Science Corporation raises $230M after retinal implant helps blind patient finish a 300-page novel

might be asking a question hey my check engine lights on or how do I put this in track mode and then it knows that you actually got the GT3 RS with the YOC package. So who knows anyway I think we have resolved our issues. Uh we will return to our discussion with Max Hodak. How are you doing Max? Think I know.

Yes, I think so.

Let's uh let's let's let's restart with uh

start over from the beginning because I was I I could hardly hear you, but what I was hearing I was quite excited about.

Yes.

So, we make brain implants. Our main our main one is a retinal prostthesis. So, it's a chip implanted in the back of the eye. Finished a major clinical trial last last summer

in age related macular degeneration, which is one of the leading causes of blindness. And in the trial, we were able to take patients that like there's one patient that hasn't been able to recognize faces in over a decade. I found out that she finished a 300page novel recently.

Oh, that's insane.

Wait, so so uh couldn't recognize faces. That means that like if I'm trying to imagine what that looks like, it's effectively take a picture of what I'm seeing and then just blur blur blur until I can't even recognize a face. Is that is that basically what the person's experiencing?

So in this disease, yeah, they lose their high resolution central vision. So there's like like you when you look online you might see an image of like a big black splotch in the center of your vision and the patients don't experience a black slot splotch. So it's like a little bit we think describe it as missing. So it's kind of like like a big extended blind spot

but um yeah I mean all of your high resolution high acuity central vision gets uh destroyed in in this disease geographic atrophy. And so we have a chip that uh we implant under where the dead retina is and can stimulate this retina directly

bypassing the dead cells to get the signal back into the brain. So like if you I mean humanity is empirically not that good at drug discovery

but the brain is a computer and there's a cable that carries the visual signal into the brain and if we can get the signal into the cable then you can get vision into the brain.

Wow, that's remarkable. Uh, how much is the device that you implant like a camera sensor that I would see in an iPhone?

The So, there's a wearable. So, the GL the patients wear glasses and on the glasses there's two parts implant, but then there's the glasses and the glasses have a camera that is just like an iPhone camera and there's a laser projector that projects the image onto the implant.

Okay. But the the cool trick here is the implant is it's kind of all these tiny little solar panels.

And so it is powered by the image that's projected onto it. And so this means that there's no battery, there's no cable coming out of the eye. It's just that little chip because it is like when you project the infrared image onto the implant that powers it at the same time to stimulate the retina.

Wow. Um this feels like an invasive process. How invasive is it? Is this an expert surgeon doing the work? Are you thinking about robotics? uh what's involved in actually obviously the reward is so high that it's worth it but uh what does that look like now what will it look like in a decade

yeah I mean this is done by by human surgeons it's done by vitra retininal surgeon yeah

um and it's a but it's a simple outpatient procedure so you can do this under local anesthesia so you make an injection next to the eye it goes darken for a couple hours

they front of the eye it looks I mean like watching a video looks kind of like

um I don't know that you would do this electively but the uh It's It's really not a big deal. It's an about an hour and a half and the patient goes home and a day later they're they can take a bandage off.

That's remarkable. Um and so so uh walk me through like

post implant postsurgery monitoring. Do you have benchmarks? I mean you mentioned like a 300page book that seems like okay it worked. But uh are you giving them the normal like eye test? What's the what what what what what are all the different batteries that you use to understand the effect of the uh of the device?

Yeah, I mean there's a bunch there's we looking at both efficacy or things like how many letters of uh can they read on the ITAR.

Yeah.

And improvement in that that was one of the main trial end points. So we wanted at least two lines of improvement. We got on average better than that.

And the uh and then you're also looking at measures of safety. So for when is there anything worse off for having this like in their they have the chip in but the the laser is not on is there residual vision because many of these patients have residual peripheral vision that they can use to walk around. It's just that they can't they can't read. They can't definitely can't drive.

Um and you don't want that to get worse. And so this is there's various ways you can you can measure that. You can measure the range of the the visual field. Um but yeah I mean the eye chart is a very standard test. We we use that as a standpoint.

Um, sorry, J.

Uh, what uh you've raised a bunch of new money like what what is this going to go towards? Is it you know continuing to focus on on the the core product? Like walk us through.

Yeah. So the so we we announced yesterday that we raised $230 million series bring the total amount of capital to a little under half a billion which is like you know it's like a lot of money. A lot of money, our

founder,

a lot of money for an incredible incredible cause. Congratulations.

