June 30, 2026 • 44 mins
When Dr. Lara Jehi began treating epilepsy patients in the 2000s, critical surgical decisions were driven more by clinician intuition and expertise than data. Today, she is a leader of IBM and Cleveland Clinic’s Discovery Accelerator, using advanced AI and quantum computing to transform how researchers analyze data, simulate molecules, accelerate drug discovery, and develop more precise treatments. Malcolm Gladwell talks with Dr. Jehi about how quantum computing is changing biomedical research, and what these breakthroughs could mean for the future of healthcare and life sciences.
This is a paid advertisement from IBM. The conversations on this podcast don’t necessarily represent IBM's positions, strategies or opinions.
Visit us at https://www.ibm.com/think/podcasts/smart-talks
See omnystudio.com/listener for privacy information.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
I'm Malcolm Glabwell and you're listening to Smart Talks with IBM.
When doctor Laura Jahi began treating epilepsy patients in the
early two thousands, she noticed something unsettling. Different surgeons could
look at the exact same case and recommend completely different treatments.
One surgeon might remove one part of the brain, another
(00:24):
a different part, and a third might not operate at all.
Doctor j Hi believed there had to be a better way,
one grounded in data. That conviction set her on a
path that would lead her to become Chief Research Information
Officer at Cleveland Clinic and the executive program lead for
the Discovery Accelerator. The Discovery Accelerator is a ten year
(00:47):
partnership between Cleveland Clinic and IBM where researchers are using
AI and quantum computing to make incredible discoveries in healthcare
and life sciences. I sat down with doctor j High
to explore what's happening now, what's possible with quantum computing,
and where this next era of biomedical discovery is headed.
(01:10):
Epilepsy was your kind of specialty within neurology?
Speaker 2 (01:13):
Correct? Yes?
Speaker 1 (01:13):
What led you to that?
Speaker 3 (01:15):
I was always fascinated by the brain. You know, it's
the part of our body that leaves still to this day,
the most to be discovered. So I was always intrigued
by areas that leave more for discovery, and epilepsy was
my pragmatic side, wanting to choose a subspecialty where the
(01:38):
problem can be fixed. In epilepsy, there are many medications
that are very effective, and there's a brain surgery that
we can do to stop seizures when medicines don't work.
Speaker 2 (01:50):
That attracted me, you.
Speaker 3 (01:52):
Know, compared to other areas in neurology, like stroke, for example,
or dementia, where usually the damage is more. I wanted
to be able to tell my patients that you have
a big problem, but here's what I can do to
fix it, and epilepsy offered me that.
Speaker 1 (02:10):
But what would there must have been interesting and intriguing
unsolved problems in neurology.
Speaker 3 (02:19):
Oh my gosh, it's the whole brain, isn't it. Before
starting to deal with artificial intelligence and research. Right in
the neurology, I'm dealing with real intelligence, the human brain
and how it works and how we think and how
we make decisions, and how it can grow and evolve,
(02:40):
and so there is many untaped questions in neurology, and
that's part of the fascination in it. In epilepsy in particular,
it's an electrical disease in the brain, it's actually one
of the conditions in neurology. That's a perfect alignment of
(03:05):
all of the scientific disciplines. It's biology and physics and
chemistry all working together to make us who we truly are. Right,
So every other discipline, if you think of computing, for example,
it's purely electricity. Or if we do drug development or
(03:27):
drug discovery, that's mostly chemistry. Experiments in the lab, that's
mostly biology, But the human brain is all of those
put together.
Speaker 2 (03:39):
The cells in our brain.
Speaker 3 (03:41):
Secrety is chemical substances that diffuse everywhere and hook up
where they need to to trigger certain circuits and then
trigger some effects afterwards.
Speaker 2 (03:52):
So it was just.
Speaker 3 (03:54):
An elegant science that has big impacts.
Speaker 1 (04:00):
We're shortly going to get there and talk about you've
taken on this kind of very technology focused roller Cleveland Clinic.
I'm curious about if we go back to when you
were just starting out at Cleveland Clinic. Yeah, how much
were you thinking about this sort of technology piece about
what technology could do for your specialty, about was that
(04:21):
on your mind or is this something you've come to
more recently.
Speaker 2 (04:26):
Well, it's been a journey.
Speaker 3 (04:27):
Right when I started training as a clinician, my priority
was just to learn how to better care for my
patients and how to be a better physician.
Speaker 2 (04:39):
Right, and then I.
Speaker 3 (04:40):
Realized that clinical practice provides an immediate reward. I'm interacting
with a human being and helping them in the moment,
so there is that immediate reward that comes with that.
But that wasn't enough. I wanted something more. So then
I learned biomedical research. Practic this is and research offered
(05:03):
me this path towards a future. You know, the reward
there is more long term. I'm studying discovering things that
could help many people in the future, even though I
will never get to see them or meet them, or
you know, have that immediate satisfaction. So that shifted me
from being a pure clinician to being a clinician scientist.
(05:26):
And as that journey progressed, it became very clear, as
medicine evolved over the past twenty years, that we cannot
do any good biomedical research without understanding data and technology.
You know, the balance is shifting from most of the
research is happening we call it, you know, on the
wet bench with actual experiments, physical experiments, to a place
(05:51):
where most of the work is happening through compute and
simulations and data. So I became more involved for my
personal research in building big data models and learning about
AI and you know, learning about technology in general. And
then as that journey progressed, I was fortunate enough to
(06:16):
be an enroll for Cleveland Clinic, where my job is
to bring that technology and bridge it to research for
all researchers across our healthcare system.
Speaker 1 (06:28):
When you were talking about how in your own research
you were moving in that direction, what was your own
research focused on. What were you looking at?
Speaker 3 (06:36):
I was looking at brain surgery for epilepsy. It's an
intervention that's been around for decades actually, but when I
started practice, I was shocked by, you know, the practice
that we had where making decisions around surgery like, you know,
what patients should get it versus not, how likely is
(07:00):
it to work, what part of the brain should we remove.
All of those decisions were at the time and the
early two thousands driven by clinical opinions right. You know,
you have an experienced surgeon, they decide to do this.
Somebody else might decide to do something completely different, and
I didn't feel that that was the right.
Speaker 2 (07:21):
Way to practice medicine.
Speaker 3 (07:22):
You know, that we needed to be more evidence based
and data driven. So I went in the business and
research of building models, predictive models that can ingest data
from all of the tests that we would do about
on these patients before to figure out surgery. So I
(07:44):
learned how to analyze all types of data, from genetic
data and individuals to pictures to electrical recordings, and then
combine those into these prediction.
Speaker 1 (07:58):
Models you're looking at You're taking large numbers of surgeries
for epilepsy, and you're seeing what kind of connection there
is between the success of the surgical intervention and the
underlying presentation.
Speaker 2 (08:11):
Of the patient exactly.
Speaker 3 (08:13):
So I would be telling the patient, what is your
specific chance of becoming seizure free with surgery instead of
giving them statistics about you know, like you know, in general,
how well is that going to be effective? So that
piece of individualizing medicine.
Speaker 1 (08:32):
So Cleveland Clinic decides to create a post called Chief
Information Officers.
Speaker 3 (08:36):
If research information we have research. Yes, always chief information
officer runs it.
Speaker 1 (08:42):
Oh, yes, chief research information.
Speaker 2 (08:44):
Yeah, so it for research.
Speaker 1 (08:46):
This is parenthetically a huge job. Yes, So you apply
for this, do you know that you're going to be
thinking and talking and dealing with quantum.
Speaker 3 (08:57):
I started in January twenty twenty, and like every leader
who's put in a position, remember, everybody tells you should
read that the first ninety days great how to plan.
So I was reading that and doing my listening tours
to understand, and then COVID hits and I got a
(09:20):
call from our executive suite about all, there is this
thing called COVID that's coming. We will start testing people
in four days. People will want to do research with COVID.
Speaker 2 (09:33):
Make it happen.
Speaker 3 (09:34):
So it was there's no chapter in the book about that,
you know. So so I then it really it was
like a pressure cooker, you know.
Speaker 2 (09:47):
Test with.
