ACM ByteCast - Mary Lou Jepsen - Episode 67
Episode Date: April 3, 2025In this episode of ACM ByteCast, our special guest host Scott Hanselman (of The Hanselminutes Podcast) welcomes Mary Lou Jepsen, CEO and Founder of Openwater, a technical executive and inventor in the... fields of display, imaging, and computer hardware with about 300 patents published or issued to her name. She founded and led two moonshots at Google X and was later an executive at Facebook/Oculus VR. Prior to this, she was a professor at the MIT Media Lab, where she co-founded and was the first CTO of One Laptop per Child (OLPC), and later founded Pixel Qi in Taipei, Taiwan, focused on the design and manufacture of displays. Jepsen has been named as one of the 100 most influential people in the world by Time Magazine (“Time 100”), CNN’s top “10 thinkers” in science and technology and has won numerous awards numerous from professional societies in the fields of optics, display, and electronics. She’s a frequent keynote speaker, has given two highly viewed TED talks, and is frequently featured in top global press publications. Mary Lou discusses her work with Openwater, a startup working on innovative imaging technology using infrared light, ultrasound, and electromagnetics to diagnose and potentially treat diseases, and aims to leapfrog traditional drug development. She and Scott talk about the role of patents in manufacturing, and regulatory and technological barriers in healthcare innovation. They also dive into the advantages of the company’s open-source model, both for its software and hardware designs. Mary Lou highlights some of their breakthroughs, including stroke detection and non-invasive cancer treatment. She also talks about reducing cost and scaling production, next steps in clinical trials, and future possibilities with open source.
Transcript
Discussion (0)
This is ACM Bytecast, a podcast series from the Association for Computing Machinery, the
world's largest education and scientific computing society.
We talk to researchers, practitioners, and innovators who are at the intersection of
computing research and practice.
They share their experiences, the lessons they've learned, and their own visions for
the future of computing.
I'm your host today, Scott Hanselman.
Hi, I'm Scott Hanselman,
and this is another episode of Hanselman.
It's in association with the ACM Bytecast.
Today, I have the distinct pleasure of chatting with
Dr. Mary Lou Jepson, CEO and founder of OpenWater. How are you?
I'm great. How are you?
Thanks for having me today.
I'm just fantastic. I just love talking to
cool people that are doing cool stuff.
I tell you though, I get a little intimidated because when you read up, you do all your
research and 250 patents and 50 plus products.
And it's like, do you have more hours than we do?
Or are you 300 years old?
How are you getting all this fun work done?
You know, we all work.
I've looked at your resume too.
It's pretty astonishing all the work that you've done.
So what I work in is a little bit different.
The patents are really useful when you're designing mixed signal and analog things because
you use contract manufacturing and it's sort of a fig leaf that helps you indemnify the
contract manufacturer so you can ship.
And it's just very helpful to have a lot of patents.
And when you use contract manufacturing and not build your own factory, you can
ship a lot more product more quickly.
So that's the secret.
Interesting.
Cause I think that the lay person who may only have one or two patents, actually
I have a patent, I didn't even think about this until we started talking.
I have a patent on the Microsoft band.
If you remember that little device that would be on the, you know, failed
at Microsoft Apple watch, which had sensors
and such that were trying to give us information about how we moved.
But that patent just kind of just happened.
It was one of those things.
And we always think as a lay person about how, why do people make patents?
And you're right.
In the course of your daily life, it would be important for you to get patents in order
to indemnify the manufacturers and things like that.
It helps us get indemnification because we can show it's unique because the claims are like,
the claims are the things that make the valuable that you can write like how it works and then you
write these claims and combining this with this in this circumstance for this purpose you own.
So that's very helpful in being able to ship stuff using contract manufacturing.
Yeah. So in the healthcare space, I am not a person that works in this space, but I'm a fan.
And I'm a fan because I've been diabetic for over 32 years and I use an open source artificial
pancreas. So I've been attached to an open APS now for almost 10 years. So this is off-label,
not FDA approved, very kind of suspicious type stuff. And we're always promised non-invasive.
