:35
She grew up in a small town with small
10:38
village with a population of 400. Her
10:40
parents didn't know how to read, right?
10:42
>> Uh couldn't even sign their names,
10:44
right? And from there
10:46
um you know one of the first ones to go
10:48
to med first female to go to college and
10:52
med school in the 50s then go abroad to
10:55
UK to get trained like that was like and
10:58
then as I told you last time we spoke
11:01
the amount of women who come to me and
11:03
tell me like who are like now 50s and
11:05
60s that
11:06
>> the reason my parents did marry me off
11:09
at age 17 18 is because of your mom
11:12
because she created this role model for
11:13
the whole community that you know you
11:17
can go and achieve uh things. Um and uh
11:22
it it has been and
11:24
>> what does that feel like when someone
11:26
says that to you?
11:29
>> Um very proud.
11:32
>> Um I would say that right you get that
11:35
feeling like you know that you were
11:38
privileged to
11:40
born to somebody who kind of created
11:42
that kind of impact in the world and
11:43
then you start thinking like am I going
11:45
to be able to do the same thing or not?
11:47
Right? So it kind of like my my son said
11:50
that to my my my mom won the my got the
11:53
uh gold award which is like every four
11:56
years given to one or two uh women in
11:59
OBGYn.
12:00
>> I think they're only like
12:02
>> eight or 10 12 given since the beginning
12:06
of FIGO and she's one of the one who
12:07
gotten it and when she received it um my
12:11
my younger son goes grandma you're
12:13
raising the bar too high for us. I have
12:15
to win a Nobel Prize to achieve that
12:18
kind of impact. So,
12:21
>> yeah. Yeah. Or you can just be like, you
12:24
know what, I'll have more of that. And
12:26
to know that you're living uh a, you
12:30
know,
12:32
basically someone who's like a
12:35
superhero is living in your under your
12:38
roof. You know, we we we we talk a lot a
12:40
lot a lot about superheroes, but um
12:42
oftentimes the superheroes
12:44
um you know, if you're if you're born
12:46
into it and you have you're very
12:47
fortunate to have a great mom and dad,
12:49
those are the superheroes. It's kind of
12:52
cool.
12:53
Uh I want to shift gears for a minute
12:55
because you know in our one conversation
12:57
we had which is super memorable and I
12:59
still ponder about it is
13:01
at some point your mission became to
13:05
impact a billion people
13:08
like that's not a casual number. Um can
13:11
you make can you do you mind sharing
13:13
that story uh about that?
13:16
>> Sure. So I built my uh first enterprise
13:21
software in high school uh for exam
13:23
management system still in Pakistan. It
13:25
was like DOSS based systems. This is 80s
13:27
right?
13:28
>> Yeah.
13:28
>> Um I think I was 16s at that time 15 16
13:32
and that thing got acquired
13:36
um and as a young you know teenager in
13:38
Pakistan getting a any kind of check for
13:42
anything was a very game life-changing
13:46
thing. So one of the things and I was a
13:48
boy scout in Pakistan and was involved
13:51
in a lot of the exchange activities that
13:53
were happening. So I decided to come to
13:55
the US as an exchange scout and just
13:58
travel on my own as a 16-year-old.
14:01
Um that was another we could have a
14:04
whole episode talking about adventures
14:06
of 16year-old Pakistani traveling alone
14:08
in the US.
14:09
>> Sounds like sounds like a book by the
14:10
way.
14:11
>> Yep. So, so this is 1988, you know, and
14:15
I landed in Boston and pretty much, you
14:19
know, uh, use the cheapest possible
14:22
routes, ways to move around to New York,
14:25
to Philadelphia, all the way down to the
14:27
east coast. My plan was the whole go
14:28
down south, back up the West Coast, and
14:31
then go back type of thing. And I would
14:33
either stay with some scout family or
14:36
somebody uh along the way. and then
14:40
landed in Miami and um uh was it a
14:44
staying at a YMCA and the whole
14:46
discussion with other kids started
14:48
happening what should I be doing and you
14:50
know idea of going to Key West to see
14:52
the southernmost point of US came up
14:54
>> and it's like right which is the
14:55
Greyhound bus that's going to take me
14:56
there or the train and they were like
14:59
you can do this they take a do adventure
15:00
take a boat
15:03
so I rented a 35T sloop and a captain I
15:07
mean obviously for you know the captain
15:09
was four year old. So like again also
15:12
not that mature adult but for a 16 year
15:15
old that was you know somebody who's 8
15:17
years older than you um you know sounded
15:20
like he's an adult. Um so you know we
15:24
sailed down to Key West and going
15:26
through um Gulf of Mexico at the time on
15:29
a moonless night.
15:31
>> First time ever saw the Milky Way.
15:34
>> Never seen it this clear. Right.
15:36
obviously you're so far away from
15:37
civilization too.
15:39
>> Um and then the captain goes to me like
15:42
what does this make you feel? And I go
15:45
like, you know, I thought about it and
15:48
I think my answer was insignificant,
15:52
>> right? The scale of you know billions
15:54
and billions of stars and galaxies and
15:56
it's like you're not even a single peck
15:58
of sand in the entire
16:00
>> you know beach around the world. Like
16:02
that's how small you are. And then he
16:04
goes, "What are you going to do about
16:06
it?" And that just somehow
16:08
figured something in my brain.
16:11
>> And that that led to like, "All right."
16:14
And obviously, you know, I had role
16:15
model at home. Probably that had
16:17
something to do with it. But
16:18
>> right,
16:19
>> we I woke up next morning and said like,
16:21
I got to go back and I got to go figure
16:23
out how I want to create an impact. And
16:26
at that time I made, you know, at that
16:27
time the world population was 4 billion.
16:29
I thought like I'll impact a billion
16:30
populated people before I die. That
16:33
became my goal as a 16 17 year old. I
16:35
didn't know how to do it, what to do.
16:37
Yeah, I was too immature, right? As a
16:39
16, 17 year old, you don't know
16:40
anything.
16:41
>> Well, but when you said that out loud
16:43
and someone heard you, like who who
16:45
thought you were crazy?