Speaker 3 (09:49):
Yeah, I have to create this access to data, structured
you know, resources, so that we can learn from it
as quickly as possible. And so that was a catalyst.
So then comes twenty twenty one. That was the year
of our centennial one hundred years for Cleveland Clinic. So everybody,
(10:10):
not just me, we were in a mindset where we
were thinking long, long term, you know, like what made
us specialized? Now, how do we stay relevant? Where is
the world going to be ten years from now? And
what should I get going right this moment to shape
that and be ready for it. And that's when quantum
(10:31):
came in my mind, where unless we invest in it
now twenty twenty one, we will not be ready for
this next computer revolution that's coming after AI.
Speaker 1 (10:45):
Everybody in the world right now is doing nothing but
talking about AI, and you are already thinking one step
beyond the quantum. What's different about the opportunity that quantum
creates for a medical research than AI?
Speaker 2 (11:02):
Biology?
Speaker 3 (11:03):
The human body, by definition, is much closer to fundamentals
of quantum, you know, to quantum mechanics and quantum physics
than it.
Speaker 2 (11:16):
Is to AI.
Speaker 3 (11:17):
AI is a classical computing approach that in essence, reduces
every piece of data to a black or white binary
categorization of a one or a zero. At its core
nature around us, the human body, there is nothing categorical
(11:40):
about it. It's that whole, you know, continuum of colors
of life. Quantum its principles are that, you know, so
there's all the scientific principles about quantum physics and superb
position and in tankerment, you know, all of these complex
(12:03):
things that people have a hard time with, but for
in essence, it really is much more aligned. Like if
I want to draw a colored picture, I will not
go and pick up charcoal.
Speaker 2 (12:17):
You know, it's much easier for.
Speaker 3 (12:19):
Me to draw it if I had a colored palette
with me, and quantum offers that. There is plenty of
situations in medicine that are just intractable, you know, meaning
it's not an issue just of it being AI being slow,
or it doesn't have enough data, or if only we
(12:40):
got more GPUs, you know, we can answer those questions.
There are some problems in medicine that are fundamentally such
that even if you give me all the GPUs in
the world, there is no way that AI can model
accurately how these mollyuels in the body are interacting among
(13:03):
each other, or how compounds electrons are moving within the mitochondria.
These are the engines within our cells. There's these fundamental
things in biology that AI and classical computers are just
not built to be able to simulate.
Speaker 1 (13:22):
So you must go to genner parties. Doctor j. High
people must ask you what is quantum computing? What do
you tell them?
Speaker 3 (13:31):
Yes, Although I often, you know, we talk about other
things at dinner.
Speaker 1 (13:35):
Parties, eventually it comes down to you. If I was
at that dinner party with you, I would ask you
what is quantic computer?
Speaker 3 (13:44):
I would say, depends on how much time we have
at the party to explain it.
Speaker 2 (13:48):
The short answer would.
Speaker 3 (13:50):
Be, it's a completely different way of working with computers
than what we're used to. Right now, you can imagine
that AI as being like a car that you take
to go from one place to another. No matter how
fast that Ferrari can get and how much fewer you
(14:10):
put in it, it's never going to be a fighter jet.
Speaker 2 (14:14):
It's never going to be a plane.
Speaker 3 (14:16):
AI is the car, the plane is quant They are
ways to get from point A to point B, but
they work very differently, and we always use them in together.
If I'm flying from Shaker Heights to Yorktown Heights, I
drive to the airport, get on the plane, then take
(14:38):
an uber to get to Yorktown Heights. In research, we
will do the same. We do some piece of it
in AI, some piece of it in quantum, and then
go back and forth, back.
Speaker 1 (14:48):
In twenty twenty one, Cleveland Clinic and IBM announced that
they were starting something called the Discovery Accelerator. What is that.
Speaker 3 (14:56):
It's an initiative, a program, a partnership really that is
designed to bridge advanced computational tools and technology through the
leader in that space, IBM, with biomedical science and research
(15:17):
and life sciences problems, and that is a Cleveland Clinic.
We called it a Discovery Accelerator because that was our goal.
We were both on both sides challenged to the fact
that discovery in medicine was just taking too long. The
classical example that really brought it to life to us
(15:39):
then was drug discovery that it took over a decade
and it still does actually over a decade. From the
moment that there is a compound that someone in a
lab biomedical lab thinks it would be effective to treat
a certain disease, it takes about ten to thirteen from
(16:00):
that moment to when that drug is on a shelf
for a patient to get to from a pharmacy. And
that was just too much of a gap to allow
when we have so many health conditions that we needed
to address, and a big part of that gap could
be computationally solved better simulation of compounds, designing drug trails
(16:22):
that are more efficient, you know, that would finish faster.
So that was the motivation to bring computational tools and
technology closer to biomedical researchers.
Speaker 1 (16:36):
Cleveland Click and IBM team up. And I'm assuming the
IBM quantum guys and other people descend on Cleveland and
you have your first meeting with them. Are they telling
you things you would never thought of? Or I mean,
I'm just curious about what's the difference between what you
thought was the potential was and what you discovered the
(16:58):
potential was.
Speaker 3 (17:01):
That is an excellent question. If I had to prioritize.
One lesson that I learned over the past few years
of doing this, it is that you can never know
what your you know where your brain is going to
go and discovery until you talk. You know, you have
(17:23):
to open it up and really listen to try to
learn what the other people are saying.
Speaker 2 (17:29):
It's it went both ways.
Speaker 1 (17:31):
Can you give you an example.
Speaker 3 (17:32):
So, okay, we built a quantum right, so it took
it took like eight months to get this machine put together,
and we put it in our cafeteria.
Speaker 2 (17:41):
That's the whole other story where you.
Speaker 1 (17:43):
Put the quantum machine computer in your cafeteria. Yes, yes, yes,
can you see it when you're eating lunch?
Speaker 2 (17:52):
Yeah?
Speaker 3 (17:52):
Yeah, we have people eating lunch around it all the time.
We wanted to have a machine that people can see.
Otherwise it it the program wouldn't launch properly, so I
needed to.
Speaker 2 (18:03):
Have it physically.
Speaker 3 (18:05):
So we had to look at retrofitted in existing space,
and the cafeteria space worked out. It was far enough
from the street, you know, there was no vibration that
it was a double floor, you know, so the ceiling
was high enough. It just like technically fit all of
those requirements. And it was either there or we put
(18:28):
it in our data center, which is a building and
like another city close to Cleveland, and we picked the cafeteria.
Speaker 1 (18:37):
Yeah. Now, why I know this is sort of this
is kind of hilarious, But there's a serious point I
I want to touch on, which is why do you
need it on premises?
Speaker 2 (18:46):
On the premises Cleveland Clinic research?
Speaker 3 (18:49):
We needed to change how we think about research and
shift the mindset of all of our researchers, and all
of our researchers I'm talking about three thousand individuals who
are one hundred percent doing biomedical research in Cleveland Clinic,
(19:10):
two hundred and thirty labs individual pis. So that's the
scale that I'm talking about that we had to create
an impact on.
Speaker 2 (19:21):
So having it, having it.
Speaker 3 (19:24):
Be there was as much for inspiration and to trigger
our motivation to change as it was a you know,
a practical solution, say, because we had the space and
the connections and all of that.
Speaker 1 (19:41):
I saw in it. IBM headquarters. They're beautiful.
Speaker 3 (19:44):
Their works of art they are They're gorgeous, and the
one that we have is the most gorgeous.
Speaker 2 (19:53):
One of all. This is not me, you know, the mom.
Speaker 4 (19:57):
Bias talking about it.
Speaker 3 (20:00):
You know, they got an award, the Red Dot Award
for Design went to IBM and Cleveland Clinic for our
quantum and the quantum that IBM built for RPI a
couple of years after hours. They modeled it after hours,
(20:21):
not after the you know the ones from before.
Speaker 1 (20:24):
Just just so people know, we're talking about basically a
small garage.
Speaker 3 (20:30):
But that size eleven foot by eleven eleven feet eleven feet, you.
Speaker 1 (20:34):
Know, the cube small, So.
