October 2, 2023 • 33 mins
Imagine compressing a decade of research into mere weeks or turning a failure rate of 90% in clinical drug development on its head. That's exactly what's on the horizon with the incorporation of artificial intelligence (AI) in the realm of drug discovery, and the person leading this charge is none other than Dr. Max Jakobs, CEO and co-founder of DeepMirror. Prepare to be blown away as we navigate the potential AI holds for revolutionizing healthcare, making it possible to sift through massive datasets with unprecedented speed and accuracy.
In our exciting conversation with Dr. Jakobs, we dig into the gritty realities of drug discovery in its current state and the formidable issues researchers face without the aid of AI algorithms. Get ready to discover how AI not only hastens the drug-to-market journey but also boosts the clinical hit rates, saving ample time and a fortune in the process. Plus, we probe into the daunting challenge of FDA approval and how DeepMirror is working to dismantle the bias inherent in traditional drug development.
To conclude, we delve into the groundbreaking advancements in small molecule drug discovery and how cloud-based AI is honing laboratory processes. Dr. Jakobs illustrates how DeepMirror is utilizing AI to unearth the most promising drug candidates more efficiently. We bring to light riveting customer success stories and the promising future that lies ahead with AI-based data analysis in the pharmaceutical industry. Brace yourselves for a riveting exploration of AI's impact on drug discovery and healthcare.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
Actually 90% of all drugs fail.
So basically you're spendingfive years and then you have
another five years and 90% ofthe time will fail.
Speaker 2 (00:09):
In a world where 90% of clinical drug development
falters, companies invest adecade for the slim chance of
success.
We dive into this high stakesjourney with Dr Max Jacobs, CEO
and co-founder of DeepMirror.
Speaker 1 (00:28):
Because, on one end, you can use machine learning,
artificial intelligence, to takeyour past data on drug
compounds that you testedagainst the target and then
basically try to predict othercompounds that might even be
more potent.
Speaker 2 (00:42):
AI is compressing years of research into weeks,
tackling vast datasets andpromising to accelerate
pre-clinical work, but thechallenge is steep.
Studying clinical outcomesremains a formidable task.
Speaker 1 (00:59):
Individual drug.
You have to first make it, testit, and so on.
It might take two, three weeksor so.
So testing tens of thousands isreally not possible.
Speaker 2 (01:09):
The goal, Level the playing field, increase
competition and challenge norms.
With technologies from the1960s still prevalent, the
industry is ripe for disruption.
Speaker 3 (01:23):
The pacemaker has remained the same from 1960 to
now.
Largely, it seems like there isa certain level of bias that
sticks around that is very hardto think completely out of the
box.
Speaker 2 (01:38):
From small molecules to advanced therapies.
We're at the cusp of anexplosion of ideas and
innovation.
Speaker 1 (01:45):
The biggest challenge for any software company in
this place has always been toresist digging for gold almost.
Speaker 2 (01:55):
Join us on Things of Change podcast for a deep dive
into the future of drugdiscovery and healthcare
innovation.
If you had known how importantthe technology economy was 20
years ago, would you have donethings differently?
The internet, cell phones, thecloud and data Things have
(02:21):
changed.
We're here to talk about it, hi.
Speaker 3 (02:25):
I'm Jed, hi, I'm Shikhar.
Welcome to Things of Change,your new economics and
technology podcast.
Healthcare in general is hard.
Biotech, healthcare, lifesciences industries are one of
(02:46):
those industries where there'sjust a massive amount of data.
These are far too long Justbecause experimentation.
It costs a lot and you need tobe precise because ultimately,
the end users are the patientsand consumers are there.
Having technology come in andimprove and influence this level
(03:07):
of data that researchers,scientists and companies biotech
companies, small and largeoperate with is a big boon.
Your algorithms could be likeyour buddy that could just help
uncover some patterns that youwere not seeing before, or even
identify potential drug targets.
It seems like healthcare ingeneral is one of those
(03:31):
industries that could massivelybe influenced by the wave of AI
that we're seeing today.
Today, we're super excited tohave Dr Max Jacobs, co-founder
and CEO of DeepMirror, a healthtech company building software,
ai software platform that helpsresearchers accelerate
(03:54):
experimentation testing and, bydoing so, unlock creativity.
Think of chat, gpt or GitHubco-pilot, but co-pilot for your
research, like research GPT thatnames you to be trademarked.
Speaker 1 (04:08):
Thank you for having me.
We're excited to be here, guys.
Speaker 3 (04:12):
I touched on it a bit , but what are your thoughts as
to why this is needed inhealthcare in general?
Speaker 1 (04:18):
Yeah, I think it's very important to maybe
appreciate that in relationshipto other industries in health
tech and biotech and pharmathese data-driven approaches are
still very experimental.
Some of the larger companiesmight have some internal
research teams already that dothis kind of work, basically
(04:41):
providing algorithmic solutionsto other parts of the companies,
but it's really not yet in thehand of end users really so
these kind of data-driventechnologies.
It is really at a stage nowwhere people maybe until a few
years ago didn't really believethat there could be some sort of
vertical solution in theindustry, because everything is,
(05:03):
as you said, so complex, sodifficult and so specialized to
the particular application thatnothing can really help everyone
in a way.
But now we're seeing more andmore approaches and attempts to
actually do this, to bring thesekind of technologies that look
maybe very difficult into thehands of people.
(05:23):
I think you mentioned chatGPT.
That has been demonstrated byespecially these kind of
approaches.
Prior to that, people workedwith large language models
chatGPT is based on, but thesewere technical people.
It didn't really get to anysort of end users, but this kind
of leap of design almost madeit possible for the first time.
(05:44):
This is what we're trying toachieve as well, like having a
little leap of design and userexperience to finally bring
AI-powered algorithms to allthese labs that are not really
using them at this stage because, de facto, it's in the hand of
a select few companies in someway that have successfully
applied it.
But apart from that, mostcompanies are like basically
(06:08):
thinking oh, maybe I should getin on it, but I'm not sure how
yet.
Speaker 2 (06:13):
Yeah, one of the things we were really curious
about and you talked about it alittle bit was how does current
state look?
What are the kind of issuesthat researchers are facing
today with regards to coming upwith drug discovery, and all
these areas are really importantfor the farm and biotech
industry space.
Can you point out some reallyhigh-level stuff that's really
(06:33):
difficult for researchers todayto achieve without this
technology that we're talkingabout?
Speaker 1 (06:38):
So it is basically around the idea of using your
past data to try to somehow makebetter decisions, right?
So at the moment, it takesaround 10 years or so, and maybe
a billion or two billiondollars to take a drug to market
.
Wow, and is that the average?
(06:59):
That's the average.
Yeah, oh my.