Diabetics, for the 30 years I've been diabetic, have been promised non-invasive solutions. And I,
when I give my talks on it, always say, you have to poke the meat bag if you want to know what's
going on inside the meat bag.
But it seems to me like open water medical technology does not believe that.
You think we don't have to necessarily poke this giant meat bag that we are to get good
data out of this body.
By poking the meat, do you mean cut or?
Inject, poke, push in non, you know, like I have multiple holes in me every day in order
to get to the interstitial fluid underneath the skin or push in insulin.
But Apple has been promising me, you know, with light, they'll tell my blood sugar or
with light, they can tell my pulse ox and with light, they can do this.
And I don't think that we understand do we or do we not have to poke the thing in order
to get good data?
I think we're missing a lot.
Certainly people have been working on blood sugar measurements and blood pressure measurements
and there is progress.
The issue I come to is $658 million in 13 years. That is the average amount of time and capitalized cost
it takes to get a regulatory approval
to say this really works
for something that's novel and therapeutic
as averaged in a comprehensive paper,
which I can share a link for you for.
But that's the problem.
And what that also means is you're building 10 units a year.
You don't get to use Moore's law.
You don't get to use contract manufacturing.
They're not interested in building 10 units a year.
They make the consumer electronics of the world.
And so, yeah, you can use AI, you can use supercomputers, but if you actually
want to get the data in different ways, you've
got to get through those proofs.
And so I love the OpenPancreas project because it is open source, but it's sort of limited
on a certain organ.
I wish they'd sort of expand to more organs because then you get to make custom chips
that can send light and ultrasound and electromagnetics into your body and then read it coming out.
And we get to use the ever shrinking transistor size to get ever shrinking resolution of rather
than just measuring the intensity of light or sound or electromagnetics, we can modulate
the phase and the interaction of these light and sounds
in different ways.
And so that's why I left my cushy position
at what was then called Facebook,
running a large aspect of advanced virtual reality
and augmented reality, and seeing what our checkbook
could fund in collaboration with the work
of the mighty Microsoft and my former
employer Google and certainly what Apple was doing and Samsung. I could see this stuff was
coming down the pike and no one wanted to touch the fact that maybe we could leapfrog drugs and
be able to not just even diagnose but also treat hundreds of diseases that kill us. Not just we've got
great results on glioblastoma, but the implications of the results are for all cancers. We're
moving mice into remission of glioblastoma. It's a big deal. It's a 100% deadly form of
human brain cancer. The implications for all cancers are pretty profound. We've got great
results on mental disease, severe depression
in humans. We're now doing work with amyloid microclots, which are the hallmark of acute
COVID and type 2 diabetes and neurodegenerative disease and aging itself. So the implications
for using light, infrared light that penetrates our body, ultrasound that penetrates our body and electromagnetic
at diagnostic levels to make general purpose machines
that can be configured, not the software layer.
So we're trying to just kind of make like Android
where the apps are regulatory approvals
that our customers get.
And with that, we can make scale.
So we are making custom lasers,
custom ultrasound transducers, custom chips, all in a
custom hardware. So we've started by doing this. We started the company nine years ago next month
by building out rooms with lots of equipment to see what worked using phantom tissue, going to buy
kidneys at the Chinese, over near Chinatown in San Francisco,
so we get a lot of range of dead meat to try this on.
And then moving into carts and got into hospitals
for about four years, starting at the beginning of pandemic.
And now we've shrunk it down again to make a system
that we can ship right now for $10,000 a piece,
but the cost of it at volume can be the cost of a smartphone.
It's fully open source, like real open source, not whatever.
It's AGPL, Creative Commons 4.0, share alike.
It's all of that.
Open source is a distribution model.
It's a trust model.
And it's an innovation model.
And the health care industry doesn't mean to be, but the healthcare industry is anti-innovation.
When it takes 30 to 40 years to ship a new product, you just got to call a spade a spade.
It's a lot of well-meaning people.
No.
Well, that very much resonates with me because, again, I've been in diabetes for 30 years.
When I was diagnosed 32 years ago, they said, I think we'll get this thing kicked in five years.
And that's the joke, is that every five,
every year they tell you five more years.
