Luke: You’re listening to a new episode of The Brave Technologist, and this is an exciting one ~~this week. ~~I’m super excited for you all to, tune into this week we have Sean Hehir, who’s CEO of Brain Chip, the company behind Aita, the world’s first neuromorphic AI processor that mimics how the human brain works to bring together intelligence directly onto your device.

In addition to innovating an ai, Sean spent decades building global partnerships at Microsoft hp, and at Cisco. In this episode, we discussed neuromorphic computing, what it is, why it’s important, and how brain ship is innovating in this area, how edge and cloud AI are. Different in how they’ll work together over time.

How Sean and Brain Chipp worked with Answer Technologies to unveil AI powered wearable glasses that predict epileptic seizures before they happen, and how Brain Chipp is working across health tech. Automotive and consumer electronics to make brain inspired AI a reality in the products we use every day.

Uh, This is a super cool episode. I think, [00:01:00] it hits on some really practical things that are getting innovated on that will help people’s quality of life. And I’m really excited for y’all to check it out. So, without further ado, now for this week episode of The Brave Technologist.

Sean, welcome to the Brave Technologist. How are you doing today?

Sean: I’m great. How are you, Luke?

Luke: Doing well. I’m doing well. We really appreciate you making the time to join us. I’m excited for this episode. And for listeners who are kind of new to the concept, how do you kind of explain neuromorphic computing and in simple terms and what kind of drew you to this area?

Sean: You know, it’s the simplest way to think about neuromorphic computing is probably through an illustration, right? Humans are the most efficient computational engine around, and the way I describe it is like I just met you a few moments ago, right? And so next time I meet you, I’ll know you. Why? The first thing I notice, I look at the hair color.

Your eye color. You know you wore glasses. I don’t try to memorize the face structure ‘cause most [00:02:00] people have two eyes and nose and a mouth. And what I’m illustrating is ignoring things unless they change, right? Hmm. And so, nor for computing, it’s just that it’s slimes level. We only compute when things change.

So it makes it very efficient. So when I meet people, the person, oh, that he has a beard. Oh, he has glasses. And so it’s very efficient. You’re not spending all this synergy. You imagine how exhausting it would be to wake up every day, meet a new person, say. And try to memorize the entire facial structure.

It’d be exhausting and, and very inefficient. And so that’s how you look at neuromorphic computing. It’s just a way to compute when things change, which makes it very efficient. I’ll make one more comment and I’ll pause when you go. Traditional computing, they’re doing that kind of brute force computing everything all the time, so you, you get the contrast.

And I can get a lot more technical than that, but that’s the layman’s way to think of it.

Luke: No, that’s awesome. I think that that helps, to really kind of, paint a clear picture too. What, when you, when you think about it from that perspective, like what, what problems are ones that you’re spending most of your time on focusing [00:03:00] on with this technology and, who kind of benefits from those.

Sean: You know, that’s, it’s almost impossible to limit it because it, because it, it’s one of the great pleasures of my day and my job as I come in. I get to talk to many prospects in every customers, and I’m constantly amazed about the creativity and the use cases that people are coming up with. The AI revolution is for sure.

Here and it’s impacting all industries. So if you look at even traditional ways of doing ai, it’s impacting it. and there’s really no limitations for art style of computing. So you can apply it to anything. Now let me be a little more precise to answer your question though. If you’re that efficient, like us, what makes it so unique is the ability to do that where there’s no internet connection to do it locally.

And where power really, really matters, right? So things that you maybe wear on your wrist, things you wear on your face, things you put into space, things you put on a drone, things you put in [00:04:00] an electric car ‘cause you’re trying to suck battery out. That’s where we see the most impact, where battery life matters.

Having said that, we see many, many use cases and deployments where people using a power cord, because the compute profile is so powerful and it works incredibly well. It’s not limited to only no power, but the ideal environment is where power is really limited.