16:48
>> Um,
16:49
you know that that's the funny thing.
16:52
I mean, even this was 24 year old
16:56
um young man, he didn't say that,
17:00
right? He his was his his was more
17:02
acknowledgement sad acknowledgement.
17:04
Yeah, that probably makes sense, right?
17:07
And um I think he's also probably was
17:09
naive not understanding what I'm saying.
17:12
>> Yeah, like it doesn't compute humans are
17:15
hard very hard time to understand large
17:17
numbers. You know, we think them
17:20
linearly but they're not. They're
17:21
exponential uh kind of activity that
17:23
used to happen to get them to a billion.
17:25
You you can't do linearly.
17:27
I have to double double double double to
17:30
get to that number and that's
17:32
exponential growth and humans are very
17:34
bad at rec you know thinking about uh
17:36
things exponentially. So
17:38
>> so then what does that force you to do
17:40
differently?
17:41
>> So I think the only thing I could think
17:42
of to create a social impact was
17:44
becoming a physician. I still didn't
17:45
know how I'll impact a billion people
17:48
but I knew that doing something in
17:49
healthcare would would be easier to
17:53
understand how I'm impacting the
17:55
population. Right. So even though
17:56
everybody thought in my family that I'm
17:58
going to go into do some kind of
17:59
engineering, computer science or
18:01
something like that because you know I
18:03
learned how to code in sixth grade and
18:04
that's what I've been doing.
18:06
Um but that shift in mindset and
18:10
everybody is surprised that I went to
18:11
med school.
18:14
>> Yeah. you mentioned uh Microsoft and
18:18
balance, but I just want to touch on
18:19
that because you've had multiple
18:22
tours of duty, I'll call them, you know,
18:24
in your career, diff different chapters,
18:26
uh different problems.
18:29
>> When you when you look back, how how do
18:33
you decide when it's time to stay versus
18:38
when it's time to move on?
18:43
I wish there was easy answer,
18:47
right? Um I I'll be honest with you
18:51
it was less about
18:54
in hindsight looks like amazing decision
18:57
in the moment it looks like you're
18:59
insane right
19:02
>> type of thing. So long time ago
19:06
>> yeah long time ago mentor
19:08
>> let me give you some context here right
19:09
long time ago mentor told me
19:12
>> if you don't do anything the answer is
19:15
no right default is it's not going to
19:17
happen
19:18
>> right
19:19
>> by even doing one effort one time try
19:22
you increasing the odd of thing getting
19:24
done I procrastinate a lot right so yeah
19:27
>> that was kind of his guidance so I've
19:29
kind of taken that like if you're not
19:30
calling somebody
19:32
>> to ask a question the answer is no.
19:34
Right? So if you want something
19:36
>> the fact that you made that call you
19:39
increase the likelihood like costing you
19:41
a coin 50% right chance that somebody
19:43
may say yes. So when I was in my
19:47
residency
19:48
you know um and again my wife has always
19:52
given me this critical feedback uh on
19:56
timely kind of nuggets of wisdom that
19:59
have um changed the direction you know
20:03
career. So I I was very into MR doing a
20:05
lot of MR research doing pediatric and
20:08
fetal cardiac MR you know ended up doing
20:10
the first world's first fetal cardiac MR
20:12
in 2000 before anybody else did and so
20:15
one day you know she goes like there
20:17
probably seven people in the world who
20:18
care about your cardiac MR research
20:21
but you go to this IT committee meeting
20:23
at Guisinger health system one day a
20:25
week and the entire health system sings
20:27
your songs. You walk in a hallway and
20:30
everybody knows who you are. Why are you
20:32
resisting this informatics
20:34
route? Like I don't know what the career
20:36
over there is like. Like like I don't
20:39
even know like like I don't have any
20:40
training. Yeah, I learned how to code
20:42
but I didn't go through engineering
20:43
school. I didn't go to anything. So this
20:45
is just my general knowledge and I don't
20:47
know what what kind of career that'll
20:49
be. And she goes like that hasn't
20:51
stopped you from figuring it out. Right?
20:54
So, so that led to me
20:58
thinking there needs to be a training
21:00
program that combines IM, you know,
21:02
imaging and informatics and nothing
21:04
existed at that time. This we're talking
21:05
about 2000 2001, right?
21:08
>> Yeah.
21:08
>> And led to me calling every single
21:11
chairman and and anybody had anything to
21:14
do with informatics to ask if there's
21:16
any training program I can go into. The
21:18
answer was no. Great idea, but nothing
21:21
exists. Somebody should start one.
21:22
Unfortunately, we don't have funding or
21:24
the people to do it. Called everybody
21:27
from MGH, Bighgam, Mayo, Cleveland
21:29
Clinic, Stanford, you know, UCSF, you
21:32
name it.
21:33
>> Everybody said no. Finally doing that
21:36
for like 2 3 weeks, I finally, you know,
21:39
maybe I talked to
21:41
50 55 people.
21:43
Um, and they all said no. Um 3 weeks
21:48
later, University of Maryland chair uh
21:52
vice chair Elit Seagull Pages calls me
21:55
back. He's like, you know, the new
21:57
chairman's coming on board. His name is
22:00
Ruben Message.
22:02
Um and him and Dean and I were having
22:05
dinner. I brought up this your call
22:08
again. Elliot didn't know who I was. It
22:10
was random call. He didn't know what I
22:11
looked like, anything, right? It was a
22:13
random call. There's another whole story
22:14
on how that happened.
22:16
and he goes like we would love to do
22:18
this but we want you to come and start
22:20
the program. So I was still a resident,
22:23
right? Um so they wanted to not just
22:25
train but come and start the informatics
22:27
program and you know when it starves the
22:30
line so Elliot's undergrad was in
22:32
computer science ended up going to med
22:34
school at Maryland. Ruben Mezri was an
22:36
engineer at RCA on working on LCD
22:38
displays and at age 37 decided to go to
22:41
med school ended up doing uh landing at
22:44
Brigham as a fellow when the first MRI
22:48
was deployed there.