Speaker 3 (07:02):
God, it's not a.
Speaker 1 (07:03):
Gaussian distribution , necessarily, but it's a lot of
money and a lot of time, andthis is because there's so many
things you have to do right.
Drug discovery goes frominitial research of trying to
figure out what can I attack totreat cancer, to then trying to
find agents like, for example,small drugs that might attack
this target and might diminishits activity, then trying to
(07:26):
make sure that these agents arevery potent, that they are safe
for people, that they can beingested by people, to then
finally putting them in theclinic and by that point you're
already like five years down,and then you have another five
years of clinical work and thenthis clinical time actually 90%
of all drugs fail.
So basically you're spendingfive years and then you have
(07:49):
another five years and 90% ofthe time will fail.
But this kind of influences thethinking of a company is quite
dramatically right.
And then anything that reducesthe time to get a gold shot,
which basically means shootinginto the clinic, right.
So actually trying clinic onceor improves the outcome in the
(08:10):
clinic is super, super valuable.
And this is what it converges toright, because on one end, you
can use machine learning andartificial intelligence to, for
example, take your past data ondrug compounds that you tested
against the target or somethinglike this, and then basically
try to predict other compoundsthat might even be more potent,
(08:32):
instead of trying all thesecompounds.
Right, so that obviously speedsyou up because you wouldn't
have to do so many experiments,because testing sometimes an
individual drug, you have tofirst make it, test it, and so
on.
It might take two, three weeksor so.
So testing tens of thousands isreally not possible, whereas
you can maybe test a thousandover a few years, and then you
(08:53):
really have to be smart whichones you test.
That's one way in which thesealgorithms can help.
The other way is, of course,once they go into the clinic,
you have very little control,right?
You do not know what is goingto happen.
However, many companies alreadygenerated some data on clinical
outcomes of drug compoundbecause they have tested quite a
(09:15):
few over the last hundred years.
Right, see, and this is anotherway you can actually improve
this you can try to predict whatwould happen in the clinic
given a certain drug, but therewould really have very small
data and also, in the other case, these data set are very small.
You really have to leverage alot of tricks from very nice
(09:36):
machine learning technologies tomake these things work and
actually pick up what happens inthe clinic is much harder than
just trying to predict stuffthat happens before the clinic.
So I think we're now this daywe can try to accelerate the
work before the clinic orpreclinical drug discovery work.
What the stage of trying toimprove clinical hit rates.
Speaker 3 (09:56):
That's still very difficult and the companies that
were successful also were moretowards the preclinical work
there's a direct link betweenthem spending all this time, all
this money and still havingjust At best ten percent chance
of getting a home run.
All that is still linked tohigher drug prices because they
(10:20):
need to recoup those cost thatthey put for ten years and on
top of actually reducing costsfor farmer.
Speaker 1 (10:27):
That itself wouldn't really help because it could
still charge whatever they want.
But if you increase competitionbetween companies as well by
actually enabling everyone toget to the same level of
capability, that's also key forthis and in a way, that is also
the thing we're trying toachieve a reliving the plane of
these kind of technologies.
Speaker 3 (10:47):
A large part of what's in the medical device
industry, farm industry andbiotech are Improvements or
iterations from what was createdin the nineteen sixties, of the
seventies.
And you smile over there.
I was so surprised because Iremember my first day at walking
with an abit and they were like, okay, this is not like rocket
(11:12):
science.
The pacemaker has remained thesame from nineteen sixty to now,
largely.
But I'd like to get your takeon that statement that I made.
It might not be completely true, but it seems like there is a
certain level of bias thatsticks around that is very hard
to think Completely out of thebox, because this is tried and
(11:36):
tested and has been used withthe public.
You have public health data forforty, fifty years.
So you're like, okay, you don'twant to deviate from that
normal.
Speaker 1 (11:45):
It's difficult to deviate from the normal if you
have to, in the end, get fdaapproval, and it's easier Get
approval for something that issimilar to something that is
known.
But that's, of course, only oneIssue.
I think this ties very wellinto what we're mainly doing
right now, which is so, firstway people build drugs.
(12:05):
So what's more?
Molecules, which is like tinymolecules that easily go into
cells.
They have been used.
There were the first drug thathave been used for the years and
farm company is developed quitea few of them over the last
century or so, and initiallythese were found mainly from
(12:26):
natural analog.
So people are just looking atplants trying to figure out what
the ingredient was that you,just you are the certain disease
of cured headache.
Then people start deviatingfrom that to be to try to find
analogs of these like slightchanges.
So you can think of this almost.
As this is huge desert ofmolecules, most of them are
toxic to people.
(12:46):
There's a few islands in therethat have promising molecules,
but the desert in between is bigand vast, so finding these
islands is extremely difficult.
That's why you Tend to often gofrom natural analogs right then
that was the status quo.
Then in the 80s and 90s peoplestarted doing more with high
(13:08):
content screening, so they wouldjust generate millions of
different compounds and justscreen them over some sort of
target that they wanted toattack.
That of course gave many moreresults, but often these
compounds in the end tend not tobe great in the clinic Because
the experiments that people didon these compounds were actually
(13:29):
not very predictive of whathappened later on in the clinic.
So it was a bit more difficult.
And then, especially now aftercovid, we see quite a lot of
development, in particularpeople called advanced therapies
, which are like therapies basednot on small molecules but
based on RNA, cell therapy,antibody therapy, peptide
(13:53):
therapy and so on.
So suddenly this whole zoo ofnew options that pops up and the
whole space became way morecomplicated.
And also people are going backto the small molecules again and
realize, oh shit, the ones thatwe actually work with we're
very similar with each other,like all the different ones that
pass through the FDA, but oftenvery slight modifications just
(14:16):
of the same molecule.
And then a company such as insilico, for example, come around
the corner and start designingmolecules that To see.
If this weird because theyhaven't really been done before,
but clearly they work.
So there is some, in some ways,you can harness, like these
generative approaches to come upwith new ideas, new solutions
(14:39):
to all the problems, and it's abit like a An explosion of ideas
that's happening right now, ofdifferent viewpoints you can
apply to all solutions again.
Speaker 3 (14:49):
So it's extremely exciting, can you provide
customer success stories orsomething where you are able to
just take some sample data andwas able to Improve an influence
there.
Speaker 1 (15:01):
research we did a lot of consulting work prior to
actually embarking on productdevelopment, and some of the
consulting work, for example,was image analysis, where we
essentially help peopleinterpret microscopy data In
early drug R&D.
And there it's straightforwardbecause typically it takes
(15:22):
people a few minutes to analyzean image where it takes an AI
working a few seconds.
So you're really talking aboutFor the fifties speed up here,
so that's very straightforward.