Luke: Hmm. Interesting, interesting. and I know you’ve, you’ve kind of, has this been an area that you always have had an interest in with your company and, kind of personally, or is this how, how’d you kind of find your way to this area?

Sean: Yeah, so I’m a lifelong tech exec. You and I were chatting a moment ago when we first met. I spent many years in the valley. If you look at my resume work, some of the biggest companies in the valley, some smaller ones as well. And of course when AI started to become a, a really important compute element, a few years back, or let’s say 10 years back, I became more and more interested in it.

My what draw me to brain chip was really simple. When the opportunity to come here being in the valley, I had access [00:05:00] to some of the greatest technologists in the world. I showed ’em the company, I showed ’em the technology, and everyone said, this is pretty amazing. Secondly, when I met the team and I realized we had some of the best scientists and engineers in the world, and again, I’ve worked with some amazing companies, so amazing technology, amazing talent in the company.

And then when I got to talk to some of the potential prospects and customers and use cases, I, I just couldn’t resist. It was drawn to this because it’s a point in time. I said a moment ago. What’s going on around the economy, around the world, around AI is so true. It’s impacting all industries. I knew I wanted to take this opportunity and position this company to be the leader in edge computing.

Luke: Interesting. Yeah. and one thing that kind of comes to mind when you kind of were giving that illustration too, of, neuromorphic computing, at least from it, is an area you kind of, I feel like we heard a bit about a while ago. and I’m just kind of curious if it’s on your radar as much.

It seems like for things like kind. Early detection types of illnesses or things like that, that this is an area that [00:06:00] could be like a really great application for what you guys are doing. Is that stuff you guys are seeing more of or is that area of, because you hear about medical kind of being this area where AI could be a huge disruptor or get great progress on.

I’m just kind of curious just from your point of view, if, if you’re seeing that happen. Now.

Sean: Yeah. Yeah. Absolutely. By the way, you had kind of two, you had a statement in there and a question, I wanna talk about your statement first. You said Sure, sure. You heard about your, you heard about this a while ago.

You know, it’s interesting when you hear about neuromorphic kind of technology. It has been going, let’s say, for several years, been out there, but most of it’s been contained to labs in research organizations. Right. Why is that? Often the implementations are clunky. They don’t use industry frameworks, like they don’t support models developed by things like PyTorch or Tensor TensorFlow, or it’s analog or it’s hand programmable.

So it’s interesting, but clumsy not ready for mainstream use. Well, we did A brain chip is, is what makes it unique, the ability to apply that [00:07:00] kind of technology, but make it easy to adopt, right?

Luke: Mm-hmm.

Sean: And so. We do support models developed in those very standard industry frameworks we have. We have a piece of software, a software stack, that allows you to take those models and convert ’em and deploy ’em on our hardware.

So, you know, if you’re an enterprise, you don’t wanna retrain all your engineers into something. They know these frameworks, let ’em use their frameworks. Then you just deploy it on our hardware. So we did the best of both worlds and we, and we, we strive for that, which is. Technology, breakthrough technology excellence, yet easy to adopt because in the end you’re in this for people to put it in their products and change the world.

That’s awesome. Now, now let me answer your question a little bit about, you know, the second part of the question, we talk about medical and early detection. Absolutely. We see that everywhere. And I know we got a question coming up later about a very specific use case, which we can talk about.

But yeah, we are getting a lot of inquiries for, from, wearable device companies, medical companies, because it’s ideal when you talk about kind of computing or AI [00:08:00] things that differ or anomaly detection for lack of a more, you know, make it really simple. really, that’s really what medical early detection is, what is different, what is wrong.

So you, if you’ve got something like us that you can do at a very low power in a wearable device, it’s ideal for that. And then of course, with the illustration I said on the front end, you’re only looking for the differences. You’re, if you’re really just looking for something where your health may be a little bit different than it was in the last.

10 years. It’s ideal for that ‘cause it says, oh, we know this. We know where your patterns have been. Something’s wrong, you should get it looked into. So we are seeing a lot of interest and yes, it’s a perfect use case.