We got ourselves the privilege of working really hard. So I I uh can you take me back a little bit to how you discovered this particular problem to solve because it feels like you've throughout your career you've taken like a sort of general approach but then uh there there's always a question of like what's the most impactful uh quickest way to market most you know capital efficient there's a whole bunch of different uh trade-offs that I imagine you wrestled with before you landed on this particular product. So we have there we have three programs at science.

There's our work in the retina. There's our biohybrid neural interface technology which is a general purpose kind of what people think of as like a a neural interface but a little bit uh different than what most other people do. And then we have uh improvements in critical critical care life support technologies. All third I mean I think one way to conceive of what we're doing is like if Metronic was started in 2021 what are the projects they would be taking on like the next 20 years. And when you think about from a brain computer interface perspective, I mean I beyond like like like what is a BCI? This has become kind of synonymous for many people with paralyzed people controlling computers. That is what many of companies are doing. But I we think that coar implants are BCIs. Yeah.

I mean those have been around for decades and the and

it's so the retinal prosthesis is really in the tradition of coar implants. There's a million people plus with those out there. It's just until recently the technology hasn't really worked. And so when you think about what is the most valuable uh BCI product that it's an unmet need, restoring vision to the blind is really up there. And so the need is is one of those things that is kind of obvious in restoring vision to the blind is a quest humanity has been on for thousands of years. And then if you were to do that, when I looked at when me and my co-founders looked around at the state of the world in early 2021, we it was pretty easy for us to build conviction that that that one way or another by stimulating either the optic nerve directly or the the layer of cells in the right just before that using either a gene therapy or an electrical stimulator, one way or another, you kind of have these like four boxes. It's really like three boxes of those trade-offs. One of those would work. and we went out and did an exhaustive survey of the state-of-the-art in uh in kind of each of those three options and zeroed in on on what we're doing now.

And there's I think really like a first principles engineering uh approach to how you tackle the problem

and uh that has so far paid off. We're not the market yet, but I think we're getting pretty close. And the next step, this capital is to

um we've submitted for marketing approval in Europe. hope to be uh commercially approved there in the in the near future. Not not quite there yet. Um and then start to be able to commercialize it.

Talk about resolution. I imagine that the camera on the glasses is something like 1080p, something along those lines. But I imagine that the uh the actual implant needs to serve something that's maybe lower resolution or translated. you you're not you're not like running a yeah you're not running like an image to text pipeline and then piping that into the brain. Uh but what what is the input and then what goes actually into the brain?

So it's important to understand that only the central couple degrees of your visual field is high acuity color.

Okay,

you're constantly looking around to try to like reduce the uncertainty in the brain. What you you don't see the image on the retina directly. What you experience is a world model constructed by the brain.

It's highly semantic. Like you can look at something and kind of know will that hard or soft if I touch it. That isn't obvious but we're experiencing this world model. And so as the eye is moving around it's filling in areas of uncertainty. And so the implant you're right it's not to we're not restoring high resolution color vision today.

Um but it is what we are getting is this this intuitive form image that patients can look at. It can it strings together shapes into letters and letters into words and the brain can apprehend that intuitively and that had never really been done before. And so that is like a really clear statement that we're on the right track.

But the next step is to make the implant larger so that this the device that's in now is 2 millimeters by 2 millimeters which covers it's like looking through a straw and the electrodes are.1 millimeters which means that you stimulate a bunch of cells at a time. And so what that means is it's not that it is pixelated. They don't see it as pixelated, they see it as sharp, but they only get a small amount of detail at once.

Now, it does better than you would expect for the for the number of electrodes because as the eye moves around, it fills in this information. And so, it really it's a little bit

that's fascinating.

Yeah, you can't really think of it like camera in that way,

but we actually have the next version of the chips already in hand. The electrodes are much smaller. We're going to go from 400 to a couple thousand. Yeah.

Um and those are we're working on the the clinical study for those over the next year.

Yeah.

And hopefully every like two years or so we'll have versions coming out and we actually think that it it could be upgradable although we'll have to prove that of course.

Yeah. based on based on uh the kind of early uh early patients like how do you expect this market to evolve and how quickly uh how quickly uh people that suffer from blindness will adopt something like this once you continue to prove out the product. My assumption is that there's like an extreme eagerness to uh find a solution again for something that a solution that has never existed. But uh you know what do you think the adoption curve will look like?