In in other ways, we did somework on optimizing molecules, as
I mentioned, and there, as youcan see, with some of the bigger
(15:43):
companies, you can speed thesethings up by what five to ten X
like you can find the mostoptimal drug candidate about
five to ten times faster if youuse some sort of AI powers
decision making and In turn andthis is also where we see our
initial product now, which wehave been developing last year,
(16:06):
because people have been workingwith small molecule in the dawn
of time, right, and this isalso the modality that most
people have experience with.
So we now pushing out or initialproduct exclusively for small
molecule drug discovery, wherewe essentially take customer
data.
Let's say somebody testedtwenty, fifty compounds in the
(16:29):
lab, then we would they uploadthe data to our app and then
they can predict the sameproperties that they measured
for these compounds for ten,twenty hundred thousand other
compounds.
Then prioritize the next step,take that back to the lab, get
the results back into the app,then close the design, may, test
and analyze cycle and this isreally what we're doing.
(16:53):
So it's this idea of laboratoryoptimization, by reducing the
amount of work you have to doand suggesting the work you
should do next, and there we dida few case studies, which you
can also find the no web page,where we think it can speed up
by about four, x and it'scurrent state.
But we're still working on somebenefits there as well.
(17:14):
Yes, really exciting, becauseyou go, you design a compound in
the app which is similar toyour compounds.
You potentially make it, thenyou test it, then you go back
again and design the next onebased on the information you
have before.
Yeah, instead of like sometimesstabbing a bit in the dark, you
can really use this to guideyour thinking and also,
(17:35):
sometimes something might comeup which you would not have
thought about.
We have molecule that maybe youhave never really synthesized
before, but make sense and thenyou might find something which
is maybe a new way, says thispast desert.
Speaker 3 (17:48):
Very interesting.
So the way you're describing it, it's like an application that
researchers can basically logonto on a browser, get into the
are, into your portal and inputtheir data set.
And is it something that theyget right away?
Or is it one of those thingswhere, okay, they input it and
(18:08):
then they head back and thenyou're doing the post processing
for a day or two and then theycome back, see all okay, what
these are the potential outcomes, and then test it out.
Is that the workflow thatyou're seeing?
Speaker 1 (18:20):
It has to be faster than that because otherwise,
okay, becomes more of aconsulting business again and we
really want to make theseinside available like almost
instantly, because in the endpeople do a lot right.
They might try this and theymight try this.
They want to play a bit aroundon the app.
(18:42):
So, yes, it's a cloud, simpleclouds, vertical SAS app you
would call it, and we initiallywhen an interface with which
people log in the upload a bitof data that they click a button
, predicts or other data thatthey might have provided
themselves, what generated usinggenerate a guy on our platform
(19:03):
and then it should take the pain.
On the data set size, anythingbetween four, what's a two
minutes to a few hours for thereally big ones.
But then nobody in early drugdiscovery has big data sets
because you would start withMaybe tens of compounds, that a
couple hundred compounds.
(19:23):
Once you have a thousandcompounds, you either failed or
one in the clinic already, butthat you don't need a I am
anymore because you already hadyour shot.
Speaker 3 (19:34):
That's really cool the people who are ultimately
their researchers.
Speaker 1 (19:37):
And on top of the game, they come into your
platform knowing maybe one ofthese hundred might work, so
they've already narrowed theproblem space and then finding
the oasis in that Problem spaceis a lot quicker and probably
gets more signal yes, sobasically, initially it's just
(19:58):
about finding something thatmight what we just call the hit,
which is like maybe one, two,three compounds, so, and then,
once you have a hit, try togenerate a few compounds that
are like in this molecule space,to just have a look, and then
at that point we, like alreadyhave narrowed into the space a
(20:18):
little bit.
I really want to look into thatregion, which could still be
massive right, and then generatea few in that area.
Then it's really where I was,so where would come in?
because it would basically takethe initial Results and then go
from that to generate newsuggestions.
That might make a lot of sense,because going really from zero
(20:41):
to something here Is it would bereally interesting.
But in terms of what ML AI iscapable these days it's really
not that straightforward becausethere's basically infinite
molecules you can make right andthen trying to do that.
We have a grant now pending.
(21:02):
That tries to do something likethis is a bit more, let's say,
high risk, so let's see if itworks out.
Speaker 3 (21:08):
But yeah, maybe we'll know next year's like a
baseline, and then you walkthrough it and you paper it.
Now you can only do a.
Speaker 1 (21:16):
I am that's awesome once you know something and
where you get that somethingfrom is really key.
Speaker 2 (21:22):
Does it get better as you onboard more customers onto
the platform, or are thesecompanies like super touchy
about their data sets that youcan never use everybody else's
data sets to establish baselinesfor new customers?
Speaker 1 (21:37):
Let's depend.
some customers are reallyinterested in that, because
particularly academic customersare very interested in making
data more widely available sothat sometimes even ask Whether
this is an, and of course it canbe done because our database
can distinguish between, say, hiLee, it's the tiny secret, just
(21:59):
part of the database where veryhappy, happy and I could never
be used for anything but servingto a customer.
But other customers can also.
We call centers Multiplayer.
They can engage in multiplayermode, which basically means that
they provide some data and thenthat data might be used For
other people as well, which isnot so interesting when you try
(22:23):
to make drugs against theparticular target, because it
really tends to work ondifferent things anyways.
But it becomes reallyinteresting for the things I
mentioned before the idea oftrying to predict what happens
in the clinic, because if youhave, if you have a few
compounds and you know whetherthese metabolize well or whether
(22:43):
these are toxic, group thatwith many other drugs of which
you know how they metabolize orhow toxic they are, that's
really cool and there's not thatmuch data on that right now.
If people are willing to sharethat, that's something people
(23:09):
often willing to engage with.
It also helps the problem,because typically people do not
want to share the structure of acompound because that's IP,
because as soon as that becomespublic knowledge, it can be
patented anymore.
But if you just say, all we canuse that model to build better
models for everyone, but yourstructure will never be
retrieval from that model,that's really something some
(23:30):
people are interested in so, max, the biggest news, at least in
the US, is ozampic and thoseweight loss drugs.
Speaker 3 (23:38):
It's like the biggest news.
Everyone's always hang on, sonow I can.
It's a really cool thingbecause you lose the weight and
then it's easier to control itthen.
But it's interesting that itstems from diabetes drugs that
we've used all this one.
When I was reading through yourblogs and the papers, it just
(24:00):
feels hey, hang on.
Something like this where youcould expand what a drug could
do or just think of it as a newway, like a new paradigm through
an AI, would be a lot fasterthan 60 years of trial over the
world's population.
So that seems like somethingthat we might just start
(24:23):
uncovering, where we usetraditional drugs and running
the co-pilot and we're like hangon, this could also do this and
go from there.