Luke: That’s fantastic. ‘cause I feel like it’s one of those things where, you know, for people that are maybe less in the trenches on this stuff.

you hear like a lot of doom and gloom or a lot of oh bubbles or birds. but really, you know, these are powerful technologies and a lot of this grid is there. People have pacemakers, people have a lot of, you know, wearables now. And and you always. we’re dealing with like aging populations and one of those look back [00:09:00] questions is always like, gosh, if I only would’ve known about this sooner, if I only would’ve, maybe they didn’t contact the doctor.

Well, maybe you have something that’s helping you to know, to call the doctor. You know what I mean? Like, it seems like very, what you guys are doing seems like very, very, you know, related to making these things actually happen, which is super exciting, like from my point of view. But but yeah, that’s really interesting to hear that

Sean: you look at anything you read nowadays.

Longevity and wellness is, is in the forefront of everybody’s mind. Mm-hmm. Across the world, you know, in the United States, it’s clearly in our forefront. Right. And so people are just looking for tools to, to help ’em on their goals for sure. But absolutely is the number one thing that people think about in the United States Longevity and wellness.

Luke: Absolutely. and you’ve said before too, the future of AI kind of isn’t in a cloud. Like what are the biggest limitations from your point of view around cloud-based AI today and, and how does Edge solve them for ai?

Sean: Yeah. Instead of saying it like a limitation of cloud, what I always like to say is the right tool for the right job.

Right?

Luke: Hmm.

Sean: You [00:10:00] know, you talked about my background earlier. You asked about it. You know, I’ve been in the industry for many decades and when you, when you look, all compute models always start the same, regardless. They always start centralized because you know, it’s easy to control you, you’re developing and things like that.

So in this case, the cloud, but eventually you distribute those workloads to some centralized, some decentralized. So I’m often asked, do you ever see the day where the cloud is abandoned? Of course not. I think what I call is the right tool for the right job. these two kind of, let’s call it, approaches Will, will coexist and just, it just makes good common sense.

Right? And by the way, there’s some really compelling reasons why you don’t want to do things in the cloud. You know, if you’re, you know, we just talked about health and wellness. If you’re concerned about your privacy, you probably don’t want that stuff going up. So say you wanted to have an early detection on something.

And you could do it locally You can then control your decision when you want to notify a doctor who’s, you know, bound by HIPAA [00:11:00] and everything else here in the United States, versus it goes in and you don’t know what the data breach, what could happen. Mm-hmm. So there’s some really compelling reasons to do things locally like that.

But you know, of course I talk about the workload. Power distribution. It makes sense. Why do something you don’t need to in the cloud, you introduce latency into the cloud. You introduce security into the cloud. It doesn’t mean it doesn’t have the right, for certain aspects, it’s fine, but for others, no. The edge is much more appropriate.

Luke: and how do you kind of see that relationship between Cloud and Edge evolving? I mean, you know, I, I know in the past local, like there were, hardware, limitation or, constraints, you know, just based on, you know, okay. the population has like these types of devices with this much power and, and this much ability, like, how do you see this evolving over the next several years?

Sean: Yeah. You know, it, it’s interesting, you again, you made a statement there, which I like to pick up on, which is. The devices in the past were limited, but when you look at a company like us, that’s exactly what we’re doing. We’re, we’re removing those limitations. We’re increasing the capacity and the [00:12:00] capabilities, right?

And so what was impossible five years ago is not only probable but possible and deployable today, and that. Pace of change is not slowing down. It’s accelerating. You know, I’ve been in this company about three and a half years and there’s model types right now that weren’t even on the drawing board when I joined this company that are mainstream.

That’s the pace. That’s the pace of change. Wow. That’s the pace of change that we’re in right now, and so this is what we have to deal with and we want to deal with because. This innovation, you know, that tech generally does, but AI does is specifically a company like us, which is a leader or the leader on the edge, we are gonna push innovation nonstop and all those boundaries go away, quite frankly, over time.