Speaker 3 (20:36):
It's a it's a glass cube, and the glass comes
all the way from Italy. It's the same glass that
protects the crown jewels and you know the Mona Lisa
and all that. So so there's a glass all around,
and then there is the tube that you see, the
(20:58):
stainless tube.
Speaker 2 (20:59):
That's shiny, you know, and clean.
Speaker 3 (21:01):
But the technologies are inside of it, you know, the
chandelier and then the processor and the bottom and it's
just a fascinating thing to watch and it hums. It
makes the sound so even it sounds alive.
Speaker 1 (21:19):
You have a great deal of affection for your computer.
Speaker 2 (21:22):
And we love our machine.
Speaker 1 (21:24):
I want to go back to something you said before,
which is when you had your initial conversations with IBM,
both sides learned things that they hadn't previously thought of.
Give me an example of something that either side hadn't
realized could be done with this new technology.
Speaker 3 (21:42):
Sure, I mean on the equivalent clinic side, we thought
that what quantum should be good for is that it
would be a faster computer, right, so that we have
these big data sets that are requiring much more compute
power and you know, GIP used than what we have
and we should just run them on the quantum. And
(22:04):
what we came to realize is that Quantum actually does
not do well with these large data sets. What it
does well with our smaller, better defined data sets, but
ones where simulation is more important. You know, we have
to run them through multiple models, multiple simulations of how
(22:27):
they interact. As an example, one of the very first
projects we throw at Quantum was wanting to predict complications
cardiac complications from general surgery using electronic health record data.
So data from our electronic health records are by definition
these really big data sets.
Speaker 2 (22:47):
You have in them every single thing.
Speaker 3 (22:49):
That you know about the patient that you've collected, and
we thought that Quantum would help us build models, better
models with this data.
Speaker 2 (23:00):
And it failed miserably. It wasn't good, you know, at
doing that thing.
Speaker 1 (23:05):
And why didn't it like that problem because it's too
it's too bigger than wieldy.
Speaker 3 (23:10):
Well, because the hardware isn't tready right, So the hardware
with Quantum, that issue was back then. You know, now
we've upgraded our processor. But still Quantum now is limited
by noise and by its ability to correct for errors
when it's doing computation. And the more data that you
(23:34):
throw at it that you require it, you know, to
put in its system, you know to model interactions, the
more errors it's likely to make, so then the harder
it is for it to get to an answer that
you can trust. Now, we made a lot of progress sense,
which I'm sure we'll get to I hope we'll get
(23:55):
to with some recent breakthroughs that we made in that
space with modeling large compounds and interactions. But the way
that got us to where we are now, where we
could model large interactions, it took us figuring out that
we shouldn't be doing everything on quantum.
Speaker 1 (24:15):
So give me an example of a problem that quantum
is ideally suited for that the quantum really.
Speaker 3 (24:22):
Likes in drug discovery, for examine chemistry. It likes chemistry
a lot because in chemistry what it has to what
we wanted to model is how a drug that we
put in our body is going to interact with We
say the protein you know, the target, the ligand that
(24:43):
it needs to buind too. Like you know, the drug
is a key and it needs to fit in a
lock in certain parts of the body to open it,
get in, do it stay. And there are many keys,
many potential compounds that we could test for many parts
of our body. We don't really know how they're going
(25:03):
to interact. Traditionally, what we do is we have to
build all the keys. We have to manufacture all these
compounds and then do actual trial clinical trials. Put them
in people, put them in animals, and see what happens.
Speaker 1 (25:14):
See which one is best?
Speaker 3 (25:15):
See yeah, which like two fit best together. So we
use AI to try to help us with that. But
AI can only model what it had learned, right, So
for rare diseases conditions that there isn't enough information out
there on what the you know, the locks look like,
(25:37):
it's really hard to then model it with AI and
get an accurate prediction of whether there's a fit or not.
Quantum does not rely on the previous data for its modeling.
Quantum does its modeling purely based on the physical characteristics
of the compound of the key you know that you're designing.
(26:00):
So that makes it ideal because it's then untethered with
It's not limited by do we have enough samples or
don't we have enough samples, or you know, what the
previous studies find or not. It's purely based on those
physical properties, those quantum properties. So we found ourselves in
situations where we were able to predict the fits between
(26:26):
certain targets and where in conditions like Alzheimer's disease. For example,
we published where the quantum based modeling of the compound
was much better than what we would have gotten with
what we got right, Like, we did it both ways,
(26:46):
and the quantum one was the better fit than the
AI generated one.
Speaker 1 (26:51):
This is fair. Quantum is a little more of an
artist and a little less of a of a kind
of nerds nerdy to me, what seems like creative and artistic?
Speaker 2 (27:03):
Creative?
Speaker 1 (27:04):
Yes, creative, it's like people.
Speaker 3 (27:07):
Yeah, it's it opens up pass that you never knew existed.
That's why when I get asked about what do I
see is the best, you know, the the most important
breakthrough that quantum is going to allow us to do,
my answer is I really don't know, because we I
(27:28):
you know, when other people invented these new technologies, I
don't think they really knew that they're you know, like
think of laser. I don't think the person who invented
laser thought that they will be used to scan groceries
at the grocery store.
Speaker 2 (27:44):
You know.
Speaker 3 (27:45):
But so technology developing technology, the way I think of
it is like having a baby you know, you raise
it as best you can, but then they're going to
go off and do their thing, and you will be
tying them down if you restrict them to just what
you thought they should do, you know.
Speaker 2 (28:04):
So it opens up that.
Speaker 3 (28:09):
Space, that creative space for us to ask questions differently
than we used to. We should train our mind to
stop starting from classical and then trying to squeeze it
into quantum. We have to learn how to think quantum
up front, right from the beginning, which we haven't really
(28:33):
been doing as a scientific community since our inception. We
were trained, and how do you convert what you're thinking
into a formula that you can ask from a computer
which is a classical right computer with quantum because of
how it works, it can answer questions, It can look
(28:54):
at problems in a very different way. So we have
to think differently about the questions and how we ask them.
Speaker 1 (29:02):
With IBM, you recently modeled protein with over twelve thousand atoms.
Talk to me about that and why it's so meaningful
for job discovery.
Speaker 3 (29:12):
So in October of twenty twenty four, so just eighteen
months ago, the largest compound biological compound that could be
simulated with quantum was ten atoms big, and it was
unfathomable back then that we will get past the thousand
(29:35):
or few thousand atom simulation in the foreseeable future. And
what our team with IBM and with Rieken and Japan
published last month April twenty twenty six is a simulation
of the electrical properties of an enzyme trips in the
(30:00):
body that is twelve thousand, six hundred atoms bait for
you know, orders of magnitude beyond what anybody thought was
possible in that span of time. And the reason why
that happened was because of a you know, innovations in
(30:21):
the technology itself where the teams stopped thinking of quantum
and AI as competitors and instead thought of them as
different members.
Speaker 2 (30:35):
Of the same team.
Speaker 3 (30:37):
Right, you're we're now in basketball season in the US
the NPA, and you need the you need defense, but
you also need your center, somebody to shoot.
Speaker 5 (30:49):
You need all of the pieces to work together. So
with this, it was figuring out where do I put
you know, when do I put AI on the field,
When do I put Quantum on the field, and how
do I tell them.
Speaker 2 (31:04):
To work together?
Speaker 3 (31:06):
It's the scientific terms the quantum centric super computing. So
quantum is at the center, but we're using our supercomputing
tools AI classical to interact with it and split that
big problem of the twelve thousand, six hundred atoms into pieces,
(31:27):
where some pieces are best served with quantum and others
are best served with classically.
Speaker 1 (31:33):
This distinction that we now cling to AI and quantum
are these very different things develop by different people for
different purposes. What you're suggesting is that's probably going to
go away. Yeah, in the future, these things will work together.
It's going to become teamwork and not one on one
competition exactly.
Speaker 2 (31:54):
And it is that now in Keeveland Clinic.
Speaker 3 (31:56):
I mean, the way we've evolved our priority is with IBM.
It's a realization that both organizations have come to where
really too for progress to happen, we should stop seeing
them as separate. We we should put them together and
work to the best of what each piece of technology
(32:19):
can provide.