22:50
>> Right. Well,
22:52
>> right. So, uh, so we have a two
22:55
engineering background becoming a
22:56
radiologist recognizing what is needed
23:00
and decided to go bet on this one kid
23:03
from Pakistan to make it happen,
23:06
>> right? So, so that kind of led to
23:09
me coming there and then obviously you
23:11
know like once somebody gives you that
23:13
kind of a chance you would be an idiot
23:15
to waste it
23:18
>> and I just put thousand% in it to make
23:21
it happen and uh look where we are.
23:25
>> You mean you also generated that
23:27
opportunity. It's not like you sat back
23:29
and it just landed on your lap. Like you
23:32
went after it,
23:34
>> you know? That's uh that's really
23:37
interesting.
23:39
I want to talk about Higgy for a moment
23:40
because
23:42
>> you work there was such a massive scale,
23:45
you know, real people, data, real
23:48
consequences. I'm just sort of curious
23:50
what what did that scale teach you that
23:54
you know no white paper course could
23:58
ever teach you?
24:01
>> Yeah. So
24:05
let me tell you why Higgy right I was a
24:08
radiologist practiced
24:10
>> you know at Hopkins and Maryland before
24:12
I did all this stuff. It's all about my
24:15
inner why, right? My inner why has
24:17
become I want to solve problems the
24:19
people I love have and create a social
24:20
impact at the same time.
24:22
>> Both my parents had hypertension
24:25
and um I knew I'm going to get it and my
24:27
kids going to get it. So I've always had
24:29
that's why I focus on cardiac MR as my
24:32
uh work is even in radiology. Um
24:36
so I always was
24:39
trying to figure out how to how I'm
24:41
going to solve this problem. my billion
24:42
impact kind of question came in broad
24:44
and I realized that all just
24:46
>> I don't know how unless I'm inventing
24:47
something big and doing something it's
24:49
not going to be creating a huge impact
24:52
and my work in card how slow it was
24:55
moving was kind of also showing me that
24:57
it needs to go faster than this
24:59
so when I was at Microsoft this idea of
25:02
consumer health kind of started started
25:04
company this product called health vault
25:06
that was trying to do this stuff
25:08
>> um I had a chance to work with the Xbox
25:10
team so saw saw in real time what scale
25:14
meant, right? Xbox at that time, this is
25:16
we're talking about 2010
25:19
>> uh when connect Xbox connect came out,
25:21
we had 350 million users, 55 million
25:25
daily active users.
25:27
That scale doesn't exist in healthcare
25:29
even today, right?
25:30
>> That's great.
25:30
>> And you're deploying software in 110
25:32
countries in 33 languages every two
25:34
weeks.
25:36
>> Like that's kind of gives you the
25:38
perspective of what real scale is. And
25:40
then you realized that healthcare is
25:42
that's not a scale, right? That's like
25:44
one hour worth of activity on
25:46
Microsoft's platform, right? So
25:49
kind of started and that kind of gave me
25:51
this idea that then when you're building
25:54
software, if you're not thinking scale
25:56
up front, you're not going to be able to
25:58
scale it harder. Technical debt will be
26:00
too high for you to do the stuff. And
26:01
that's where a lot of the fail failures
26:03
happen once you know rubber hits the
26:05
road. Even if you have a great product
26:07
because of your lack of ability to scale
26:09
it fast, uh you lose your reputation,
26:12
you lose your credibility and then
26:14
things fizz. There's so many amazing
26:15
ideas come up and never
26:18
become successful is one of those
26:19
reasons. So one thing I learned up front
26:21
was, you know, number one is obsessed
26:23
with the problem, right? What problem
26:25
you're trying to solve. Number two is
26:28
um align with your inner why.
26:31
Number three is the right people. People
26:34
process and uh make make or break the
26:37
whole thing. Number four is user uh
26:40
obsession with user of the user
26:42
experience.
26:43
>> Number fifth is thinking of building a
26:45
scale from day one.
26:47
>> Um and the last one is you have to have
26:49
say thesis on how you're going to make
26:51
money. But anyway, so so we could have,
26:55
you know, took us about 6 months to get
26:58
the platform ready for Higgy when we
27:00
started.
27:01
>> Yeah.
27:01
>> And I would always get this question
27:03
like, we're only doing two kiosks right
27:05
now as a pilot. Why are you spending so
27:07
much time building all this
27:08
infrastructure? Like if I don't do this
27:11
base like right now,
27:13
we're going to be crippled. It's not
27:15
going to work. Right? So I would always
27:18
get the board will always you investors
27:21
would always get upset like you're
27:23
delaying all the releases and all this
27:24
stuff blah blah blah but
27:25
>> but but that was because you saw what
27:27
was going on at Microsoft so you
27:28
obviously gained that and that brought
27:30
that into Higgy. So you kind of knew
27:32
like hey this is not a nice to have.
27:34
We're not just doing this for
27:35
>> this is an absolute mustave. This is a
27:38
critical choke point if we if we don't
27:40
>> one of the things that one of the things
27:41
we talk about is patient engagement
27:43
right so it is really hard to create new
27:47
way to engage because the behavior
27:49
change
27:49
>> if you can tie to some activity that's
27:51
already happening then you taking
27:54
advantage of existing behaviors so when
27:57
one of my interns who came up with the
27:58
idea of the kiosk right it's funny thing
28:00
like we weren't we trying to build a
28:01
watch at Iggy we weren't trying to build
28:05
a a kiosk and we were just couldn't get
28:06
the price down because our focus was
28:08
elderly and low income and it just
28:09
couldn't get the price down. And one of
28:11
this intern on one of the brainstorming
28:13
session goes like, "What about blood
28:14
pressure kiosk?" I was like, "What do
28:15
you mean?" He goes, "Those things in the
28:16
grocery stores and pharmacy like they're
28:18
always broken and never work." He goes,
28:19
"Exactly."
28:23
Right. So, and we walked down, literally
28:25
the whole team got up and from the
28:26
conference room in Chicago walked down
28:28
to the C CVS. I still remember it's on
28:30
um
28:32
uh Dearbornne and that corner. Kenzie's
28:35
Kinsey and Dearborn right next to the
28:37
merchandise mart.