Speaker 1 (24:31):
Yeah, Things like we are doing is, let's say, you
have a bit of information,initially like again 20, 50
compounds and so which arecompletely new and have never
been done before, but they'renot really that potent.
They're quite toxic, so they'requite annoying, but you can use
the information from them tobasically go and screen public
(24:51):
databases.
There's millions of compoundsout there that you can just buy,
so you can then basically usethis information to try to
repurpose one of them.
For example, I don't know,these 50 compounds are really
good against killing thisparticular cancer, but they are
also doing all kinds of otherbad stuff.
So maybe I use this informationfrom these compounds to then
(25:14):
screen other compounds whichthen flag up and maybe these
compounds have never been usedagainst this disease, and then
you can start just repurposingthem because they're already FDA
proof.
So that's pretty cool andhistory is littered with these
kind of, let's say, superversatile compounds, right?
Doxycycline is one of them.
(25:34):
I tend to discover a newapplication of that every month
by chance.
Then Viagra is actually one ofthe most famous ones, because
Viagra was initially.
I can't remember what theclinical trial was about Hard,
hard.
Speaker 2 (25:49):
It was the hard.
Speaker 3 (25:50):
Yeah.
Speaker 1 (25:51):
There's also so many drugs that kind of failed
clinical trials, but they werealready shown to be safe for
patients and we don't even knowif they can do anything yet.
So that's also then really sadand often.
These then sometimes becomeavailable and you can maybe ask
a farmer company if you can lookinto them on top of this vast
(26:11):
chemical space and say wehaven't yet explored.
We still don't really fullyunderstand how, first of all,
some of the drugs we useactually work and then, second
of all, how we can actuallyrepurpose them to do even
completely different things,right, because we already know
that they're safe.
Right, so they can do so manythings potentially.
It was really exciting.
(26:32):
We have medicine.
We only scratch the surfaceuntil now, and now we like
starting developing all thesenew therapies, all these
different approaches.
Cell and gene therapies arereally awesome, especially for
like cancer.
Speaker 3 (26:44):
Yeah, yeah, it's super exciting.
Speaker 2 (26:46):
One of the things we wanted to ask was as you
envision the future of deepmirror and the problems that
you're solving today.
You look five, 10 years downthe line.
How do you see that?
What are the challenges thatyou foresee right now that
you're solving, and how you'llget to the future state that you
want?
Speaker 1 (27:06):
The biggest challenge for any, let's say, software
company in this space has alwaysbeen to resist digging for gold
, almost.
Because once you have somethingthat kind of works, everybody
will try to tell you oh, now youhave to go all the end to the
patient.
But then you tend to put allyour money onto a single
(27:27):
component at some point and thenagain 90% failed and then you
might have your cool platformsoftware which sped up your
preclinical work.
Let's say it's five, x orsomething like that.
But that still doesn'tnecessarily mean that afterwards
you will have an higher successrate necessarily.
But people still think in thisold school biotech way where you
(27:52):
basically go and okay, so youplay around a bit, you find
something, and then you put allyour hosts and then you become a
proper biotech company thatthen gets bought by a farmer or
your partner with some farmercompanies and so on, and staying
clear of that like trying toresist this digging and really
trying to stay a serviceprovider and delivering value to
(28:12):
everyone.
That will be quite tricky, butwe see other companies in this
space that kind of manage tobecome a fully vertical Sastau
solution.
Benchling is maybe one of themost famous one.
Nobody thought that somethinglike Benchling would work when
they came around becauseeffectively it was like a
workflow solution forresearchers, and why would
(28:34):
researchers pay for that?
But then over time peoplerealized researchers go to
farmer companies, have bigpockets and suddenly Benchling
has contracts that are more thana million and annual revenue.
Speaker 3 (28:45):
You said something that is so interesting.
I see the promise of thistechnology.
I'm thinking, oh, massiveamounts of data.
You put AI there.
We will be able to uncovercertain patterns.
But you mentioned something sointeresting where sometimes it
might just be helpful to get acleaner set of data, a smaller
(29:06):
set of data, and then expandfrom there.
So that is really interesting Ihave not thought of that before
and how AI can help with.
Speaker 1 (29:15):
Sometimes the less is more You're also defining the
question very well for thatparticular data set.
Like in our early days, we oncegot contacted by a company that
essentially asked us oh, wehave these millions of images
here.
Can we pipe those through analgorithm and learn something?
And then again, initially youthink, OK, maybe, but then you
(29:41):
look into this and then, as longas you don't really ask a
proper question about the data,you just end up with nothing,
Like if you don't ask a questionyou don't know.
you won't find an answer becauseeverything will be
statistically significant, andthen you might be amplifying
noise as well, because it's abig data set and you can't just
have a look at it, and you mightnot learn anything.
(30:04):
It's like this whole idea oflooking at each grain of sand on
a beach and then trying tounderstand what a beach is.
It doesn't really help.
You have to have some sort ofquestion about the data you
might analyze.
There's something similarhappening in academia now, where
people were thinking a lotabout something called spatial
transcriptomics, which isbasically the idea of taking
(30:25):
images of tissues and then alsolooking at genetic information
in these images and basicallygenerating not even terabytes
but petabytes of data on this.
And still they don't reallyknow how to apply this, because
just by looking and collectingdata without narrowing the
question, we don't know.
This is where we come in, rightthat people still are very much
(30:47):
required to define a question,define the hypothesis and then
help get a machine to help withthe number crunching.
Right Can't replace it yet.
Speaker 3 (30:57):
Yeah, max this was really great.
Before you leave us, we alwayslike to give our guests the mic.
You already have a mic, but wewant to pass the virtual mic to
you so that you can give a shoutout to your team, the work that
you're doing and where peoplecan reach you, because we have a
few OVCs and founders wholisten to us and they're always
(31:18):
interested to hear what peopleare building and how they can
reach them.
Speaker 1 (31:22):
Yeah, people can reach me straight on the max at
deepmirrorai, and without theteam, nothing of this would have
been possible.
I'm quite fortunate because Ihave two co-founders One is more
on the technical side, one ismore on the product side.
I think the synergy between thetwo of them is pretty much
amazing.
We also over the last yearhired two more people who help
(31:47):
us more with customer serviceand things, like Cecilia and
Jacob, who's like basicallytaking charge of all the machine
learning, and potentially inthe next year we'll grow the
team to 10 people.
So I think we might be lookingto hire soon.
That's very exciting and at themoment we're very much
bootstrapped for our consultancycontract.
(32:09):
We have a bit of investment.
We might be looking for somelater this year to re-scale up
the product Nice.
But, yeah, that's all.
Very.
Sometimes I would say we're thepre-seed just looking to find
the product with which we'll goon the seed rocket.
But you really have to.