And then you start, after 30 years,
you start getting a tinfoil hat.
You say, well, it's cheaper to keep me diabetic
than it is to fix this thing.
Well, right.
There's a lot of stories about that,
particularly with the, oh, the kidney,
the transfusion stuff.
Yeah, yeah. There's a great book written
about it recently about the racket of, I want to remember the two names, Friesenius and Davita,
where they do apparently do that according to a lot of the lawsuits. Yeah, cash cow. You're a cash cow of the industry. I am too. I had
brain surgery in 1995. Hardest thing I've ever done. It took 17 years to diagnose that
I've taken a dozen medications every day for the last 29 years to stay alive.
Wow. You can really relate and you're thinking yourself-
Unwittingly become a neuroendocrinology expert. Maybe this work can help for that, but actually maybe having treatments for all cancers
and all mental diseases and addictions
is faster and easier.
So you gotta sort of say, wow, let's pursue that.
And so that's really what we decided to pursue.
Oh, and by the way, the number two killer in the world,
stroke, we have blood flow level detection that's about 20 times better than multimillion dollar
MRI.
And that really matters for diseases of blood flow.
And the number two killer in the world is stroke.
That's a blood flow issue.
It either stops or it explodes.
And when you can see blood flow, actually blood flow precisely, that's a big deal.
So we've been in hospitals for three years studying stroke patients of the most severe
type of stroke called large vessel occlusion stroke.
When a large vessel is occluded, it blocks everything downstream.
So, you know, if you don't get the right procedure within two hours, if you live,
you probably won't either walk or talk again.
You certainly won't have a job again.
You probably won't go home again.
My dad had two TIAs last week and they're ocular TIAs.
And we're trying to figure out is that the harbinger of the big one?
And 30% of the time it is.
Yeah.
The repeat strokes.
And so if you had a way to know if you were having an LVL stroke, you could go
to the only 5%
of the hospitals.
This is the worst one, the big killer.
It kills 6.5 million people a year globally, number two cause of death globally.
5% of the hospitals can do the treatment.
So by chance, you have 5% chance of getting there because it's not apparent what you have.
The gunshot victims, we know what they have.
They get to go to the right treatment.
Even heart attacks, because there's an ECG you can put on your chest and find out if
you've got a heart attack or not.
So we've basically made a version of an ECG for your forehead that tells you if you're
having an LVO stroke, it could be used in a portable in an ambulance to direct the ambulance
to the right hospital and notify the hospital.
So in parallel, they're preparing the cath lab. The solution is to string a catheter up your
carotid artery and pull the clot out because it's a plumbing problem. If you can get that
done within two hours of stroke onset, there's a 90% chance of no neural deficit at all.
Time is brain cells.
But this is also useful potentially for TIAs.
Those are transient isochemic attacks.
What we're seeing in amyloid microclots is,
there's a theory it's clogging the capillaries,
which are the smallest diameter vessel in your body.
This is happening in acute COVID and long COVID,
as well as diabetes and neurodegenerative
disease and aging.
And what we've shown is using ultrasound as well.
And I can tell you the frequencies, we're open source.
We're using 150 kilohertz, 10% duty cycle, and we're breaking apart those microclots.
So there's sub five microns in size, which means the capillary at its thinnest width is 5 to
10 microns.
So the theory is they should be able to get through.
We're in pre-clinical with it, but we're sharing that we're just writing up our first
results right now.
We also think we should be able to see those microclots, which may give rise to TIAs.
It's not fully understood TIAs, but we can make these better non-invasive tools that
help us have better understanding of it and so forth.
You had a round last year of $100 million.
How do you raise money and also give away things as open source?
I have found that that is not necessarily compatible.
I made the case, to be fair, $54 million last year.
Oh, my apologies.
But $100 million over the life of the company.
In total funding, yeah.
In total funding.
That it's a better business model than $658 million in 13 years for a single rare disease.
That if you get it to work, then you still have to get through standard of care approval
and reimbursement.
Minimum, a billion dollars. Okay,
it's single rare disease. Let's take glioblastoma. I happen to know about 5,000 people in the US
have glioblastoma right now. A new 5,000 people will have it next year because these ones will
be dead. 5,000 people, how many have health insurance? Half? Maybe?