28:38
>> We went in there and we spent the whole
28:39
day in that CVS. We couldn't believe
28:41
there were people using these broken
28:42
kiosks
28:44
>> and talked to the pharmacist. He was
28:45
like, "Yeah, there lines of people
28:46
standing there." And there was one time
28:47
there like 10 people waiting to use that
28:49
blood pressure machine. I was like,
28:50
"This habit already exists. We just need
28:52
to replace with a better machine,
28:54
>> right?" And then we looked at the
28:56
numbers and it was insane amount of
28:58
usage on these things, right? So, so I
29:00
knew that the usage is there. you just
29:02
need to replace it with a better
29:03
machine, more connected and all this
29:04
stuff and you could take advantage of
29:06
existing habit that already exists and
29:07
it was a demographic we were targeting
29:09
was all elderly people who were using it
29:10
and low income were using it.
29:13
So I think building the foundation up
29:15
front I mean the first kiosk were
29:17
deployed in CVS in downtown Chicago on
29:19
July 6th 2012. Guess what?
29:22
>> That's obviously etched in your mind
29:23
that date.
29:24
>> Yeah. Because we all were down there
29:26
doing it ourselves, right?
29:27
>> Guess what? All those kiosks still exist
29:30
in those 17 locations. How many years
29:33
later?
29:34
>> Wow.
29:35
>> They've g they've gone through three
29:36
operating system updates, right? Like
29:39
Windows from 2012 don't exist anymore,
29:42
right? We've done three oper system
29:43
updates every two weeks software updates
29:46
and they're still there.
29:49
That was the I mean our goal was five
29:51
year life of these kiosks.
29:54
No, none of us expected to run 15 years.
29:58
>> That's pretty phenomenal,
30:00
>> right? And I'll stop next time I'm down
30:02
there.
30:03
>> Yeah. I mean, we deployed 12,000 of
30:05
those, you know, in North America, both
30:07
Canada and the US. Uh and was
30:09
unbelievable. ended up doing I think
30:11
when I left the we had crossed 700
30:14
million screenings.
30:18
>> That's that's wild.
30:21
So now thinking about this type of scale
30:24
um because you've seen healthc care at
30:27
the edges at the center what what what
30:29
do you see what's the common thing that
30:33
smart leaders um
30:36
still misunderstand about about data.
30:44
I think people get stuck in the
30:47
buzzwords,
30:51
right?
30:54
I mean like you know
30:57
when we started higgy but when we got to
31:00
the scale where we got to the scale it
31:02
was very obvious if we had a data breach
31:05
there's no way to recover from it
31:08
>> right the because we are holding PHI
31:10
right the it first of all it will be a
31:13
DOJ investigation type of thing right
31:15
it'll be a DOJ you know $100 per events
31:19
when you have 70 million patients VHI
31:22
you there's no way to recover from that.
31:25
>> Yeah.
31:26
>> Right. You're you're going to jail and
31:27
everybody's going home basically. That's
31:29
the answer to that. And then and you
31:32
need to make sure you're thinking how
31:35
you're doing this the right way. You you
31:38
know you can't take that lightly.
31:41
So
31:43
um
31:45
and there are ways to you know like we
31:49
solved this problem in 2012.
31:52
like how do you secure data? I still
31:54
like I literally last by week I got an
31:56
email some data hack happened for some
31:59
stupid thing you know it's like in this
32:03
day and age right like we never took
32:07
Higgy down for maintenance
32:10
that wasn't a thing like when was the
32:13
last time Gmail went down for
32:14
maintenance
32:16
>> or Facebook for maintenance
32:19
>> never right like I still don't
32:21
understand when EHR has to go down for
32:24
maintenance.
32:26
I I I just don't get it right. These
32:28
problem are solved problems. They were
32:29
sort of solved problems 15 20 years ago.
32:32
It's not that people these are and
32:34
they're not secret IP proprietary stuff.
32:37
This is common knowledge how to secure
32:38
this stuff, right? It's just it's all
32:41
about the right people, right processes,
32:43
right problems to focus on, you know,
32:44
but what people don't do is
32:47
>> go down to first principle thinking,
32:51
right? What is first principle thinking?
32:52
what are the first problems and how do
32:54
you handle them at the atomic level of a
32:56
particular problem these are the
32:58
learnings I learned at Microsoft right
33:00
uh is how do you actually go to like I
33:02
would have you asked the five W's of why
33:05
we're doing it um and um I think that's
33:09
what ends up happening where
33:12
with data especially
33:15
>> right I personally believe uh patients
33:18
have the right to control what happens
33:20
to the data right If they are making you
33:23
the custodian of the data then it's your
33:24
responsibility to ensure uh you know
33:27
something nothing bad happens like he
33:29
was the first one that actually created
33:30
receipts when you share data with
33:32
somebody so he had an API and you know
33:35
other people could build apps to pull
33:36
the data around the stuff but we would
33:38
actually send a patient an email saying
33:40
that this entity requested this
33:43
information and here was the reason they
33:46
said they wanted the data right like you
33:48
probably acknowledge agreed you know we
33:49
look at the terms and we say agreed
33:52
and then we disappear but then what some
33:55
breach happens and how do you know
33:58
that was something you agreed to or not.
34:01
So we actually would cra capture that
34:03
like why this data was shared what was
34:05
the use with this other third party app
34:08
was asking for and made sure patient had
34:10
a record of it when it happened what it
34:12
was what the timeline was and all of
34:14
that stuff so that if they ever want to
34:15
go back after that thing they had a
34:16
receipt for it. So they basically were
34:19
like, you know, they click the terms and
34:20
condition and that's a way of them
34:21
saying, you know, consent. Yeah.
34:24
>> Consent into this,
34:25
>> but they but there's no receipt for it.
34:28
>> You don't get an email confirmation like
34:30
you shared this data and here was the
34:32
reason you shared this data and here's
34:34
what the entity said they will do with
34:35
the data and here's a link on the
34:37
privacy policy. Here's a link of the
34:38
stuff that now you have, you know,
34:41
record of.
34:42
>> Yeah.