In such a complex space, as wementioned, you really have to
watch out that the thing you'rebuilding is actually something
(32:31):
useful.
Speaker 2 (32:33):
I wanted to say thank you for coming on the show.
It was really nice to meet youand hopefully we have you on
again.
Check in a couple of years' time.
See how deep mirrors do weleave you with thought.
The future of thepharmaceutical industry is on
the brink of transformation.
From AI-powered drug discoveryto the emergence of novel
(32:56):
therapies, the landscape isshifting rapidly.
These innovations usher in anew era of health care.
Only time will tell.
Stay tuned, stay informed andjoin us again to explore how
things are changing around theworld.
Until next time, stay curious.
Actually 90% of all drugs fail.
So basically you're spendingfive years and then you have
another five years and 90% ofthe time will fail.
Speaker 2 (00:09):
In a world where 90% of clinical drug development
falters, companies invest adecade for the slim chance of
success.
We dive into this high stakesjourney with Dr Max Jacobs, CEO
and co-founder of DeepMirror.
Speaker 1 (00:28):
Because, on one end, you can use machine learning,
artificial intelligence, to takeyour past data on drug
compounds that you testedagainst the target and then
basically try to predict othercompounds that might even be
more potent.
Speaker 2 (00:42):
AI is compressing years of research into weeks,
tackling vast datasets andpromising to accelerate
pre-clinical work, but thechallenge is steep.
Studying clinical outcomesremains a formidable task.
Speaker 1 (00:59):
Individual drug.
You have to first make it, testit, and so on.
It might take two, three weeksor so.
So testing tens of thousands isreally not possible.
Speaker 2 (01:09):
The goal, Level the playing field, increase
competition and challenge norms.
With technologies from the1960s still prevalent, the
industry is ripe for disruption.
Speaker 3 (01:23):
The pacemaker has remained the same from 1960 to
now.
Largely, it seems like there isa certain level of bias that
sticks around that is very hardto think completely out of the
box.
Speaker 2 (01:38):
From small molecules to advanced therapies.
We're at the cusp of anexplosion of ideas and
innovation.
Speaker 1 (01:45):
The biggest challenge for any software company in
this place has always been toresist digging for gold almost.
Speaker 2 (01:55):
Join us on Things of Change podcast for a deep dive
into the future of drugdiscovery and healthcare
innovation.
If you had known how importantthe technology economy was 20
years ago, would you have donethings differently?
The internet, cell phones, thecloud and data Things have
(02:21):
changed.
We're here to talk about it, hi.
Speaker 3 (02:25):
I'm Jed, hi, I'm Shikhar.
Welcome to Things of Change,your new economics and
technology podcast.
Healthcare in general is hard.
Biotech, healthcare, lifesciences industries are one of
(02:46):
those industries where there'sjust a massive amount of data.
These are far too long Justbecause experimentation.
It costs a lot and you need tobe precise because ultimately,
the end users are the patientsand consumers are there.
Having technology come in andimprove and influence this level
(03:07):
of data that researchers,scientists and companies biotech
companies, small and largeoperate with is a big boon.
Your algorithms could be likeyour buddy that could just help
uncover some patterns that youwere not seeing before, or even
identify potential drug targets.
It seems like healthcare ingeneral is one of those
(03:31):
industries that could massivelybe influenced by the wave of AI
that we're seeing today.
Today, we're super excited tohave Dr Max Jacobs, co-founder
and CEO of DeepMirror, a healthtech company building software,
ai software platform that helpsresearchers accelerate
(03:54):
experimentation testing and, bydoing so, unlock creativity.
Think of chat, gpt or GitHubco-pilot, but co-pilot for your
research, like research GPT thatnames you to be trademarked.
Speaker 1 (04:08):
Thank you for having me.
We're excited to be here, guys.
Speaker 3 (04:12):
I touched on it a bit , but what are your thoughts as
to why this is needed inhealthcare in general?
Speaker 1 (04:18):
Yeah, I think it's very important to maybe
appreciate that in relationshipto other industries in health
tech and biotech and pharmathese data-driven approaches are
still very experimental.
Some of the larger companiesmight have some internal
research teams already that dothis kind of work, basically
(04:41):
providing algorithmic solutionsto other parts of the companies,
but it's really not yet in thehand of end users really so
these kind of data-driventechnologies.
It is really at a stage nowwhere people maybe until a few
years ago didn't really believethat there could be some sort of
vertical solution in theindustry, because everything is,
(05:03):
as you said, so complex, sodifficult and so specialized to
the particular application thatnothing can really help everyone
in a way.
But now we're seeing more andmore approaches and attempts to
actually do this, to bring thesekind of technologies that look
maybe very difficult into thehands of people.
(05:23):
I think you mentioned chatGPT.
That has been demonstrated byespecially these kind of
approaches.
Prior to that, people workedwith large language models
chatGPT is based on, but thesewere technical people.
It didn't really get to anysort of end users, but this kind
of leap of design almost madeit possible for the first time.
(05:44):
This is what we're trying toachieve as well, like having a
little leap of design and userexperience to finally bring
AI-powered algorithms to allthese labs that are not really
using them at this stage because, de facto, it's in the hand of
a select few companies in someway that have successfully
applied it.
But apart from that, mostcompanies are like basically
(06:08):
thinking oh, maybe I should getin on it, but I'm not sure how
yet.
Speaker 2 (06:13):
Yeah, one of the things we were really curious
about and you talked about it alittle bit was how does current
state look?
What are the kind of issuesthat researchers are facing
today with regards to coming upwith drug discovery, and all
these areas are really importantfor the farm and biotech
industry space.
Can you point out some reallyhigh-level stuff that's really
(06:33):
difficult for researchers todayto achieve without this
technology that we're talkingabout?
Speaker 1 (06:38):
So it is basically around the idea of using your
past data to try to somehow makebetter decisions, right?
So at the moment, it takesaround 10 years or so, and maybe
a billion or two billiondollars to take a drug to market
.
Wow, and is that the average?
(06:59):
That's the average.
Yeah, oh my.
Speaker 3 (07:02):
God, it's not a.
Speaker 1 (07:03):
Gaussian distribution , necessarily, but it's a lot of
money and a lot of time, andthis is because there's so many
things you have to do right.
Drug discovery goes frominitial research of trying to
figure out what can I attack totreat cancer, to then trying to
find agents like, for example,small drugs that might attack
this target and might diminishits activity, then trying to
(07:26):
make sure that these agents arevery potent, that they are safe
for people, that they can beingested by people, to then
finally putting them in theclinic and by that point you're
already like five years down,and then you have another five
years of clinical work and thenthis clinical time actually 90%
of all drugs fail.
So basically you're spendingfive years and then you have
(07:49):
another five years and 90% ofthe time will fail.