Speaker 1 (32:21):
One theme running through a lot of your your what
you've been talking about is that the arrival of this
new technology requires the researcher to behave differently and that's
one of the reasons why you want the quantum machine
(32:42):
and the on full display, and then you were talking
about how you have to ask different kinds of questions.
I'm curious, can you can you? Can you elaborate on
that a little bit? So take me back, for example,
to your earliest days. If I had given you all
these tools that people have now, how would your research
(33:02):
have proceeded differently? What would you have done differently?
Speaker 2 (33:07):
You know, I would have looked at the molecular.
Speaker 3 (33:17):
Components of the human brain as they relate to outcomes
of brain surgery way earlier than I did. The first
ten years of my career doing research was all spent
building models that were purely based on brain waves and
pictures of the brain. It wasn't until after I hit
(33:43):
a wall with my models aren't getting past that eighty
percent accuracy threshold that I started thinking, oh, you know,
there must be something genetic or you know, more biological
that is influencing this. Had I been exposed to quantum
(34:03):
at least as a concept, right to quantum computing and
what quantum science is back then, I think it would
have opened up my mind to realize that it's not
just about what I see. You know, there is hidden
relationships that exist within the human body.
Speaker 2 (34:26):
And that's our genetic.
Speaker 3 (34:27):
Makeup and our chemical makeup that influence what comes to
the surface that urge to dig deeper. The other thing
it would have changed is it would have made me
reach out to engineers and physicists and mathematicians much earlier
in my career.
Speaker 1 (34:48):
Yeah, yeah, yeah. And how would it have changed is
a very kind of prosaic question, but just kind of
the day to day life of someone doing better research.
I mean, the oh wow, you walk into the office
in the morning, How does your day proceed differently when
you're when you have these kinds of tools at your fingertips.
Speaker 3 (35:12):
That's the fundamental question in biomedical research right now, and
it's it's less to do with quantum, more to do
with urgentic AI, right these agents that we can work
with now to help us how to think more creatively
and how to do work that up until now was
(35:36):
more like Scott work that the researchers had to do,
whether it was you know, so the hypothesis generation has
always been the most creative part of aspect of scientific research.
But what comes after it with the data collection, for example,
that was always such a repetitive, you know, exercise, and
(36:01):
then the analysis after that was something that was very
resource intensive and you had to try so many different
approaches before you get to an answer, and that was
fairly complex.
Speaker 2 (36:16):
Now, with access to.
Speaker 3 (36:20):
Urgantic AI and you know, some quantum, of course, we
can spend more of our energy on the creative thinking
part of the aspects of the work and less on
the you know, just that repetitive.
Speaker 1 (36:39):
When you look around, I'm assuming you walk around Cleveland
Clinic and you have lots of conversations with some of
the most brilliant medical researchers in the world. Are you
satisfied with how quickly and aggressively they are adopting these
two technologies or do they still need encouragement from you
to do? You have to say people, wait, I've got
(37:02):
this machine in the cafeteria. You should be using it
for this problem. How much are you doing that kind
of encouraging in cheerleading or is it unnecessary?
Speaker 3 (37:11):
No, there's plenty of cheerleading that's necessary people. You know,
humans don't like to change. It's hardwired in us. So
there is plenty of cheerleading. But what happens is, I mean,
the way we built our program to where we are now,
(37:32):
so Cleveland clinic now is pretty much winning every global competition.
And Quantum for life sciences, whether that's the welcome trust,
you know, Quantum for biological applications. We our partner. There
was a startup in Finland, Algorithmic.
Speaker 2 (37:50):
They're brilliant.
Speaker 3 (37:51):
We worked with them to develop a photodynamic drug therapy
for cancer, or whether it is the NIH with they
had an expert price challenge for Quantum or universities that
we're collaborating with. We have a program that's bringing startups
to our ecosystem.
Speaker 2 (38:10):
We give them access to the.
Speaker 3 (38:11):
Machine if they have a question that is significant enough
for life sciences, universities that we're building a bachelor's, master's,
PhD programs with on quantum computing, we developed that whole
ecosystem around it. If that's not cheerleading, I don't know
(38:33):
what else would qualify for cheerleading. But then what happens
is these early adopters, the risk takers, become the cheerleaders themselves, right,
and then they work with it, they achieve success, and
we're nothing but competitive in medicine and science. Right, So
then it becomes okay, so and so I did this
(38:54):
with this machine, let me learn it. So I can
do the same, and it becomes this verse virtuous cycle
of then innovation and growth and people wanting to work together.
Speaker 2 (39:08):
It's just been fascinating to watch.
Speaker 1 (39:11):
You said that people don't like change, but you clearly do.
Speaker 2 (39:15):
I do. That's the mindset of researchers, right.
Speaker 3 (39:19):
A researcher is someone who is not afraid to fail,
actually sees failure as a chance to learn something right,
to do better the next time. So we get rejected
all the time and we don't care, right, we keep
moving on with papers, grant applications.
Speaker 2 (39:37):
So that mindset is what all of research is built around.
Speaker 3 (39:44):
So it's a very forward looking mindset and leveland clinic
as the health system, we wouldn't have survived one hundred years.
We wouldn't have done all of the firsts that we had.
Serotonin the chemical that drives the whole all science of
psychiatry and neuroscience that was discovered in Cleveland clinic. So
(40:07):
as an organization, we have enough of people who think
that way that they will be the early adopters who
will pull the others with them.
Speaker 1 (40:21):
Yeah. One last question, look look ahead tenures, pat me
a picture of what quantum and related technologies look like
for a place like Cleveland Clinic and tenures.
Speaker 3 (40:36):
Well, what I hope is that ten years from now,
if I if I'm seeing a patient in clinic who
has bad epilepsy and for the love of you know,
I can't figure out what medicine do I need to
prescribe to them to make them seizure free, I will
be able to send them to get a blood test
(40:58):
that we can then run through some analytical platform that
leverages both quantum and AI that can model exactly for
that patient, what compound, either existing or not to be developed.
It's some new chemical that we haven't tested for that
(41:21):
indication yet is going to treat them, you know, make
them seizure free. It's and I think quantum is is
ideal for a condition like mine epilepsy, because we are
a rare disease. And I think that benefit that it's
(41:41):
going to have will start with rare diseases, you know,
as I explained earlier, So take any other rare disease,
we should start there and then expand from that to
more complex things like cancer for example, and others. But
those rare conditions where we are left now completely scratching
(42:02):
our heads and going by intuition, it will make our
care truly precise, and it will make drug development a
more tailored exercise than the way it is now, where
it's like a hammer that's looking for nails?
Speaker 1 (42:22):
Am I right in thinking that of all of the
over the one hundred year history or more now of
Cleveland clinic, this sounds like the absolute best time to
be a Cleveland clinic.
Speaker 2 (42:31):
Yeah.
Speaker 3 (42:31):
I love it, no complaints, no complaints, And I am hiring.
Speaker 4 (42:39):
I need I need those quantum researchers, those people who
are wanting, you know, to apply quant genetic research, imaging research,
every single aspect of biomedical research.
Speaker 3 (42:54):
We can't afford to just wait on the sideline and
until the technolog she's ready, and then you know, then
it was teach us. It is ready now that twelve
thousand atom simulation wouldn't have happened just a year and
a half ago.
Speaker 2 (43:10):
It's quidinn.
Speaker 1 (43:12):
So the headline of this conversation is we're hiring, we hired.
Speaker 2 (43:15):
Yes, Yeah, that's a good headline. We're growing, how about that.
Speaker 1 (43:24):
Smart Talks with IBM is produced by Matt Ramano, Amy
Gains McQuaid, Trina Menino, and Jake Harper Engineering by Nina Bird,
Lawrence Mastering by Sarah Buguerer, music by Gramoscope, Strategy by
Cassidy Meyer, Sophia Derlon and Tatiana Lieberman. Special thanks to
doctor Laura jay Hi, Alicia Real Cooney, and the Cleveland
(43:45):
Clinic team. Smart Talks with IBM is a production of
Pushkin Industries and Ruby Studio at iHeartMedia. To find more
Pushkin podcasts, listen on the iHeartRadio app, Apple Podcasts, or
wherever you listen to. I'm Malcolm Gladwell. This is a
paid advertisement from IBM. The conversations on this podcast don't
(44:07):
necessarily represent IBM's positions, strategies, or opinions.