34:42
>> And we were doing this in what, 2013,
34:45
2014? Like 12 years ago. Come on.
34:49
>> People still I'm still mind boggled that
34:52
these simple things that you could do
34:53
people still not doing.
34:55
>> Yes. I mean, but everybody wants their
34:57
systems to talk. I mean, what are they
34:59
what are they not willing to give up to
35:01
make that happen?
35:05
>> No, I mean like there nothing to give
35:08
up.
35:10
>> Yeah.
35:11
>> You need to make sure you know you've
35:13
thought through all the problems, you've
35:15
addressed the you know need. I mean at
35:16
the end of the day
35:18
um like I'll give you an example of
35:20
things we did at a gate again like this
35:23
was this is not secret stuff right this
35:25
is wellnown thing these are best
35:26
practices
35:29
only four people on the team had access
35:31
to production environment that is it
35:36
was me my the CTO my VP of engineering
35:39
my head of platform and one of the that
35:42
is it rest of the team never knew
35:46
touched the production environment. They
35:48
always work in and staging environment
35:50
only for people. So if a breach
35:52
happened,
35:53
>> guess what? Who got compromised? They're
35:54
only four people,
35:58
right? And if anything happened to all
36:00
four of us, there was access in the bank
36:03
in evolved where our CFO
36:06
or the attorneys could go and get access
36:10
to it and then could give it to somebody
36:11
else, right? is a physical way how you
36:14
can actually go access u something that
36:17
right and uh and that's it right so we
36:21
had operational controls we have
36:23
contractual controls right where you
36:26
know we did a lot of stuff uh and we
36:28
have technical controls around this
36:29
aspect of it audit logs and all that
36:31
needs to happen and um everything that
36:34
and we decided like okay fine
36:37
can we store data that is not human
36:39
readable
36:41
can we store in a format that is not
36:43
human readable right so we so stored
36:45
everything in yeah I mean you can easily
36:47
convert it all the stuff but like it's
36:48
just extra step to add to it right so
36:51
you can save everything in binary JSON
36:54
then even if somebody gets access to you
36:56
encrypt the data and do all the stuff
36:58
you just make layers and layers and
37:00
layers of complexity open I I remember
37:02
the the best example right so one of my
37:04
engineers came up with this idea that if
37:06
you you know on a kiosk remember we're
37:08
talking about kiosk 2012 iPad has not
37:10
come out yet right Right?
37:12
>> Screens aren't that great. So the more
37:14
characters you have in a password, the
37:16
horrible the experience was. Right? So
37:19
we found the six characters were the
37:20
optimal where people were happy with it.
37:22
Eight characters or any kind of weird
37:24
characters, they would just
37:27
not go by it. And I said like you know
37:29
like the N standard is eight character
37:31
passwords.
37:33
How do we do this the right way? And so
37:36
what's the first principle? Why eight
37:38
characters? Right? So you go to first of
37:39
thinking,
37:40
>> right? Then you realize it's about the
37:41
entropy like how long does it take to
37:43
brute force 8 character password right
37:46
because like what are the other ways you
37:48
can increase entropy of a password
37:50
decryption other than this
37:53
and we realize that it's all about time
37:57
how long does it take right so eight
37:59
character is leading to a time like is
38:03
there any other way to induce time in
38:05
decryption
38:07
>> so we came up with this uh uh
38:10
architecture called brypt which induces
38:12
time for decryption.
38:13
>> Okay.
38:14
>> And we realize if you add 10
38:16
milliseconds
38:18
as a lag time to decryption and double
38:21
it every time you guess the password
38:23
wrong. So 10 then 20 then 40 then 80
38:25
>> 10 millconds for a human is very fast.
38:27
You don't even know happening right?
38:31
But for a automated brute force attack
38:33
thing, it's crippling
38:36
and after 10 10 tries, you just lock the
38:38
account down.
38:41
>> Oh, right. So,
38:44
so very simple thing. So, we were had
38:46
six character passwords with two
38:48
millennia entropy.
38:51
>> So, look to to will take you 2,000 years
38:55
for brute force figuring that place
38:57
thing out, right?
38:59
Yeah,
39:02
>> like you know like I'm just talking
39:03
about things that are when you think
39:05
about first principal thinking like why
39:06
are you doing certain things? People
39:07
don't ask like those things.
39:10
Um
39:11
>> isn't that amazing? You know we spend
39:12
all of our time doing training but when
39:16
it when it comes to basing things off of
39:19
a very clear set of principles which
39:21
stand the test of time. I think it's
39:24
even the same thing with like judgment
39:25
and decision- making. Like how do people
39:28
get good at making like good decisions?
39:30
Like I I mean I I never like took a
39:32
class on that. Like they just came
39:34
through a lot of trial and error and
39:35
like thinking through it like what do I
39:38
what do I actually use as a a matrix to
39:42
to to make a good sound decision.
39:45
>> Yeah.
39:45
>> You don't ask the why enough, right?
39:49
Why um you know why is like I remember
39:52
my my first experience I first think
39:55
thinking was I was on call as a resident
39:59
and there was a a patient who has a
40:02
young kid I think it was 10 yearear-old
40:04
12 year old who got a CT scan
40:07
um for abdominal pain and the appendix
40:10
appendix was 1 cm in diameter
40:15
and the but the problem was
40:17
The kid didn't have right abdominal
40:20
pain.
40:21
>> Yeah.
40:22
>> And attorney goes like, "Oh, there's
40:24
appendicitis." I like, "But the symptoms
40:26
don't match what we're saying." It's
40:29
like, "No, but the appendix is large."
40:31
It's like, "Why is appendix 1 cm? You're
40:33
calling it appendicitis." Like, well,
40:36
that's the cut off. It's like, but how
40:38
did that come up? Then you realize that
40:39
the cutoff came from a paper written
40:42
1960s in which they autopsied
40:45
100 uh people and did a normal
40:48
distribution of apprentices and did a
40:51
95% like student vision cut off as the
40:54
normal
40:56
>> appendic size and that has stayed
40:58
forever and we applying the same
40:59
principle to kids and adults and all the
41:02
same. It's like oh
41:04
>> crazy.