Maybe.
And how many will get denied? Let's call it 1,000, 2,000. Through 2,000, you've spent a
billion dollars to get there. What do you charge? Well, most people charge us close to a million
dollars. There's new treatments right now that are at $3 million, some of the new CRISPR treatments. So it shouldn't take me to say it's an understatement
to say that a million dollar treatment is not affordable to the vast, vast majority of humanity,
nor Americans. So why are we funding R&D that if it works, or companies that if it works,
it's going to be a million dollar treatment. Why? So the argument
of this general purpose device, and it's not we stand on the shoulders of giants. We've done some
really innovative things. The hardware design is really amazing. We're lowering the cost of the
system. We've lowered it from a million dollars to $10,000. It'll go down to another thousand dollars.
But there's been a million papers published in the last 20 years alone about using infrared
light or ultrasound or electromagnetics or in combination to treat thousands of diseases.
People get their PhDs.
People get tenure.
Maybe they get into the national, maybe they become an IEEE, I'm sorry, ACM fellow, whatever.
But it's a rounding error to say zero of it gets into the healthcare system because no one can afford the $658 million and you don't get to shrink the cost of the device first because
the FDA considers a quality bill of 10 units. I was a CTO, a group CTO at Intel in 2021 years ago.
A sample size was 10,000 units for us, a sample size.
There's nothing quality build about 10 units.
And so you build these big carts for a million dollars
and you get a regulatory approval, you get reimbursement
and you get to build 12 carts then
and have them in 12 hospitals.
Like it's just, until you can spend another $658 million
shrink it and get it to production.
So why not just say, look, we have enough data
to demonstrate that this is viable?
Because if you look at it,
if you make more of something, it's cheaper.
It's a slight exaggeration to say, if you make you make more of something, it's cheaper. It's a slight exaggeration to say
if you make 10X more of something, it's 10X cheaper,
but it's only a slight exaggeration.
So you make a lot more, you collect a lot of data.
We're making everybody that buys our devices
and you can make your own if you want to.
Good luck with the laser.
It's really hard.
So it's the ultrasound.
It's really hard to do this mixed signal sort of stuff.
So, but the plans are there. You can do it or you can buy this. And if it keeps
us honest, if you get approval, we're ISO 1345 approved, you can just, we're IRB ready,
FDA ready, all that stuff for doing clinical trials. But you can also just take the design
and have it built wherever you want to. So why is that a good business model?
Because we still ship a lot more of them. We're a lot more profitable.
We'll ship a few thousand units this year. We'll go profitable.
Just even getting it into the R&D community, a fully reconfigurable,
low-intensity focused ultrasound system that's wearable and small,
as well as this blood flow detection unit that's wearable and small,
as well as this blood flow detection unit that's also useful for any movement.
We're seeing seizures and we think kidney flow and lymph flow and anything you want.
You can change the pulse duration of the laser.
We make basically a hologram on camera chips shipping in every smartphone in the world.
Okay, so I want to understand,
you said shrink it down a couple of times.
MRIs, CTs, big giant donuts that spin,
these are giant million dollar machines that are scary.
Then we have these pocket supercomputers.
Help me understand why medical devices need to be so big,
and why you're deciding that no,
they really don't need to be so big. Well, you need a big magnet to see the, I mean, it's one of the most complicated products
I know ever shipped.
It saved my life.
Size and cost hasn't changed in the 29 years since.
I wasn't scared of it.
People are.
But yeah, you have to lie inside of the magnet and the magnet is a very specific thing. It's actually,
we can get into the physics of it. But yes, it hasn't changed. This is my frustration as well.
And yet, look at what we can make in the silicon fabs and the semiconductor fabs within
a variety of materials. I started life as a teenager falling in love with making holograms.
You can now make holograms on these camera chips that are shipping in smartphones that have high quantum efficiency in the near infrared.
The near infrared, that's the thing that goes through your body, like the invention of the fire was really important to humanity.
You see the light, but you feel the heat.