But this kind of influences thethinking of a company is quite
dramatically right.
And then anything that reducesthe time to get a gold shot,
which basically means shootinginto the clinic, right.
So actually trying clinic onceor improves the outcome in the
(08:10):
clinic is super, super valuable.
And this is what it converges toright, because on one end, you
can use machine learning andartificial intelligence to, for
example, take your past data ondrug compounds that you tested
against the target or somethinglike this, and then basically
try to predict other compoundsthat might even be more potent,
(08:32):
instead of trying all thesecompounds.
Right, so that obviously speedsyou up because you wouldn't
have to do so many experiments,because testing sometimes an
individual drug, you have tofirst make it, test it, and so
on.
It might take two, three weeksor so.
So testing tens of thousands isreally not possible, whereas
you can maybe test a thousandover a few years, and then you
(08:53):
really have to be smart whichones you test.
That's one way in which thesealgorithms can help.
The other way is, of course,once they go into the clinic,
you have very little control,right?
You do not know what is goingto happen.
However, many companies alreadygenerated some data on clinical
outcomes of drug compoundbecause they have tested quite a
(09:15):
few over the last hundred years.
Right, see, and this is anotherway you can actually improve
this you can try to predict whatwould happen in the clinic
given a certain drug, but therewould really have very small
data and also, in the other case, these data set are very small.
You really have to leverage alot of tricks from very nice
(09:36):
machine learning technologies tomake these things work and
actually pick up what happens inthe clinic is much harder than
just trying to predict stuffthat happens before the clinic.
So I think we're now this daywe can try to accelerate the
work before the clinic orpreclinical drug discovery work.
What the stage of trying toimprove clinical hit rates.
Speaker 3 (09:56):
That's still very difficult and the companies that
were successful also were moretowards the preclinical work
there's a direct link betweenthem spending all this time, all
this money and still havingjust At best ten percent chance
of getting a home run.
All that is still linked tohigher drug prices because they
(10:20):
need to recoup those cost thatthey put for ten years and on
top of actually reducing costsfor farmer.
Speaker 1 (10:27):
That itself wouldn't really help because it could
still charge whatever they want.
But if you increase competitionbetween companies as well by
actually enabling everyone toget to the same level of
capability, that's also key forthis and in a way, that is also
the thing we're trying toachieve a reliving the plane of
these kind of technologies.
Speaker 3 (10:47):
A large part of what's in the medical device
industry, farm industry andbiotech are Improvements or
iterations from what was createdin the nineteen sixties, of the
seventies.
And you smile over there.
I was so surprised because Iremember my first day at walking
with an abit and they were like, okay, this is not like rocket
(11:12):
science.
The pacemaker has remained thesame from nineteen sixty to now,
largely.
But I'd like to get your takeon that statement that I made.
It might not be completely true, but it seems like there is a
certain level of bias thatsticks around that is very hard
to think Completely out of thebox, because this is tried and
(11:36):
tested and has been used withthe public.
You have public health data forforty, fifty years.
So you're like, okay, you don'twant to deviate from that
normal.
Speaker 1 (11:45):
It's difficult to deviate from the normal if you
have to, in the end, get fdaapproval, and it's easier Get
approval for something that issimilar to something that is
known.
But that's, of course, only oneIssue.
I think this ties very wellinto what we're mainly doing
right now, which is so, firstway people build drugs.
(12:05):
So what's more?
Molecules, which is like tinymolecules that easily go into
cells.
They have been used.
There were the first drug thathave been used for the years and
farm company is developed quitea few of them over the last
century or so, and initiallythese were found mainly from
(12:26):
natural analog.
So people are just looking atplants trying to figure out what
the ingredient was that you,just you are the certain disease
of cured headache.
Then people start deviatingfrom that to be to try to find
analogs of these like slightchanges.
So you can think of this almost.
As this is huge desert ofmolecules, most of them are
toxic to people.
(12:46):
There's a few islands in therethat have promising molecules,
but the desert in between is bigand vast, so finding these
islands is extremely difficult.
That's why you Tend to often gofrom natural analogs right then
that was the status quo.
Then in the 80s and 90s peoplestarted doing more with high
(13:08):
content screening, so they wouldjust generate millions of
different compounds and justscreen them over some sort of
target that they wanted toattack.
That of course gave many moreresults, but often these
compounds in the end tend not tobe great in the clinic Because
the experiments that people didon these compounds were actually
(13:29):
not very predictive of whathappened later on in the clinic.
So it was a bit more difficult.
And then, especially now aftercovid, we see quite a lot of
development, in particularpeople called advanced therapies
, which are like therapies basednot on small molecules but
based on RNA, cell therapy,antibody therapy, peptide
(13:53):
therapy and so on.
So suddenly this whole zoo ofnew options that pops up and the
whole space became way morecomplicated.
And also people are going backto the small molecules again and
realize, oh shit, the ones thatwe actually work with we're
very similar with each other,like all the different ones that
pass through the FDA, but oftenvery slight modifications just
(14:16):
of the same molecule.
And then a company such as insilico, for example, come around
the corner and start designingmolecules that To see.
If this weird because theyhaven't really been done before,
but clearly they work.
So there is some, in some ways,you can harness, like these
generative approaches to come upwith new ideas, new solutions
(14:39):
to all the problems, and it's abit like a An explosion of ideas
that's happening right now, ofdifferent viewpoints you can
apply to all solutions again.
Speaker 3 (14:49):
So it's extremely exciting, can you provide
customer success stories orsomething where you are able to
just take some sample data andwas able to Improve an influence
there.
Speaker 1 (15:01):
research we did a lot of consulting work prior to
actually embarking on productdevelopment, and some of the
consulting work, for example,was image analysis, where we
essentially help peopleinterpret microscopy data In
early drug R&D.
And there it's straightforwardbecause typically it takes
(15:22):
people a few minutes to analyzean image where it takes an AI
working a few seconds.
So you're really talking aboutFor the fifties speed up here,
so that's very straightforward.
In in other ways, we did somework on optimizing molecules, as
I mentioned, and there, as youcan see, with some of the bigger
(15:43):
companies, you can speed thesethings up by what five to ten X
like you can find the mostoptimal drug candidate about
five to ten times faster if youuse some sort of AI powers
decision making and In turn andthis is also where we see our
initial product now, which wehave been developing last year,
(16:06):
because people have been workingwith small molecule in the dawn
of time, right, and this isalso the modality that most
people have experience with.
So we now pushing out or initialproduct exclusively for small
molecule drug discovery, wherewe essentially take customer
data.
Let's say somebody testedtwenty, fifty compounds in the
(16:29):
lab, then we would they uploadthe data to our app and then
they can predict the sameproperties that they measured
for these compounds for ten,twenty hundred thousand other
compounds.