41:04
>> Yeah, we should probably not be doing
41:06
this anymore. But anyway, that was my
41:08
first real, you know, use of why and
41:11
going and investigating like where did
41:13
the appendix threshold came from to call
41:16
something above a centimeter.
41:19
But since then, a lot has changed. But,
41:21
you know,
41:23
all right. So, it seems to me that every
41:26
one of your companies that you've
41:28
started, there's been a sort of a
41:30
personal reason behind it. Let's talk
41:32
about Hopper. Let's talk about what the
41:35
focus is, what the why is and like share
41:39
a little bit more about that.
41:41
>> Yeah. So I think as what a lot of people
41:44
would realize on this call or on this
41:47
broadcast is
41:49
um
41:52
radiology is in a crisis right now right
41:55
the imaging volume is just going up and
41:58
not because we are doing in unnecessary
42:00
imaging right I don't think majority
42:03
doctors are intend to harm the patient
42:06
right there's the reality of the world
42:08
right our population is aging
42:11
more people are surviving longer, right?
42:13
And that that's where imaging is needed,
42:16
right? Imaging has become way more
42:18
objective,
42:20
>> is providing lot more data that we never
42:22
had,
42:23
>> right? Before um and um and now drugs
42:28
are coming out that require imaging to
42:30
understand doymmetry for those
42:32
particular diseases or even looking for
42:34
complications of those diseases.
42:36
uh pressures on ER turnaround times, you
42:39
know, like you can get patient much
42:41
faster out the ER if you do imaging. So,
42:44
so uh Andy Rosen Crrance, one of my
42:46
former residents in Maryland, just
42:47
published a paper in September of last
42:49
year showing that the CT volume in ED is
42:52
actually increasing at 7% annually.
42:56
Previously assumption was 5% growth on
42:58
imaging, but we are seeing 7% and
43:00
accelerating. Right? If you look at his
43:01
papers from September September 30th and
43:04
the radiology journal,
43:05
>> it's actually accelerating and it's uh
43:08
you know um but we're not training more
43:11
radiologists.
43:13
Um so in 2017
43:17
I started getting you know people
43:19
reaching out like hey you left client
43:21
practice at different things and I'm
43:23
burning out and uh
43:26
how do I do this? And obviously my
43:28
wife's a radiologist so I I I would see
43:29
her like she joined um one of the
43:32
academic practice here in Chicago. Um
43:34
there were 12 body imaging radiologists
43:36
in 2014 when she started.
43:39
>> When she left last year, she was the
43:41
only one remaining. They were not able
43:43
to hire anybody back.
43:44
>> Right.
43:45
>> So can you imagine this is a transplant
43:47
center, cancer center and you don't have
43:50
a fellowship brain body imaging
43:51
radiologist
43:54
>> like
43:55
and a residency program. You're going to
43:57
train residents with no fellowship
43:59
trained body imaging faculty training
44:00
them like and by body imaging I mean
44:03
chest and pelvis right like most of the
44:05
body
44:08
and when she left she was the last one
44:10
who was doing this at the academic
44:13
center right and then why did she leave
44:14
like same issues the volume just going
44:17
up
44:18
>> through the roof u she was coming home
44:20
every day completely tired fatigued
44:24
um have dinner and she would still stay
44:26
till midnight finishing up a day's work
44:31
and again if you're an academic therapy
44:33
care center you're not seeing so simple
44:34
stuff I mean these are cancer patients
44:36
with multiple scans or transplant with
44:39
multiple things anatomy is completely
44:40
deranged if you don't understand what
44:42
the surgeon did in the uh you know how
44:45
did they did anastmosis what they
44:46
removed what they not removed how you
44:48
tell what is going on because anatomy is
44:50
completely different now right
44:51
>> so the amount of effort you have to
44:53
spend time understanding what is
44:54
happening is very complex you know
44:56
people think oh I can replace all this
44:58
stuff like no like the things I did when
45:01
my wife does like there's a lot of
45:03
cognitive overload to really understand
45:04
what has happened to the patient because
45:06
>> the brain doesn't just shut off you're
45:08
like constantly
45:10
>> it's constantly working
45:11
>> yeah people think it's detection it's
45:13
not right what what we're doing is we're
45:15
generating a narrative for the clinician
45:18
who ordered this exam to tell them what
45:21
is happening so it's a communication
45:22
method right we're looking at what's in
45:24
the images converting into pros that
45:27
then another human who's going to make
45:29
the clinical decision. So like when I
45:30
take care report I would say depending
45:32
on who ordered it because I've learned
45:34
over time Dr. XYZ wants to know these
45:37
things Dr. ABC wants to know other
45:40
things and I would make sure I'm
45:42
addressing who has ordered it, whose
45:44
patient it is, what service are they
45:45
doing, what are they looking for, right?
45:48
>> Or even anticipating, right? Like the
45:49
anticipation come in, oh, he's going to
45:50
probably go to the oncology service. Oh,
45:52
you're probably going to go to the
45:53
surgery service and I make sure let me
45:55
make sure I cover these things because
45:57
they would ask me immediately
45:58
afterwards, right? So that's how you
46:00
kind of think about all the stuff,
46:02
right? And so my the turning point for
46:05
me was uh so me and my two younger boys
46:09
have this text internally we called the
46:11
boys chat and uh I work out of my
46:14
basement. Uh my elder one was sitting in
46:16
his room on the on the second floor and
46:19
my second floor and my younger one was
46:21
in the kitchen and we heard the garage
46:23
door open. The elder one test texts to
46:26
the boys chat how's mama's mood? Should
46:29
I come down to greet give her a hug or
46:31
stay in my room? and they rang up and go
46:33
stay in your room. And I was just like,
46:36
my kids should not be deciding if they
46:39
want to give their mom a hug based on
46:41
how our radiology work was. I was like,
46:43
I got to go fix this thing.