That's the infrared light warming your belly like it goes through,
but it scatters everywhere.
You can unscatter it if you can record the phase of the light.
And the 5 megapixel camera pixels shipping in most all smartphones
have pixel sizes of a micron.
That's the wavelength of light.
You can then interfere if you can make a highly coherent laser
that allows you to get that information out because the phase of
light with the intensity of light is fundamentally a lot more information and
What we do is we pulse this laser that we made and when you pulse it
For anybody in the audience that ever made a hologram, you know if anything moves
When you're doing your exposure you get a bad hologram. It's either
a dim one or nothing comes out. So if we pulse the laser and something's moving, we get a
dim hologram. And if nothing's moving, we get a really bright hologram. And the gradations
that we can see give us blood flow in phantom tissue to 0.1% accuracy, which is 20 times better than anything else
we can find recorded, including CT, MRI, Doppler ultrasound,
and so forth.
Using literally a $1 camera chip in your smartphones
and a laser that was a million dollars in the store
five years ago, but we've reduced it down to something
the size of the first digit of my finger in a butterfly package.
And so that it's a Mopa laser. We use a highly coherency laser and amplify it with a tapered
amplifier. It's a little bit more tricky than that because the pulsing, when you pulse something,
you change its heat and its current, and that can change its wavelength.
And so it's really tricky on how you build that out.
But we now have it, and it's starting production next quarter.
ACM Bytecast is available on Apple Podcasts, Google Podcasts,
Podbean, Spotify, Stitcher, and TuneIn.
If you're enjoying this episode, please do subscribe
and leave us a review on your favorite platform. That's amazing.
So using lay person's analogies here for the audience,
you're using light, you're using
broads frequencies for both read and write.
You are detecting things and then changing it.
You are reaching into the body and changing it.
It's like gamma knife radio surgery.
You're doing pinpoint accuracy,
targeting things in the body from the outside. Just like gamma knife radio surgery, you're doing pinpoint accuracy targeting things in
the body from the outside without having to cut.
But with no ionizing radiation.
So we're using ultrasound to selectively kill cancer cells at diagnostic doses shown safe
on pregnant women and their fetuses in rich countries for the last 50 years and tens of
billions of people for the last 100 years.
This is not like a new drug that's never existed before and is being injected into a body,
a bunch of bodies.
Nonetheless, the way we do it is we find the resonant frequency of the cancer cells.
For example, with glioblastoma, we got 16 different lines of glioblastoma, built up
these little organoids, little tiny brains, half a millimeter in size, and ran through
frequency sweeps and rhythm sweeps to find the ones that would kill the cancer, shake
apart the cancer, but not harm the healthy cells.
And when it shakes apart the cancer, it releases proteins that vaccinate the body from the
very cancer it has.
Whether you're an anti-vaxxer or not,
you're going to be for this. It's your own protein letting your immune system see it.
So we ran through that. In our studies, we did better than any of the chemo drugs. In the
autopsies, we then replicated that in mice and moved 30 mice into remission of glioblastoma. The ones that we didn't
give the treatment, we had to sacrifice. We sent them to autopsy. Child's River found no
damaged healthy cells in the autopsy. There's no cancer treatment that doesn't harm healthy cells.
Even surgery, you cut through, but chemotherapy, radiation therapy also harms the healthy cells. Even surgery you cut through, but chemotherapy, radiation therapy also harms
the healthy cells. So it's huge. Just find the resonant frequency of the cancer, get
a biopsy. We think it's probably going to end up, we're not sure where it's going to
end up. People want us to biopsy. We're probably going to start a thing with an organoid cell
so you can send your biopsy. we can find your frequency and you know,
in a trial. We are not FDA approved yet. So, of course, of course. So pushing gently on that,
if this was a Hollywood movie and I was the scientist and I had little blastoma, I would
jump into the box and I would shoot myself with the thing. How far away is this? Because they're my neighbor, my neighbor.
Literally next quarter, literally next quarter. And I have someone in London waiting for the
treatment. That's at 150 kilohertz. So what we're trying to do is span a whole bunch of ranges.