I'm Malcolm Glabwell and you're listening to Smart Talks with IBM.
When doctor Laura Jahi began treating epilepsy patients in the
early two thousands, she noticed something unsettling. Different surgeons could
look at the exact same case and recommend completely different treatments.
One surgeon might remove one part of the brain, another
(00:24):
a different part, and a third might not operate at all.
Doctor j Hi believed there had to be a better way,
one grounded in data. That conviction set her on a
path that would lead her to become Chief Research Information
Officer at Cleveland Clinic and the executive program lead for
the Discovery Accelerator. The Discovery Accelerator is a ten year
(00:47):
partnership between Cleveland Clinic and IBM where researchers are using
AI and quantum computing to make incredible discoveries in healthcare
and life sciences. I sat down with doctor j High
to explore what's happening now, what's possible with quantum computing,
and where this next era of biomedical discovery is headed.
(01:10):
Epilepsy was your kind of specialty within neurology?
Speaker 2 (01:13):
Correct? Yes?
Speaker 1 (01:13):
What led you to that?
Speaker 3 (01:15):
I was always fascinated by the brain. You know, it's
the part of our body that leaves still to this day,
the most to be discovered. So I was always intrigued
by areas that leave more for discovery, and epilepsy was
my pragmatic side, wanting to choose a subspecialty where the
(01:38):
problem can be fixed. In epilepsy, there are many medications
that are very effective, and there's a brain surgery that
we can do to stop seizures when medicines don't work.
Speaker 2 (01:50):
That attracted me, you.
Speaker 3 (01:52):
Know, compared to other areas in neurology, like stroke, for example,
or dementia, where usually the damage is more. I wanted
to be able to tell my patients that you have
a big problem, but here's what I can do to
fix it, and epilepsy offered me that.
Speaker 1 (02:10):
But what would there must have been interesting and intriguing
unsolved problems in neurology.
Speaker 3 (02:19):
Oh my gosh, it's the whole brain, isn't it. Before
starting to deal with artificial intelligence and research. Right in
the neurology, I'm dealing with real intelligence, the human brain
and how it works and how we think and how
we make decisions, and how it can grow and evolve,
(02:40):
and so there is many untaped questions in neurology, and
that's part of the fascination in it. In epilepsy in particular,
it's an electrical disease in the brain, it's actually one
of the conditions in neurology. That's a perfect alignment of
(03:05):
all of the scientific disciplines. It's biology and physics and
chemistry all working together to make us who we truly are. Right,
So every other discipline, if you think of computing, for example,
it's purely electricity. Or if we do drug development or
(03:27):
drug discovery, that's mostly chemistry. Experiments in the lab, that's
mostly biology, But the human brain is all of those
put together.
Speaker 2 (03:39):
The cells in our brain.
Speaker 3 (03:41):
Secrety is chemical substances that diffuse everywhere and hook up
where they need to to trigger certain circuits and then
trigger some effects afterwards.
Speaker 2 (03:52):
So it was just.
Speaker 3 (03:54):
An elegant science that has big impacts.
Speaker 1 (04:00):
We're shortly going to get there and talk about you've
taken on this kind of very technology focused roller Cleveland Clinic.
I'm curious about if we go back to when you
were just starting out at Cleveland Clinic. Yeah, how much
were you thinking about this sort of technology piece about
what technology could do for your specialty, about was that
(04:21):
on your mind or is this something you've come to
more recently.
Speaker 2 (04:26):
Well, it's been a journey.
Speaker 3 (04:27):
Right when I started training as a clinician, my priority
was just to learn how to better care for my
patients and how to be a better physician.
Speaker 2 (04:39):
Right, and then I.
Speaker 3 (04:40):
Realized that clinical practice provides an immediate reward. I'm interacting
with a human being and helping them in the moment,
so there is that immediate reward that comes with that.
But that wasn't enough. I wanted something more. So then
I learned biomedical research. Practic this is and research offered
(05:03):
me this path towards a future. You know, the reward
there is more long term. I'm studying discovering things that
could help many people in the future, even though I
will never get to see them or meet them, or
you know, have that immediate satisfaction. So that shifted me
from being a pure clinician to being a clinician scientist.
(05:26):
And as that journey progressed, it became very clear, as
medicine evolved over the past twenty years, that we cannot
do any good biomedical research without understanding data and technology.
You know, the balance is shifting from most of the
research is happening we call it, you know, on the
wet bench with actual experiments, physical experiments, to a place
(05:51):
where most of the work is happening through compute and
simulations and data. So I became more involved for my
personal research in building big data models and learning about
AI and you know, learning about technology in general. And
then as that journey progressed, I was fortunate enough to
(06:16):
be an enroll for Cleveland Clinic, where my job is
to bring that technology and bridge it to research for
all researchers across our healthcare system.
Speaker 1 (06:28):
When you were talking about how in your own research
you were moving in that direction, what was your own
research focused on. What were you looking at?
Speaker 3 (06:36):
I was looking at brain surgery for epilepsy. It's an
intervention that's been around for decades actually, but when I
started practice, I was shocked by, you know, the practice
that we had where making decisions around surgery like, you know,
what patients should get it versus not, how likely is
(07:00):
it to work, what part of the brain should we remove.
All of those decisions were at the time and the
early two thousands driven by clinical opinions right. You know,
you have an experienced surgeon, they decide to do this.
Somebody else might decide to do something completely different, and
I didn't feel that that was the right.
Speaker 2 (07:21):
Way to practice medicine.
Speaker 3 (07:22):
You know, that we needed to be more evidence based
and data driven. So I went in the business and
research of building models, predictive models that can ingest data
from all of the tests that we would do about
on these patients before to figure out surgery. So I
(07:44):
learned how to analyze all types of data, from genetic
data and individuals to pictures to electrical recordings, and then
combine those into these prediction.
Speaker 1 (07:58):
Models you're looking at You're taking large numbers of surgeries
for epilepsy, and you're seeing what kind of connection there
is between the success of the surgical intervention and the
underlying presentation.
Speaker 2 (08:11):
Of the patient exactly.
Speaker 3 (08:13):
So I would be telling the patient, what is your
specific chance of becoming seizure free with surgery instead of
giving them statistics about you know, like you know, in general,
how well is that going to be effective? So that
piece of individualizing medicine.
Speaker 1 (08:32):
So Cleveland Clinic decides to create a post called Chief
Information Officers.
Speaker 3 (08:36):
If research information we have research. Yes, always chief information
officer runs it.
Speaker 1 (08:42):
Oh, yes, chief research information.
Speaker 2 (08:44):
Yeah, so it for research.
Speaker 1 (08:46):
This is parenthetically a huge job. Yes, So you apply
for this, do you know that you're going to be
thinking and talking and dealing with quantum.
Speaker 3 (08:57):
I started in January twenty twenty, and like every leader
who's put in a position, remember, everybody tells you should
read that the first ninety days great how to plan.
So I was reading that and doing my listening tours
to understand, and then COVID hits and I got a
(09:20):
call from our executive suite about all, there is this
thing called COVID that's coming. We will start testing people
in four days. People will want to do research with COVID.
Speaker 2 (09:33):
Make it happen.
Speaker 3 (09:34):
So it was there's no chapter in the book about that,
you know. So so I then it really it was
like a pressure cooker, you know.
Speaker 2 (09:47):
Test with.
Speaker 3 (09:49):
Yeah, I have to create this access to data, structured
you know, resources, so that we can learn from it
as quickly as possible. And so that was a catalyst.
So then comes twenty twenty one. That was the year
of our centennial one hundred years for Cleveland Clinic. So everybody,
(10:10):
not just me, we were in a mindset where we
were thinking long, long term, you know, like what made
us specialized? Now, how do we stay relevant? Where is
the world going to be ten years from now? And
what should I get going right this moment to shape
that and be ready for it. And that's when quantum
(10:31):
came in my mind, where unless we invest in it
now twenty twenty one, we will not be ready for
this next computer revolution that's coming after AI.