46:47
>> So that kind of became my like, all
46:49
right, I need to go figure out what's
46:50
going on. Like, you know, I started
46:53
working in machine learning, computer
46:54
vision, AI in 2003, 4. And you know this
46:57
is 2017 and in my mind all this should
47:01
have been solved by now and why why
47:03
hasn't uh so that led to me and Robert
47:06
Bakas who was my first employee at Higgy
47:08
and became my CTO eventually having a
47:11
discussion together on a napkin like if
47:14
we had to do this again like
47:17
radiology stuff like what would we how
47:18
do we think about like and I told him
47:21
about this memo I wrote in Microsoft in
47:22
2012 about building a platform to
47:25
accelerate AI development because I
47:27
didn't think we should be building
47:28
disease by disease AI models. We need to
47:30
build larger models that can do hundreds
47:32
and thousands and of things at the same
47:34
time. Uh but how do you do it at a
47:36
scale? How do you do it uh you know
47:38
different was the approach. So
47:40
describing him what I was thinking and
47:42
Robert had worked at Merge Healthcare
47:44
which is a pax company right so he had
47:46
some background in this aspect of it and
47:48
we started thinking about how we would
47:49
approach it. At the same time what
47:51
happened was a lot of the uh the
47:52
attention is all you need paper comes
47:54
out from Google and my former computer
47:57
vision machine learning team at
47:58
Microsoft who had all moved on right
48:00
some went to autonomous vehicles some
48:02
went to meta some went to deep they
48:05
start sending me hey Khan look at this
48:06
transformer architecture remember your
48:08
memo you wrote this would be a better
48:10
architecture for this not a CNN based uh
48:13
platform that you had described so that
48:15
led to me collecting all of them in an
48:17
Airbnb
48:18
and flushing this out right so I got
48:20
some former fellows from Maryland some
48:23
you know former engineers from Microsoft
48:25
some hegy people all in three days at
48:28
Airbnb like hey how what are you going
48:29
to do so the goal out came back that we
48:32
need to build a platform that we can
48:35
shorten the development time frame and
48:37
we need to provide all the components
48:39
that are needed one of the other
48:40
realization we had was that we knew when
48:43
we were training these models ourselves
48:45
that models don't generalize well unless
48:47
either you have mass amount of data like
48:49
almost like all data from all time or
48:53
you have ability to fine-tune it for
48:54
local needs, right?
48:56
>> And it was becoming very obvious that
48:59
when you train a model in a particular
49:00
data set, you go deploy another patient
49:03
demographic. Even practice patterns are
49:05
different, right? You train at Hopkins,
49:06
you will scan a patient differently. You
49:08
train in Stanford, you will do it
49:09
differently because you
49:11
>> and those models don't perform the same
49:12
way. So and it sounds really really hard
49:16
to build your own AI model like like how
49:19
do we make this simple that was became
49:20
the tool. So, and the tools you need are
49:23
number one is data, right? How do you
49:24
get the data? Like, well,
49:26
>> yeah, with Microsoft and with Higgy, I
49:28
figured that piece out, right? How do I
49:30
secure data? How do I do data? How do I
49:31
handle it?
49:32
>> Yeah,
49:32
>> we had now decades of experience how to
49:34
do this stuff. Plus, we build trust
49:36
across health systems where we'd already
49:38
deployed all these things that people
49:39
kind of trusted this team to be able to
49:41
do this.
49:42
>> Next was can we do large pre-trained
49:44
models, right? Usually when people are
49:45
building models, they're using they're
49:46
not training from scratch. they are
49:48
using some CNN or transformer
49:50
architecture that are there retraining
49:52
on top of it. So like how do you do
49:54
something that is built on uh trained
49:57
from scratch on medical imaging uh data
49:59
itself.
50:01
So now we call them foundation models
50:03
but that was the kind of idea at that
50:05
time and then the entire tooling that is
50:06
needed to train at scale deploy at scale
50:10
validate verify
50:12
uh scale right so if you do it at scale
50:14
let's say you're doing thousand models
50:16
for you know hundreds of different
50:18
companies the scale and cost you can
50:20
have is very different than one entity
50:23
going building five six seven solutions
50:26
around that you would not be able to hit
50:28
the scale to get the cost uh low enough
50:31
around that aspect of it. So that became
50:34
hopper. Um you know we started with um
50:37
focusing on developers initially
50:39
launched API and SDKs and uh foundation
50:42
models the summer of last year and then
50:44
at RSNA we did a low code version of it.
50:47
Now
50:48
>> somebody with a little bit of IT
50:49
background can without knowing any line
50:52
of code can build their own um
50:54
classification models and um rest is
50:58
history.
51:00
That's amazing because the, you know,
51:02
how you kind of laid out the story, how
51:04
you how you had these relationships
51:08
like everything basically compounded,
51:10
you know, and that that sort of
51:11
compounding is what allows you to really
51:14
get to this particular new point and
51:16
exponentially be able to help more
51:18
people.
51:19
When you think about
51:22
the billion people impact, is that is
51:25
that still true north for you always?
51:27
And do
51:29
do your teams know that? The ones that
51:31
you work with, do they know they all
51:33
know?
51:34
>> They everybody knows it. Anybody I've
51:37
helped knows this. Um because my ask
51:41
always is like uh if you're doing
51:43
something that going to impact a lot of
51:46
people, I'll help you out with no ask in
51:49
return
51:51
>> because my motivation is very different.
51:54
>> Yeah. for you significance and the
51:57
impact.
51:58
>> Whenever my new admin joins to support
52:00
me in every company, they always ask
52:02
like why do you have all these random
52:05
like people on LinkedIn reaching out and
52:06
you have take these calls? Should I say
52:09
no to them? Like no, no, no, no. You
52:12
need to have I need to have at least
52:13
three of these every week because you
52:16
never know.
52:17
>> Come up with something idea that maybe
52:19
Yeah. I mean,
52:21
>> I think I told you the story, right? the
52:23
number of people
52:25
I meet now who say how I changed their
52:28
life and I don't even remember talking
52:29
to them.
52:31
>> It's crazy,
52:32
>> right?