We're turning on and off neurons non invasively without the one inch hole in your head. I had
the one inch hole in my head. No fun. It was death or that. Of course I went for it,
but I don't see the vast swath of humanity
actually signing up for it.
Indeed they don't for deep brain stimulation
where you could get two to do it.
But non-invasively we can reach anywhere in the brain.
And in our first study of severe depression,
almost half of the patients went into remission with just 10-minute a
day doses for three weeks, five days the first week, three days the second week, three days
the third week, and they've stayed there.
So this could be an at-home system that you put on your head.
You know, honestly, we even found a way to register.
We can register precisely to where your thing is by not using a hundred thousand dollar system
in the hospital, but an Android smartphone
where we take pictures of your head,
make a mesh, a 3D mesh, register it on your scans,
match your bone structure.
When we take those pictures, you're wearing the headset.
We know exactly where those transducers are.
And with the transducers, we delay the phase from, we have these 8x8 arrays,
you delay the phase from one emitter to the next emitter so you can steer that sound
to precise focus near, far, up, down, right, left, anywhere in the body we wish to go.
So it's a very versatile system. We're making it in a tiled manner. You can buy a 1Z or 2Z or 6-pack
for your abs and 3D print your own shape so that you can experiment with whatever the ailment is.
Because the implications of focused ultrasound are so profound, we've seen really great results in high intensity focused ultrasound.
But that, like gamma-night, in the case of high, it ablates tissue by heating or even
vaporizing it. In this case, we're just looking for the resonant frequencies and with the
diagnostic level affecting selectively at a cellular level.
Do you think in some years,
this will be the kind of thing that someone could build in there?
I mean, I could certainly get a prescription for this and take it home,
but could I build this myself?
Isn't there risk in high-intensity focus ultrasound creating heat?
Well, we're doing low intensity.
Low intensity.
We're doing the diagnostic level,
achieving the therapeutic result.
Therapeutic results on low intensity.
We can ship it once the first volumes hit for about a thousand dollars a piece about the cost
of a smartphone and treatment would be like the cost of a phone call. So that really changes
things as you think of the shortages of drugs and using drugs. They just affect here, not your whole
body. It doesn't flow through your whole body.
There could be some like...
Do you ever feel like you're really close to a tricorder?
You know what I mean?
This is that Star Trek level of noninvasive diagnosis
and curative.
That's the goal.
We're starting with two modules of that,
but really when you look at the million papers
that have been published, there's a lot of talent out there.
So we hope to bring them along and get more modules in this unit or
compete to save the world, whatever.
But we think open source is the right approach because it gets the cost structure.
Right.
Like literally the wall street journal in December had a headline.
Are we justified in murdering healthcare CEOs?
Oh my goodness.
We all remember Guido there who shot the United Healthcare CEO dead on the streets of Manhattan.
Of course not, but people are frustrated. There really are death panels. It's called standard of
care. Can we move it forward fast enough? Why does so many people have to die? There are 55 million people dying every year globally.
Those bodies are stacking up and we all know those people.
Let's get going.
So open source does that.
And the ACM is perfect because like writing software,
testing this, like getting it out, it's the right way.
Otherwise it's just this, we're literally
making new therapies for the fraction of the top 1% that can afford a million dollar therapy.
It's insane.
Yeah.
Insurance.
Well, I mentioned my dad and his TIAs, we're literally waiting. We either pay $1,799,
$1,800 a month for the drug that he needs or we fight with Medicare.
Right. We go back and forth. And right now, we're in the middle of going back and forth and every day is $1,800 a month for the drug that he needs, or we fight with Medicare.
We go back and forth.
Right now, we're in the middle of going back and forth,
and every day is another day that we're wasting time.
Or we move.
Or we move. Yeah, exactly.
We move somewhere where we can get
another health care plan or
another country with another health care plan.
That's what I've done, just to stay alive.
I have health care in other countries.
It's funny that you mentioned that.
I've worked at Microsoft now for 17 years. I came here to do open source. That's funny that you mentioned that. I've been working at Microsoft now for 17 years.
I came here to do open source.
That's the narrative and that's what people say.
They know that I came to
Microsoft 17 years ago to do open source.