Then prioritize the next step,take that back to the lab, get
the results back into the app,then close the design, may, test
and analyze cycle and this isreally what we're doing.
(16:53):
So it's this idea of laboratoryoptimization, by reducing the
amount of work you have to doand suggesting the work you
should do next, and there we dida few case studies, which you
can also find the no web page,where we think it can speed up
by about four, x and it'scurrent state.
But we're still working on somebenefits there as well.
(17:14):
Yes, really exciting, becauseyou go, you design a compound in
the app which is similar toyour compounds.
You potentially make it, thenyou test it, then you go back
again and design the next onebased on the information you
have before.
Yeah, instead of like sometimesstabbing a bit in the dark, you
can really use this to guideyour thinking and also,
(17:35):
sometimes something might comeup which you would not have
thought about.
We have molecule that maybe youhave never really synthesized
before, but make sense and thenyou might find something which
is maybe a new way, says thispast desert.
Speaker 3 (17:48):
Very interesting.
So the way you're describing it, it's like an application that
researchers can basically logonto on a browser, get into the
are, into your portal and inputtheir data set.
And is it something that theyget right away?
Or is it one of those thingswhere, okay, they input it and
(18:08):
then they head back and thenyou're doing the post processing
for a day or two and then theycome back, see all okay, what
these are the potential outcomes, and then test it out.
Is that the workflow thatyou're seeing?
Speaker 1 (18:20):
It has to be faster than that because otherwise,
okay, becomes more of aconsulting business again and we
really want to make theseinside available like almost
instantly, because in the endpeople do a lot right.
They might try this and theymight try this.
They want to play a bit aroundon the app.
(18:42):
So, yes, it's a cloud, simpleclouds, vertical SAS app you
would call it, and we initiallywhen an interface with which
people log in the upload a bitof data that they click a button
, predicts or other data thatthey might have provided
themselves, what generated usinggenerate a guy on our platform
(19:03):
and then it should take the pain.
On the data set size, anythingbetween four, what's a two
minutes to a few hours for thereally big ones.
But then nobody in early drugdiscovery has big data sets
because you would start withMaybe tens of compounds, that a
couple hundred compounds.
(19:23):
Once you have a thousandcompounds, you either failed or
one in the clinic already, butthat you don't need a I am
anymore because you already hadyour shot.
Speaker 3 (19:34):
That's really cool the people who are ultimately
their researchers.
Speaker 1 (19:37):
And on top of the game, they come into your
platform knowing maybe one ofthese hundred might work, so
they've already narrowed theproblem space and then finding
the oasis in that Problem spaceis a lot quicker and probably
gets more signal yes, sobasically, initially it's just
(19:58):
about finding something thatmight what we just call the hit,
which is like maybe one, two,three compounds, so, and then,
once you have a hit, try togenerate a few compounds that
are like in this molecule space,to just have a look, and then
at that point we, like alreadyhave narrowed into the space a
(20:18):
little bit.
I really want to look into thatregion, which could still be
massive right, and then generatea few in that area.
Then it's really where I was,so where would come in?
because it would basically takethe initial Results and then go
from that to generate newsuggestions.
That might make a lot of sense,because going really from zero
(20:41):
to something here Is it would bereally interesting.
But in terms of what ML AI iscapable these days it's really
not that straightforward becausethere's basically infinite
molecules you can make right andthen trying to do that.
We have a grant now pending.
(21:02):
That tries to do something likethis is a bit more, let's say,
high risk, so let's see if itworks out.
Speaker 3 (21:08):
But yeah, maybe we'll know next year's like a
baseline, and then you walkthrough it and you paper it.
Now you can only do a.
Speaker 1 (21:16):
I am that's awesome once you know something and
where you get that somethingfrom is really key.
Speaker 2 (21:22):
Does it get better as you onboard more customers onto
the platform, or are thesecompanies like super touchy
about their data sets that youcan never use everybody else's
data sets to establish baselinesfor new customers?
Speaker 1 (21:37):
Let's depend.
some customers are reallyinterested in that, because
particularly academic customersare very interested in making
data more widely available sothat sometimes even ask Whether
this is an, and of course it canbe done because our database
can distinguish between, say, hiLee, it's the tiny secret, just
(21:59):
part of the database where veryhappy, happy and I could never
be used for anything but servingto a customer.
But other customers can also.
We call centers Multiplayer.
They can engage in multiplayermode, which basically means that
they provide some data and thenthat data might be used For
other people as well, which isnot so interesting when you try
(22:23):
to make drugs against theparticular target, because it
really tends to work ondifferent things anyways.
But it becomes reallyinteresting for the things I
mentioned before the idea oftrying to predict what happens
in the clinic, because if youhave, if you have a few
compounds and you know whetherthese metabolize well or whether
(22:43):
these are toxic, group thatwith many other drugs of which
you know how they metabolize orhow toxic they are, that's
really cool and there's not thatmuch data on that right now.
If people are willing to sharethat, that's something people
(23:09):
often willing to engage with.
It also helps the problem,because typically people do not
want to share the structure of acompound because that's IP,
because as soon as that becomespublic knowledge, it can be
patented anymore.
But if you just say, all we canuse that model to build better
models for everyone, but yourstructure will never be
retrieval from that model,that's really something some
(23:30):
people are interested in so, max, the biggest news, at least in
the US, is ozampic and thoseweight loss drugs.
Speaker 3 (23:38):
It's like the biggest news.
Everyone's always hang on, sonow I can.
It's a really cool thingbecause you lose the weight and
then it's easier to control itthen.
But it's interesting that itstems from diabetes drugs that
we've used all this one.
When I was reading through yourblogs and the papers, it just
(24:00):
feels hey, hang on.
Something like this where youcould expand what a drug could
do or just think of it as a newway, like a new paradigm through
an AI, would be a lot fasterthan 60 years of trial over the
world's population.
So that seems like somethingthat we might just start
(24:23):
uncovering, where we usetraditional drugs and running
the co-pilot and we're like hangon, this could also do this and
go from there.
Speaker 1 (24:31):
Yeah, Things like we are doing is, let's say, you
have a bit of information,initially like again 20, 50
compounds and so which arecompletely new and have never
been done before, but they'renot really that potent.
They're quite toxic, so they'requite annoying, but you can use
the information from them tobasically go and screen public
(24:51):
databases.
There's millions of compoundsout there that you can just buy,
so you can then basically usethis information to try to
repurpose one of them.
For example, I don't know,these 50 compounds are really
good against killing thisparticular cancer, but they are
also doing all kinds of otherbad stuff.
So maybe I use this informationfrom these compounds to then
(25:14):
screen other compounds whichthen flag up and maybe these
compounds have never been usedagainst this disease, and then
you can start just repurposingthem because they're already FDA
proof.