52:32
>> That's how you know you're having
52:34
immense impact. I mean, you obviously
52:36
>> to that person it's it's a it's such a
52:39
monumental a monumental moment
52:42
>> and you're like trying to database in
52:43
your head like I don't even remember
52:45
having that conversation but apparently
52:47
what I'm doing is working. So
52:49
>> yeah, random LinkedIn reach out
52:51
>> had a 30-minute call and somebody
52:54
changed their
52:56
career based on that feedback
52:59
and created huge impact like yeah that's
53:02
my goal
53:05
even though I don't remember the
53:06
discussion but the outcome happened.
53:08
That's what I care about.
53:11
>> All right let's we're getting closer to
53:13
the end here. So let's I want to do a
53:14
little quick fire to make do do a little
53:16
fun. These are some things I've been
53:17
thinking about. So,
53:19
>> um, one habit that your mom,
53:23
one habit from your mom that still shows
53:26
up in your leadership,
53:30
>> I think it's the word caring.
53:31
>> Karen. All right. What about
53:34
um, slide decks or whiteboards?
53:38
>> Whiteboards every day?
53:42
>> All right. What about uh meetings before
53:45
8:00 a.m. necessary or avoidable?
53:50
>> I'm a late guy.
53:51
>> It's, you know, my brain works after
53:54
midnight.
53:55
>> So, I think it's avoidable.
53:57
>> Okay.
53:57
>> For me, it's avoidable.
53:58
>> All right. What about um one because you
54:02
mentioned buzzwords earlier. What about
54:03
one one healthcare buzz word that you
54:05
quietly dislike?
54:09
>> What's the word I dislike?
54:12
I think it's HIPPA certification, but
54:14
there's no such thing.
54:19
>> All right. And then just last question.
54:21
I mean, if we're, let's just say we're
54:24
10 years from now,
54:26
um, if somebody looks back and studies
54:28
your career,
54:31
what do you hope that they don't miss?
54:38
Oh, that's a tough one.
54:47
I think what they don't miss
54:54
is how do you do that that all the
54:56
things you know you do you can do it
54:59
while still being civil
55:01
and still being humble.
55:05
>> Yeah.
55:05
>> And still caring for people. Right.
55:08
>> Yeah. The yeah the archetype of you know
55:11
entrepreneur is that
55:13
>> make sorry I'm using bad
55:16
>> it's your do whatever you want
55:18
>> right like you have to be you know
55:23
>> ruthless and things like that right no
55:26
you don't have to be right
55:28
>> uh it doesn't mean you know you can't be
55:31
passionate aggressive getting things
55:33
done but you can do it while being
55:34
humble
55:36
>> totally like
55:38
what is the best way for people to
55:41
follow you, your work? Maybe just kind
55:43
of give a little commercial about, you
55:45
know, Hopper where they can follow. I
55:46
mean, because you're doing you're doing
55:47
a number of different things. So, I want
55:49
to give you the floor and then I'll
55:50
we'll wrap up.
55:51
>> Hopper is easy. Hopper.ai
55:54
website. Um I'm easy to get hold of
55:57
through on LinkedIn. I respond to
55:59
messages. Um just find me on LinkedIn.
56:02
Um, you know, we have um plenty of way
56:06
to kind of contact us on Hopper's
56:07
website through Hopper's social media
56:09
channels. We're pretty active in that
56:11
aspect of it. Just tag us.
56:12
>> Yeah.
56:13
>> Uh, if you're trying to message us or
56:15
things like that, more than happy to do
56:16
it. If you have ideas to explore on
56:19
Operas platform more than we launched
56:21
our catalyst program where we're looking
56:22
for academic researchers with crazy
56:24
ideas, be providing comput resources,
56:26
data, and uh uh and expertise to help
56:30
achieve those. Um and then practices
56:33
that want to deploy their own AI models
56:35
but don't have people we have forward
56:37
deployed services uh that we bring on
56:39
board and help them do that aspect of it
56:41
too right in my mind every practice has
56:43
unique problems right so you can't be
56:46
waiting for a vendor to come up with a
56:48
product like here's a platform that help
56:50
you do that
56:51
>> so what do you want to do and uh if you
56:54
don't have the teams we'll help you
56:55
build it
56:58
>> you know that there's one overarching
57:00
theme throughout this conversation that
57:02
I'm hitting on. It's like your sense of
57:04
curiosity and adventure like you know
57:07
and it's because
57:09
you have taken those adventures that
57:12
those have led and created more
57:13
opportunities and
57:17
yeah you do care. I think uh like I
57:19
mentioned before, sometimes, you know,
57:20
we go through life and we uh look for
57:24
superheroes, but we don't actually
57:25
realize that there's superheroes living
57:29
under the roof that we're in growing up.
57:31
And you recognizing that early on and
57:33
then going out and doing something and
57:35
then having this big audacious goal of
57:38
look, your work is about significance
57:41
and impact. And those that are aligned
57:44
around significance and impact will
57:46
absolutely run through walls with you
57:48
because they know what true north is.
57:51
And so for those that are listening, if
57:54
you're listening to this, what my take
57:57
is is number one is um good guys do
58:00
finish first. You could be kind, still
58:03
have really high standards, and still
58:05
get a lot of stuff done and still make a
58:07
lot of progress and make a lot of
58:09
impact. So that that's like a really
58:12
cool thing to leave with. And then the
58:13
other thing is just
58:16
your ability to remain curious. No
58:18
matter how much uh you know worldly
58:21
success, you know, others might perceive
58:24
you to have that you've you've obtained,
58:26
there's still a level of curiosity that
58:28
you have. And like you said, be humble
58:30
um and keep moving ahead. You can always
58:32
get better get better.
58:35
So for those tuning in, if uh there is a
58:38
topic that you want us to explore or
58:41
even a a leader that um you want to hear
58:44
from, you know, drop me a note, shoot us
58:45
a comment, let us know what you think of
58:47
today's conversation. You know, I mean,
58:49
would you just mind sharing it if you
58:50
found it valuable? Yeah, share it. Let
58:52
us know. Um so this platform for me,
58:54
it's it's about community. It's not just
58:57
about content. So thank you for
58:58
listening. Keep building with purpose
59:01
and we will see you next time on the
59:03
Advancing Health Innovation Show. Good
59:05
to see you. Pleasure, Michael.