But the part that I don't usually talk publicly about is,
I also came because we have really good healthcare.
Right.
Microsoft was the cheapest place for me to get the insulin I need,
the insulin pumps, and the medication that I need.
At the time, it had a Cadillac level healthcare plan, and that tying healthcare to employment
is insanity. Right. After my brain tumor, I wasn't going to find my first startup.
I had dropped out of my PhD in physics. I was living in a wheelchair, sleeping 20 hours a day, super sick.
I got the surgery, finished my PhD, DARPA gave us $4 million and I said, I can't do
it, I need really good health insurance.
You know what my co-founder said?
You seem to know a lot about health insurance, why don't you figure that one out?
I'm like, oh yeah, I can do a startup, it doesn't have to be risky.
So everybody in all of my startups and their families get really good
healthcare because they're taking a risk being with a startup.
But yeah, American system has this and other systems have, you
know, drugs I take in other countries.
I literally have to get a script from the U S and go with a bag of money
to where it's made and get the drug that way, because it's illegal for adults.
Fine for children.
Something like a human growth hormone, for example. Yeah. I love that open source is at the heart of this though, that it's made and get the drug that way, because it's illegal for adults, fine for children. Something like a human growth hormone, for example.
Yeah, yeah, yeah.
I love that open source is at the heart of this though,
that it's about availability.
It has to be.
Yeah, it seems like it has to be,
because it's a scale thing.
Because as we get the approvals,
we want the costs to be affordable to change the outcomes.
Like, why is it novel for a healthcare company
its success to be attached to helping more people
with diseases rather than less?
Like the billionaires that funded me,
a lot of them would say, go for the single rare disease first
and that gets you, and maybe it's not a million dollars,
but it's more than $100,000 for the treatment then.
And the bankruptcies from healthcare
are staggering across the country.
It's not just not paying their credit card loans
or having avocado toast.
Like literally the bankruptcies are cost.
It's so staggering that when I talk to people
in other countries, they don't believe it.
Like, you know that the number one source of bankruptcy
in the United States is medical bills. And they're like, no, surely
no fully organized first world country would ever allow their people to die out of medical debt.
I mean, and go bankrupt, lose their homes is insane. Yeah.
And it's the reality. And so with open source, we can get more data. So we're asking everybody
to share the safety data because don't we want it to be safe? Shouldn't that be open?
Somebody dies from a Tylenol every day.
We don't hear about it because Tylenol is not a hundred percent safe.
So share that.
So then a regulator has that information, all the safety stuff, but also a doctor
and patient have the details of it and the efficacy data gets shared.
There is like, we will be a group that just shares all their efficacy data because
we have this other major tool of our time is really hard to miss, AI. Could it do more
with more data, understand more about disease states, treatments, etc.? Sure, if we create
that structure. There's a beneficial structure for everybody in this. I'm just surprised.
I can't understand why people are getting stuck.
We've come upon this general purpose thing
and we stand on the shoulder of these millions of papers.
Again, we've pushed it forward with this harmonic stuff
and blood flow.
We've got really great technology
that we've opened to everybody.
But a lot of people have really great technology
in healthcare and it just doesn't get.
I mean, the stuff I've seen in measuring glucose, there's a bunch of different approaches.
I'm very excited about some of the polarization approaches because with sugar, the light rotates.
You know, you can...
I think it's funny.
I think of polarization, rotation around glucose and light the same way I think about E-ink
displays and how E-ink displays rotate the ink to turn black
or to turn-
Sure, yeah, that's a great idea.
I remember getting a-
Yeah, there's so many different approaches
and I think people run out of gas.
They can get funded for their PhD or early research,
but then it's the companies that need to do those.
So we can take on part of that.
Others can take on part of that. Others can take on part of that.
The partnerships seem to be evolving.
I was talking to the NIH, they went a little blank.
The US government is, what did you say before?
Yeah, it's in some major transition for sure.
Indeed.
Nice this way I can put it.
But people are still dying
and what can we do to save them?
And how do we better engage the open source community?
I think we need the hardware out for them.
So that's coming out next quarter.