Speaker 1 (10:45):
Everybody in the world right now is doing nothing but
talking about AI, and you are already thinking one step
beyond the quantum. What's different about the opportunity that quantum
creates for a medical research than AI?
Speaker 2 (11:02):
Biology?
Speaker 3 (11:03):
The human body, by definition, is much closer to fundamentals
of quantum, you know, to quantum mechanics and quantum physics
than it.
Speaker 2 (11:16):
Is to AI.
Speaker 3 (11:17):
AI is a classical computing approach that in essence, reduces
every piece of data to a black or white binary
categorization of a one or a zero. At its core
nature around us, the human body, there is nothing categorical
(11:40):
about it. It's that whole, you know, continuum of colors
of life. Quantum its principles are that, you know, so
there's all the scientific principles about quantum physics and superb
position and in tankerment, you know, all of these complex
(12:03):
things that people have a hard time with, but for
in essence, it really is much more aligned. Like if
I want to draw a colored picture, I will not
go and pick up charcoal.
Speaker 2 (12:17):
You know, it's much easier for.
Speaker 3 (12:19):
Me to draw it if I had a colored palette
with me, and quantum offers that. There is plenty of
situations in medicine that are just intractable, you know, meaning
it's not an issue just of it being AI being slow,
or it doesn't have enough data, or if only we
(12:40):
got more GPUs, you know, we can answer those questions.
There are some problems in medicine that are fundamentally such
that even if you give me all the GPUs in
the world, there is no way that AI can model
accurately how these mollyuels in the body are interacting among
(13:03):
each other, or how compounds electrons are moving within the mitochondria.
These are the engines within our cells. There's these fundamental
things in biology that AI and classical computers are just
not built to be able to simulate.
Speaker 1 (13:22):
So you must go to genner parties. Doctor j. High
people must ask you what is quantum computing? What do
you tell them?
Speaker 3 (13:31):
Yes, Although I often, you know, we talk about other
things at dinner.
Speaker 1 (13:35):
Parties, eventually it comes down to you. If I was
at that dinner party with you, I would ask you
what is quantic computer?
Speaker 3 (13:44):
I would say, depends on how much time we have
at the party to explain it.
Speaker 2 (13:48):
The short answer would.
Speaker 3 (13:50):
Be, it's a completely different way of working with computers
than what we're used to. Right now, you can imagine
that AI as being like a car that you take
to go from one place to another. No matter how
fast that Ferrari can get and how much fewer you
(14:10):
put in it, it's never going to be a fighter jet.
Speaker 2 (14:14):
It's never going to be a plane.
Speaker 3 (14:16):
AI is the car, the plane is quant They are
ways to get from point A to point B, but
they work very differently, and we always use them in together.
If I'm flying from Shaker Heights to Yorktown Heights, I
drive to the airport, get on the plane, then take
(14:38):
an uber to get to Yorktown Heights. In research, we
will do the same. We do some piece of it
in AI, some piece of it in quantum, and then
go back and forth, back.
Speaker 1 (14:48):
In twenty twenty one, Cleveland Clinic and IBM announced that
they were starting something called the Discovery Accelerator. What is that.
Speaker 3 (14:56):
It's an initiative, a program, a partnership really that is
designed to bridge advanced computational tools and technology through the
leader in that space, IBM, with biomedical science and research
(15:17):
and life sciences problems, and that is a Cleveland Clinic.
We called it a Discovery Accelerator because that was our goal.
We were both on both sides challenged to the fact
that discovery in medicine was just taking too long. The
classical example that really brought it to life to us
(15:39):
then was drug discovery that it took over a decade
and it still does actually over a decade. From the
moment that there is a compound that someone in a
lab biomedical lab thinks it would be effective to treat
a certain disease, it takes about ten to thirteen from
(16:00):
that moment to when that drug is on a shelf
for a patient to get to from a pharmacy. And
that was just too much of a gap to allow
when we have so many health conditions that we needed
to address, and a big part of that gap could
be computationally solved better simulation of compounds, designing drug trails
(16:22):
that are more efficient, you know, that would finish faster.
So that was the motivation to bring computational tools and
technology closer to biomedical researchers.
Speaker 1 (16:36):
Cleveland Click and IBM team up. And I'm assuming the
IBM quantum guys and other people descend on Cleveland and
you have your first meeting with them. Are they telling
you things you would never thought of? Or I mean,
I'm just curious about what's the difference between what you
thought was the potential was and what you discovered the
(16:58):
potential was.
Speaker 3 (17:01):
That is an excellent question. If I had to prioritize.
One lesson that I learned over the past few years
of doing this, it is that you can never know
what your you know where your brain is going to
go and discovery until you talk. You know, you have
(17:23):
to open it up and really listen to try to
learn what the other people are saying.
Speaker 2 (17:29):
It's it went both ways.
Speaker 1 (17:31):
Can you give you an example.
Speaker 3 (17:32):
So, okay, we built a quantum right, so it took
it took like eight months to get this machine put together,
and we put it in our cafeteria.
Speaker 2 (17:41):
That's the whole other story where you.
Speaker 1 (17:43):
Put the quantum machine computer in your cafeteria. Yes, yes, yes,
can you see it when you're eating lunch?
Speaker 2 (17:52):
Yeah?
Speaker 3 (17:52):
Yeah, we have people eating lunch around it all the time.
We wanted to have a machine that people can see.
Otherwise it it the program wouldn't launch properly, so I
needed to.
Speaker 2 (18:03):
Have it physically.
Speaker 3 (18:05):
So we had to look at retrofitted in existing space,
and the cafeteria space worked out. It was far enough
from the street, you know, there was no vibration that
it was a double floor, you know, so the ceiling
was high enough. It just like technically fit all of
those requirements. And it was either there or we put
(18:28):
it in our data center, which is a building and
like another city close to Cleveland, and we picked the cafeteria.
Speaker 1 (18:37):
Yeah. Now, why I know this is sort of this
is kind of hilarious, But there's a serious point I
I want to touch on, which is why do you
need it on premises?
Speaker 2 (18:46):
On the premises Cleveland Clinic research?
Speaker 3 (18:49):
We needed to change how we think about research and
shift the mindset of all of our researchers, and all
of our researchers I'm talking about three thousand individuals who
are one hundred percent doing biomedical research in Cleveland Clinic,
(19:10):
two hundred and thirty labs individual pis. So that's the
scale that I'm talking about that we had to create
an impact on.
Speaker 2 (19:21):
So having it, having it.
Speaker 3 (19:24):
Be there was as much for inspiration and to trigger
our motivation to change as it was a you know,
a practical solution, say, because we had the space and
the connections and all of that.
Speaker 1 (19:41):
I saw in it. IBM headquarters. They're beautiful.
Speaker 3 (19:44):
Their works of art they are They're gorgeous, and the
one that we have is the most gorgeous.
Speaker 2 (19:53):
One of all. This is not me, you know, the mom.
Speaker 4 (19:57):
Bias talking about it.
Speaker 3 (20:00):
You know, they got an award, the Red Dot Award
for Design went to IBM and Cleveland Clinic for our
quantum and the quantum that IBM built for RPI a
couple of years after hours. They modeled it after hours,
(20:21):
not after the you know the ones from before.
Speaker 1 (20:24):
Just just so people know, we're talking about basically a
small garage.
Speaker 3 (20:30):
But that size eleven foot by eleven eleven feet eleven feet, you.
Speaker 1 (20:34):
Know, the cube small, So.
Speaker 3 (20:36):
It's a it's a glass cube, and the glass comes
all the way from Italy. It's the same glass that
protects the crown jewels and you know the Mona Lisa
and all that. So so there's a glass all around,
and then there is the tube that you see, the
(20:58):
stainless tube.
Speaker 2 (20:59):
That's shiny, you know, and clean.
Speaker 3 (21:01):
But the technologies are inside of it, you know, the
chandelier and then the processor and the bottom and it's
just a fascinating thing to watch and it hums. It
makes the sound so even it sounds alive.
Speaker 1 (21:19):
You have a great deal of affection for your computer.
Speaker 2 (21:22):
And we love our machine.