Yeah. I mean blind patients are pretty motivated for the most part. Um so age related macular generation as a whole is big but not that many people have a form that is uh late stage and serious enough that this makes sense right now for the current so that AMD globally is 200 million people but that's not really relevant number both eyes large area of atrophy um the that's the population that's really relevant for the product as it is right now that's we think that's like six 6,500 people a year in the US and backlog of a couple of maybe 150 200 thousand people across the US and Europe. Um and with the next version of the device with a larger field of view and a smaller electrode that expands to probably five million people and these are these are we're not going

we're if the next several years we're doing a thousand a year that would be those would be numbers

and um I so right now it like I said it's not high resolution color vision but is this powerful existence proof that this technology is on the right track but I I do think that the next couple gen device generations will hopefully be able to get color percepts will be able to at least get red and green and you'll be able to get close to native acuity. Um you're not going to go sometimes people ask like okay when will I get this as an augmentation like you're not going to do better than the photo receptors if you have photo receptors you should use those

but I think that we can get pretty close to what those are for you.

Last question for me smellvision when are we getting it? How will we solve the problem of teaching AI to smell? Do we need to capture the olfactory system from my brain and decode that into tokens? What if I lose my sense of smell? Will you be able to help me restore my sense of smell in the future?

I mean, this gets to a fairly deep question of just like how do you capture the structure of any phenomenal mode?

Like and hearing are different, smell is different, sematic is different. So how do you capture like we have we can write down images and audio files to computer and so we have this format that captures them but that doesn't really tell you about like the like a wave file is different than an image file and so like what is a small file like and there's actually really interesting research on like you can there's like it's it's a chemical sensor spans the space of molecules you can get this decomposition of the space of molecules um but this is yeah smell has been very um there's a lot of companies that over the years have worked on growing up neurons and express single factory receptors to try to do like chemical detection. Um because like dog noses are way they're like some of the most chemical most sensitive chemical detectors in the universe.

Yeah.

And I actually was talking to a company the other day that is training dogs to detect cancer.

Yeah.

And I went into the pitch being like I don't believe you. And I came out oh like I get like that actually is like I could see it.

Yeah. And the trick was the really interesting thing just really shortly was wasn't not it's not that you're picking up some chemical that is present if you have cancer. It's that there's thousands of chemicals that are constantly just like being emitted that are like in your blood that are being moved out. And if you can train a dog under reinforcement, what it does is it kind of upregulates some receptors, downregulates some other receptors, the brain wires together to pick out that protochemical space. And so it can pick out this exquisite balance of hundreds or thousands of chemicals to say when these are present in these proportions. Yeah, you should get

that's fascinating. So it's not like there's a cancer chemical. It's the irregulation. It's the pattern. I mean this seems perfectly tailored for artificial intelligence, but uh yeah, we don't we don't even have a smell file yet. So dogs still employed

in the singularity for sure.

Yeah. And so it's like if it's not one chemical, then you can't just massspec it. You want to detect thousands of them, which is why they're funny as dogs.

I love it. I love it. Well, thank you so much and congratulations on the progress. Thank you for everything you do. Uh what

how last question from my side. How like talk about the significance of this milestone. You obviously raised a lot of money out the gates, but is this like like hey, we actually like is like becoming a real company now and moving moving beyond uh because everybody talks about uh big, you know, there's a lot of big numbers that get thrown around in our industry. Uh if you have, you know, some AI database company with 10 million of ARR, you can raise $230 million. But this is a very different kind of type of business where you had to raise a lot of money out the gates, but you maybe wouldn't have gotten a bunch more money if you weren't really onto something.

Yeah. I mean, our focus, I I really firmly believe that there needs to be businesses in this industry. I mean, you can spend an infinite amount of money just on our research. I mean, it's an infinite well of basic research. I mean, this is like there's so much grant money that goes into neuroscience, neural engineering, but we um I don't I mean I I I just like strongly feel it's like I we're suffering from this like money cancer that every now and then you kind of goes into remission when you raise around but then it comes back and like it the feeling won't go away until you have a sustainable business where there's revenue coming in and I think that we're hopefully pretty like we're not we're close to that's like the thing we're focused on is um getting to a real business that reaches many patients and improves their lives and there's things that you can point to that are like clearly valuable um to to lots of people that allow this to become sustainable and then it's a question of how fast do you grow rather than like is is BCI going to have a winner.

Yeah, makes sense.

Very cool. Well, thank you for coming on and sharing and and for the work that you're doing.

Yeah, have a great weekend. We'll talk to you soon, Max.

Yeah, I'm sure I'm sure Max has very relaxing weekends.

Yeah,

nothing going on at the office. No. Uh, good luck with all the hard work ahead of you. Thank you for everything you do. Great talking to you soon. Goodbye.

Cheers.

Let me tell you about Railway. Railway is the all-intelligent cloud provider.

Use your favorite agent to deploy web apps, servers, databases, and more while Railway automatically takes care of