And we need to seed some hacker spaces
or some labs and so forth
to make sure people have access to do things.
I would love to help with that.
If I can work with you and get Microsoft
to give you office space
or we have these in called reactors
that you can have in different cities
where you're gonna get people together
and hack as a collective that would be super fun.
That would be great.
That'd be fantastic.
There is this issue, it's all ultrasound
after 1995 is considered class two
after the toothbrushes we all use were approved
and the beauty treatments for getting rid of wrinkles.
So as a result, to use it on people, there's some rules.
We like rules, it's okay, but they have to be followed.
Biohacking is legal, you just have to follow the local rules
on it, but if there's that, with that construct,
even doing it on anything with a backbone is illegal.
That's why you'll find in labs in
Boston they often use lobsters because they don't have a backbone. No offense. We've got crabs on
the West Coast. So there's things. But no, there's fast ways to do it. And there's IRBs. IRBs are
the ethics review panel to make sure anything you're doing with this is checked out, it's going to be saved, share the results, it's all very reasonable.
And even putting together what we're starting to do, there's different pockets where you
can find IRBs, but to help people create their own IRBs is a whole thing as well.
Like, because we want it to be safe, we want the data to be shared, but there's a huge
– so we really,
now that these are coming out,
now have to start to engage better
with the open source community.
So we would love to try the reactors
or anything else you might suggest as.
Great activities. Yeah, absolutely.
My team runs the open source programs office at Microsoft
and I'd love to chat with you.
And of course you probably know people at MSR
and I'm gonna have Peter Lee,
the head of Microsoft Research on soon.
So if you know Peter.
Right, lovely guy.
I got to end up talking about this stuff at Necker Island,
and he was there too.
That's cool.
Yeah, it was.
The speakers were on, it's called Mosquito Island,
but the mosquitoes were gone.
But yes, so I got to get to know him and his wife
very well for a week.
So in conclusion, if we can go to Open Water Health and we can learn about this, you can
click on open source business, you can learn about your patent pledge, you can learn about
how you use the AGPL and Creative Commons.
What is another recommendation and how either a hacker or a lay person can get involved
in support the work that you're doing at Open Water?
If you want a device, reach out to us and you can't afford the 10k, we might be able to pull it
and get it into a local space, a reactor like you suggest or whatever the hacker otherwise space is
so that maybe more people can have access to it. There's certain universities, but then the
universities sort of are already covered
for people outside of that who want to help.
I can see up on github.com slash openwaterhealth,
you've got the neuromodulation hardware,
the focused ultrasound toolbox in Python,
the blood flow gen two stuff.
You're continuing to release everything,
firmware, software and hardware up on GitHub.
And we will forever.
We changed our corporate rules to do that forever.
By the way, Italic Buterin, the founder of Ethereum,
wrote a $50 million non-deleted check
and we changed the structure to be open source forever.
He has veto authority if we ever change that. We had lawyers, I mean, lawyering this up, like how has everybody else gotten out of open source forever. He has veto authority if we ever change that.
We had lawyers, I mean, luring this up,
like how has everybody else gotten out of open source?
Open AI was getting out of it at the time.
And so we made the best minds in the legal field
put together the best construct they could
for no way for us to get out of this, like ever.
So we're open source always.
On open sourcing or closed sourcing anything,
that's excellent.
Right.
Fantastic.
Well, what a joy to chat with you and to learn about the great work that you're doing
over there at OpenWater.
Thank you so much for spending time with us today.
Thank you too.
I hope to work with you with the reactors and the Open Source and Microsoft community.
It'll be great.
I hope so.
We have been chatting with Dr. Mary Lou Jepson, CEO and founder of Open Water.
This has been another episode of Hansel and It's
in association with the ACM Bytecast.
And we will see you again next week.
ACM Bytecast is a production of the Association
for Computing Machinery's Practitioner Board.
To learn more about ACM and its activities, visit acm.org.
For more information about this and other episodes, please do visit our website at learning. acm.org. For more information about this and other episodes,
please do visit our website at learning.acm.org slash Bytcast.
That's B-Y-T-E-C-A-S-T.
Learning.acm.org slash Bytcast.