Speaker 1 (21:24):
I want to go back to something you said before,
which is when you had your initial conversations with IBM,
both sides learned things that they hadn't previously thought of.
Give me an example of something that either side hadn't
realized could be done with this new technology.
Speaker 3 (21:42):
Sure, I mean on the equivalent clinic side, we thought
that what quantum should be good for is that it
would be a faster computer, right, so that we have
these big data sets that are requiring much more compute
power and you know, GIP used than what we have
and we should just run them on the quantum. And
(22:04):
what we came to realize is that Quantum actually does
not do well with these large data sets. What it
does well with our smaller, better defined data sets, but
ones where simulation is more important. You know, we have
to run them through multiple models, multiple simulations of how
(22:27):
they interact. As an example, one of the very first
projects we throw at Quantum was wanting to predict complications
cardiac complications from general surgery using electronic health record data.
So data from our electronic health records are by definition
these really big data sets.
Speaker 2 (22:47):
You have in them every single thing.
Speaker 3 (22:49):
That you know about the patient that you've collected, and
we thought that Quantum would help us build models, better
models with this data.
Speaker 2 (23:00):
And it failed miserably. It wasn't good, you know, at
doing that thing.
Speaker 1 (23:05):
And why didn't it like that problem because it's too
it's too bigger than wieldy.
Speaker 3 (23:10):
Well, because the hardware isn't tready right, So the hardware
with Quantum, that issue was back then. You know, now
we've upgraded our processor. But still Quantum now is limited
by noise and by its ability to correct for errors
when it's doing computation. And the more data that you
(23:34):
throw at it that you require it, you know, to
put in its system, you know to model interactions, the
more errors it's likely to make, so then the harder
it is for it to get to an answer that
you can trust. Now, we made a lot of progress sense,
which I'm sure we'll get to I hope we'll get
(23:55):
to with some recent breakthroughs that we made in that
space with modeling large compounds and interactions. But the way
that got us to where we are now, where we
could model large interactions, it took us figuring out that
we shouldn't be doing everything on quantum.
Speaker 1 (24:15):
So give me an example of a problem that quantum
is ideally suited for that the quantum really.
Speaker 3 (24:22):
Likes in drug discovery, for examine chemistry. It likes chemistry
a lot because in chemistry what it has to what
we wanted to model is how a drug that we
put in our body is going to interact with We
say the protein you know, the target, the ligand that
(24:43):
it needs to buind too. Like you know, the drug
is a key and it needs to fit in a
lock in certain parts of the body to open it,
get in, do it stay. And there are many keys,
many potential compounds that we could test for many parts
of our body. We don't really know how they're going
(25:03):
to interact. Traditionally, what we do is we have to
build all the keys. We have to manufacture all these
compounds and then do actual trial clinical trials. Put them
in people, put them in animals, and see what happens.
Speaker 1 (25:14):
See which one is best?
Speaker 3 (25:15):
See yeah, which like two fit best together. So we
use AI to try to help us with that. But
AI can only model what it had learned, right, So
for rare diseases conditions that there isn't enough information out
there on what the you know, the locks look like,
(25:37):
it's really hard to then model it with AI and
get an accurate prediction of whether there's a fit or not.
Quantum does not rely on the previous data for its modeling.
Quantum does its modeling purely based on the physical characteristics
of the compound of the key you know that you're designing.
(26:00):
So that makes it ideal because it's then untethered with
It's not limited by do we have enough samples or
don't we have enough samples, or you know, what the
previous studies find or not. It's purely based on those
physical properties, those quantum properties. So we found ourselves in
situations where we were able to predict the fits between
(26:26):
certain targets and where in conditions like Alzheimer's disease. For example,
we published where the quantum based modeling of the compound
was much better than what we would have gotten with
what we got right, Like, we did it both ways,
(26:46):
and the quantum one was the better fit than the
AI generated one.
Speaker 1 (26:51):
This is fair. Quantum is a little more of an
artist and a little less of a of a kind
of nerds nerdy to me, what seems like creative and artistic?
Speaker 2 (27:03):
Creative?
Speaker 1 (27:04):
Yes, creative, it's like people.
Speaker 3 (27:07):
Yeah, it's it opens up pass that you never knew existed.
That's why when I get asked about what do I
see is the best, you know, the the most important
breakthrough that quantum is going to allow us to do,
my answer is I really don't know, because we I
(27:28):
you know, when other people invented these new technologies, I
don't think they really knew that they're you know, like
think of laser. I don't think the person who invented
laser thought that they will be used to scan groceries
at the grocery store.
Speaker 2 (27:44):
You know.
Speaker 3 (27:45):
But so technology developing technology, the way I think of
it is like having a baby you know, you raise
it as best you can, but then they're going to
go off and do their thing, and you will be
tying them down if you restrict them to just what
you thought they should do, you know.
Speaker 2 (28:04):
So it opens up that.
Speaker 3 (28:09):
Space, that creative space for us to ask questions differently
than we used to. We should train our mind to
stop starting from classical and then trying to squeeze it
into quantum. We have to learn how to think quantum
up front, right from the beginning, which we haven't really
(28:33):
been doing as a scientific community since our inception. We
were trained, and how do you convert what you're thinking
into a formula that you can ask from a computer
which is a classical right computer with quantum because of
how it works, it can answer questions, It can look
(28:54):
at problems in a very different way. So we have
to think differently about the questions and how we ask them.
Speaker 1 (29:02):
With IBM, you recently modeled protein with over twelve thousand atoms.
Talk to me about that and why it's so meaningful
for job discovery.
Speaker 3 (29:12):
So in October of twenty twenty four, so just eighteen
months ago, the largest compound biological compound that could be
simulated with quantum was ten atoms big, and it was
unfathomable back then that we will get past the thousand
(29:35):
or few thousand atom simulation in the foreseeable future. And
what our team with IBM and with Rieken and Japan
published last month April twenty twenty six is a simulation
of the electrical properties of an enzyme trips in the
(30:00):
body that is twelve thousand, six hundred atoms bait for
you know, orders of magnitude beyond what anybody thought was
possible in that span of time. And the reason why
that happened was because of a you know, innovations in
(30:21):
the technology itself where the teams stopped thinking of quantum
and AI as competitors and instead thought of them as
different members.
Speaker 2 (30:35):
Of the same team.
Speaker 3 (30:37):
Right, you're we're now in basketball season in the US
the NPA, and you need the you need defense, but
you also need your center, somebody to shoot.
Speaker 5 (30:49):
You need all of the pieces to work together. So
with this, it was figuring out where do I put
you know, when do I put AI on the field,
When do I put Quantum on the field, and how
do I tell them.
Speaker 2 (31:04):
To work together?
Speaker 3 (31:06):
It's the scientific terms the quantum centric super computing. So
quantum is at the center, but we're using our supercomputing
tools AI classical to interact with it and split that
big problem of the twelve thousand, six hundred atoms into pieces,
(31:27):
where some pieces are best served with quantum and others
are best served with classically.
Speaker 1 (31:33):
This distinction that we now cling to AI and quantum
are these very different things develop by different people for
different purposes. What you're suggesting is that's probably going to
go away. Yeah, in the future, these things will work together.
It's going to become teamwork and not one on one
competition exactly.
Speaker 2 (31:54):
And it is that now in Keeveland Clinic.
Speaker 3 (31:56):
I mean, the way we've evolved our priority is with IBM.
It's a realization that both organizations have come to where
really too for progress to happen, we should stop seeing
them as separate. We we should put them together and
work to the best of what each piece of technology
(32:19):
can provide.
Speaker 1 (32:21):
One theme running through a lot of your your what
you've been talking about is that the arrival of this
new technology requires the researcher to behave differently and that's
one of the reasons why you want the quantum machine
(32:42):
and the on full display, and then you were talking
about how you have to ask different kinds of questions.
I'm curious, can you can you? Can you elaborate on
that a little bit? So take me back, for example,
to your earliest days. If I had given you all
these tools that people have now, how would your research
(33:02):
have proceeded differently? What would you have done differently?
Speaker 2 (33:07):
You know, I would have looked at the molecular.