So that's pretty cool andhistory is littered with these
kind of, let's say, superversatile compounds, right?
Doxycycline is one of them.
(25:34):
I tend to discover a newapplication of that every month
by chance.
Then Viagra is actually one ofthe most famous ones, because
Viagra was initially.
I can't remember what theclinical trial was about Hard,
hard.
Speaker 2 (25:49):
It was the hard.
Speaker 3 (25:50):
Yeah.
Speaker 1 (25:51):
There's also so many drugs that kind of failed
clinical trials, but they werealready shown to be safe for
patients and we don't even knowif they can do anything yet.
So that's also then really sadand often.
These then sometimes becomeavailable and you can maybe ask
a farmer company if you can lookinto them on top of this vast
(26:11):
chemical space and say wehaven't yet explored.
We still don't really fullyunderstand how, first of all,
some of the drugs we useactually work and then, second
of all, how we can actuallyrepurpose them to do even
completely different things,right, because we already know
that they're safe.
Right, so they can do so manythings potentially.
It was really exciting.
(26:32):
We have medicine.
We only scratch the surfaceuntil now, and now we like
starting developing all thesenew therapies, all these
different approaches.
Cell and gene therapies arereally awesome, especially for
like cancer.
Speaker 3 (26:44):
Yeah, yeah, it's super exciting.
Speaker 2 (26:46):
One of the things we wanted to ask was as you
envision the future of deepmirror and the problems that
you're solving today.
You look five, 10 years downthe line.
How do you see that?
What are the challenges thatyou foresee right now that
you're solving, and how you'llget to the future state that you
want?
Speaker 1 (27:06):
The biggest challenge for any, let's say, software
company in this space has alwaysbeen to resist digging for gold
, almost.
Because once you have somethingthat kind of works, everybody
will try to tell you oh, now youhave to go all the end to the
patient.
But then you tend to put allyour money onto a single
(27:27):
component at some point and thenagain 90% failed and then you
might have your cool platformsoftware which sped up your
preclinical work.
Let's say it's five, x orsomething like that.
But that still doesn'tnecessarily mean that afterwards
you will have an higher successrate necessarily.
But people still think in thisold school biotech way where you
(27:52):
basically go and okay, so youplay around a bit, you find
something, and then you put allyour hosts and then you become a
proper biotech company thatthen gets bought by a farmer or
your partner with some farmercompanies and so on, and staying
clear of that like trying toresist this digging and really
trying to stay a serviceprovider and delivering value to
(28:12):
everyone.
That will be quite tricky, butwe see other companies in this
space that kind of manage tobecome a fully vertical Sastau
solution.
Benchling is maybe one of themost famous one.
Nobody thought that somethinglike Benchling would work when
they came around becauseeffectively it was like a
workflow solution forresearchers, and why would
(28:34):
researchers pay for that?
But then over time peoplerealized researchers go to
farmer companies, have bigpockets and suddenly Benchling
has contracts that are more thana million and annual revenue.
Speaker 3 (28:45):
You said something that is so interesting.
I see the promise of thistechnology.
I'm thinking, oh, massiveamounts of data.
You put AI there.
We will be able to uncovercertain patterns.
But you mentioned something sointeresting where sometimes it
might just be helpful to get acleaner set of data, a smaller
(29:06):
set of data, and then expandfrom there.
So that is really interesting Ihave not thought of that before
and how AI can help with.
Speaker 1 (29:15):
Sometimes the less is more You're also defining the
question very well for thatparticular data set.
Like in our early days, we oncegot contacted by a company that
essentially asked us oh, wehave these millions of images
here.
Can we pipe those through analgorithm and learn something?
And then again, initially youthink, OK, maybe, but then you
(29:41):
look into this and then, as longas you don't really ask a
proper question about the data,you just end up with nothing,
Like if you don't ask a questionyou don't know.
you won't find an answer becauseeverything will be
statistically significant, andthen you might be amplifying
noise as well, because it's abig data set and you can't just
have a look at it, and you mightnot learn anything.
(30:04):
It's like this whole idea oflooking at each grain of sand on
a beach and then trying tounderstand what a beach is.
It doesn't really help.
You have to have some sort ofquestion about the data you
might analyze.
There's something similarhappening in academia now, where
people were thinking a lotabout something called spatial
transcriptomics, which isbasically the idea of taking
(30:25):
images of tissues and then alsolooking at genetic information
in these images and basicallygenerating not even terabytes
but petabytes of data on this.
And still they don't reallyknow how to apply this, because
just by looking and collectingdata without narrowing the
question, we don't know.
This is where we come in, rightthat people still are very much
(30:47):
required to define a question,define the hypothesis and then
help get a machine to help withthe number crunching.
Right Can't replace it yet.
Speaker 3 (30:57):
Yeah, max this was really great.
Before you leave us, we alwayslike to give our guests the mic.
You already have a mic, but wewant to pass the virtual mic to
you so that you can give a shoutout to your team, the work that
you're doing and where peoplecan reach you, because we have a
few OVCs and founders wholisten to us and they're always
(31:18):
interested to hear what peopleare building and how they can
reach them.
Speaker 1 (31:22):
Yeah, people can reach me straight on the max at
deepmirrorai, and without theteam, nothing of this would have
been possible.
I'm quite fortunate because Ihave two co-founders One is more
on the technical side, one ismore on the product side.
I think the synergy between thetwo of them is pretty much
amazing.
We also over the last yearhired two more people who help
(31:47):
us more with customer serviceand things, like Cecilia and
Jacob, who's like basicallytaking charge of all the machine
learning, and potentially inthe next year we'll grow the
team to 10 people.
So I think we might be lookingto hire soon.
That's very exciting and at themoment we're very much
bootstrapped for our consultancycontract.
(32:09):
We have a bit of investment.
We might be looking for somelater this year to re-scale up
the product Nice.
But, yeah, that's all.
Very.
Sometimes I would say we're thepre-seed just looking to find
the product with which we'll goon the seed rocket.
But you really have to.
In such a complex space, as wementioned, you really have to
watch out that the thing you'rebuilding is actually something
(32:31):
useful.
Speaker 2 (32:33):
I wanted to say thank you for coming on the show.
It was really nice to meet youand hopefully we have you on
again.
Check in a couple of years' time.
See how deep mirrors do weleave you with thought.
The future of thepharmaceutical industry is on
the brink of transformation.
From AI-powered drug discoveryto the emergence of novel
(32:56):
therapies, the landscape isshifting rapidly.
These innovations usher in anew era of health care.
Only time will tell.
Stay tuned, stay informed andjoin us again to explore how
things are changing around theworld.
Until next time, stay curious.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.
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.