Speaker 3 (33:17):
Components of the human brain as they relate to outcomes
of brain surgery way earlier than I did. The first
ten years of my career doing research was all spent
building models that were purely based on brain waves and
pictures of the brain. It wasn't until after I hit
(33:43):
a wall with my models aren't getting past that eighty
percent accuracy threshold that I started thinking, oh, you know,
there must be something genetic or you know, more biological
that is influencing this. Had I been exposed to quantum
(34:03):
at least as a concept, right to quantum computing and
what quantum science is back then, I think it would
have opened up my mind to realize that it's not
just about what I see. You know, there is hidden
relationships that exist within the human body.
Speaker 2 (34:26):
And that's our genetic.
Speaker 3 (34:27):
Makeup and our chemical makeup that influence what comes to
the surface that urge to dig deeper. The other thing
it would have changed is it would have made me
reach out to engineers and physicists and mathematicians much earlier
in my career.
Speaker 1 (34:48):
Yeah, yeah, yeah. And how would it have changed is
a very kind of prosaic question, but just kind of
the day to day life of someone doing better research.
I mean, the oh wow, you walk into the office
in the morning, How does your day proceed differently when
you're when you have these kinds of tools at your fingertips.
Speaker 3 (35:12):
That's the fundamental question in biomedical research right now, and
it's it's less to do with quantum, more to do
with urgentic AI, right these agents that we can work
with now to help us how to think more creatively
and how to do work that up until now was
(35:36):
more like Scott work that the researchers had to do,
whether it was you know, so the hypothesis generation has
always been the most creative part of aspect of scientific research.
But what comes after it with the data collection, for example,
that was always such a repetitive, you know, exercise, and
(36:01):
then the analysis after that was something that was very
resource intensive and you had to try so many different
approaches before you get to an answer, and that was
fairly complex.
Speaker 2 (36:16):
Now, with access to.
Speaker 3 (36:20):
Urgantic AI and you know, some quantum, of course, we
can spend more of our energy on the creative thinking
part of the aspects of the work and less on
the you know, just that repetitive.
Speaker 1 (36:39):
When you look around, I'm assuming you walk around Cleveland
Clinic and you have lots of conversations with some of
the most brilliant medical researchers in the world. Are you
satisfied with how quickly and aggressively they are adopting these
two technologies or do they still need encouragement from you
to do? You have to say people, wait, I've got
(37:02):
this machine in the cafeteria. You should be using it
for this problem. How much are you doing that kind
of encouraging in cheerleading or is it unnecessary?
Speaker 3 (37:11):
No, there's plenty of cheerleading that's necessary people. You know,
humans don't like to change. It's hardwired in us. So
there is plenty of cheerleading. But what happens is, I mean,
the way we built our program to where we are now,
(37:32):
so Cleveland clinic now is pretty much winning every global competition.
And Quantum for life sciences, whether that's the welcome trust,
you know, Quantum for biological applications. We our partner. There
was a startup in Finland, Algorithmic.
Speaker 2 (37:50):
They're brilliant.
Speaker 3 (37:51):
We worked with them to develop a photodynamic drug therapy
for cancer, or whether it is the NIH with they
had an expert price challenge for Quantum or universities that
we're collaborating with. We have a program that's bringing startups
to our ecosystem.
Speaker 2 (38:10):
We give them access to the.
Speaker 3 (38:11):
Machine if they have a question that is significant enough
for life sciences, universities that we're building a bachelor's, master's,
PhD programs with on quantum computing, we developed that whole
ecosystem around it. If that's not cheerleading, I don't know
(38:33):
what else would qualify for cheerleading. But then what happens
is these early adopters, the risk takers, become the cheerleaders themselves, right,
and then they work with it, they achieve success, and
we're nothing but competitive in medicine and science. Right, So
then it becomes okay, so and so I did this
(38:54):
with this machine, let me learn it. So I can
do the same, and it becomes this verse virtuous cycle
of then innovation and growth and people wanting to work together.
Speaker 2 (39:08):
It's just been fascinating to watch.
Speaker 1 (39:11):
You said that people don't like change, but you clearly do.
Speaker 2 (39:15):
I do. That's the mindset of researchers, right.
Speaker 3 (39:19):
A researcher is someone who is not afraid to fail,
actually sees failure as a chance to learn something right,
to do better the next time. So we get rejected
all the time and we don't care, right, we keep
moving on with papers, grant applications.
Speaker 2 (39:37):
So that mindset is what all of research is built around.
Speaker 3 (39:44):
So it's a very forward looking mindset and leveland clinic
as the health system, we wouldn't have survived one hundred years.
We wouldn't have done all of the firsts that we had.
Serotonin the chemical that drives the whole all science of
psychiatry and neuroscience that was discovered in Cleveland clinic. So
(40:07):
as an organization, we have enough of people who think
that way that they will be the early adopters who
will pull the others with them.
Speaker 1 (40:21):
Yeah. One last question, look look ahead tenures, pat me
a picture of what quantum and related technologies look like
for a place like Cleveland Clinic and tenures.
Speaker 3 (40:36):
Well, what I hope is that ten years from now,
if I if I'm seeing a patient in clinic who
has bad epilepsy and for the love of you know,
I can't figure out what medicine do I need to
prescribe to them to make them seizure free, I will
be able to send them to get a blood test
(40:58):
that we can then run through some analytical platform that
leverages both quantum and AI that can model exactly for
that patient, what compound, either existing or not to be developed.
It's some new chemical that we haven't tested for that
(41:21):
indication yet is going to treat them, you know, make
them seizure free. It's and I think quantum is is
ideal for a condition like mine epilepsy, because we are
a rare disease. And I think that benefit that it's
(41:41):
going to have will start with rare diseases, you know,
as I explained earlier, So take any other rare disease,
we should start there and then expand from that to
more complex things like cancer for example, and others. But
those rare conditions where we are left now completely scratching
(42:02):
our heads and going by intuition, it will make our
care truly precise, and it will make drug development a
more tailored exercise than the way it is now, where
it's like a hammer that's looking for nails?
Speaker 1 (42:22):
Am I right in thinking that of all of the
over the one hundred year history or more now of
Cleveland clinic, this sounds like the absolute best time to
be a Cleveland clinic.
Speaker 2 (42:31):
Yeah.
Speaker 3 (42:31):
I love it, no complaints, no complaints, And I am hiring.
Speaker 4 (42:39):
I need I need those quantum researchers, those people who
are wanting, you know, to apply quant genetic research, imaging research,
every single aspect of biomedical research.
Speaker 3 (42:54):
We can't afford to just wait on the sideline and
until the technolog she's ready, and then you know, then
it was teach us. It is ready now that twelve
thousand atom simulation wouldn't have happened just a year and
a half ago.
Speaker 2 (43:10):
It's quidinn.
Speaker 1 (43:12):
So the headline of this conversation is we're hiring, we hired.
Speaker 2 (43:15):
Yes, Yeah, that's a good headline. We're growing, how about that.
Speaker 1 (43:24):
Smart Talks with IBM is produced by Matt Ramano, Amy
Gains McQuaid, Trina Menino, and Jake Harper Engineering by Nina Bird,
Lawrence Mastering by Sarah Buguerer, music by Gramoscope, Strategy by
Cassidy Meyer, Sophia Derlon and Tatiana Lieberman. Special thanks to
doctor Laura jay Hi, Alicia Real Cooney, and the Cleveland
(43:45):
Clinic team. Smart Talks with IBM is a production of
Pushkin Industries and Ruby Studio at iHeartMedia. To find more
Pushkin podcasts, listen on the iHeartRadio app, Apple Podcasts, or
wherever you listen to. I'm Malcolm Gladwell. This is a
paid advertisement from IBM. The conversations on this podcast don't
(44:07):
necessarily represent IBM's positions, strategies, or opinions.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Hey Jonas!
Hey Jonas! The official Jonas Brothers podcast. Hosted by Kevin, Joe, and Nick Jonas. It’s the Jonas Brothers you know... musicians, actors, and well, yes, brothers. Now, they’re sharing another side of themselves in the playful, intimate, and irreverent way only they can. Spend time with the Jonas Brothers here and stay a little bit longer for deep conversations like never before.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by Audiochuck Media Company.
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.