And we’re gonna talk, I believe, a little bit later on, about some things where things just weren’t there. Now we can do ’em. it’s an amazing, amazing environment.

Luke: And, and, and you mentioned privacy earlier too. I’m just really curious. You know, are you [00:13:00] hearing more from people you know, companies you’re working with maybe customers around privacy concerns now than you were like six months or a year ago?

Sean: Yeah, absolutely. Absolutely. It, we, we hear it all the time, right?

Luke: Mm-hmm.

Sean: You know, not everybody is very secured about their, you know, very concerned about their data. Not only their health data, their financial data. Yeah, there’s a lot of privacy concerns there for sure.

Luke: what does generative AI look like at the edge?

You know, what kind of experiences become possible when these models can run directly on the device? Yeah,

Sean: when you said when they can, they can now

Luke: or they can now. Hey.

Sean: Yeah.

Luke: Informing us. Right.

Sean: You know, and, and Brain Shift is a leader there, and without getting overly technical, we, you know, we are pushing those boundaries.

We’ve public have said, you know, we’ve got an underlying technology. On our accelerator to support that. And we will be commercializing that in the subsequent, quarters. We’ve got a set of models that we have developed [00:14:00] that, you know, are generative AI LLMs that run exclusively on the edge, that have unbelievable performance.

And again, I don’t want to quote numbers and things like that here. Sure, sure. But stay, stay tuned. So basically everything that you can see in the cloud. Maybe not everything, but it might be. Everything can be done soon. On the edge. The models are getting more efficient, smaller, and on the hardware that can support it with some unique architectures.

There’s some very interesting things that are coming through. I would say simply stay tuned to us and watch what we’re gonna announce later this year. I think you’ll be amazed about the functionality we’re, we’re gonna be talking about. And, you know, later this year is not that long as we are already in October, later this year, early, early, later this year, early next year.

There’s a lot of things we’re gonna be announcing that you’ll see the possibility. So the short answer is almost everything you can see in the cloud will be available on the edge. I, I strongly believe that.

Luke: That’s fantastic. I mean, like, it’s great for us to hear too, like, you know, we’re a browser company, [00:15:00] a search company, and, and a lot of what we’ve done has been our whole, our whole kind of mantra has been around keeping your information on your device.

And I know when our team was working~~ earlier on~~ earlier on, you know, we, we, we hit some limitations around local models and stuff. So it’s super exciting to hear that you guys are, you got a lot of cool things happening and on the horizon too.

Sean: Yeah, I would, I would just probably add what you’re not gonna see as a.

Full blown kind of chat GPT thing because that’s just, you know, that, that is so big and, but you know, the, kind of the core functionality, sure you’re gonna see a lot of very interesting use cases, but not something that requires that much data and that, that large of a model, not in the short term, maybe someday, but you know, other than those extreme cases, you’re gonna see some pretty interesting stuff moving to the edge.

Luke: That’s what’s exciting too. I think, touched on this earlier too, the right, the right tool for the right job, and, and I think that right now we’re starting to hit this point where at least we’re, we’re doing some genic things from the browser and, and doing other things where there’re known unknowns on our side around what use cases are actually.

You know, what, which ones are more kind of hand wave showman types of use [00:16:00] cases that look cool in a presentation versus what are the things that are actually gonna help users with their everyday lives and bring convenience and bring, you know, more efficiency and productivity and all of that too.

Which just kind of leads me to like, I heard that earlier this year, brain ship was collaborating on some AI powered glasses that can kind of predict epileptic seizures. What does this tell us about kinda the future of health tech and wearable ai?

Sean: I think it’s just a wonderful example of that, right?

And it’s ideal because if you think about these glasses, if you saw them or the videos or read anything about ’em, it’s simply some sensors in the frame right about here. And, you know, your, your brain puts out certain signals or patterns prior to having a sign a seizure. And you know, again, an ideal use case, you have a very small piece of silicon.

That can last for a really long period of time. That does one simple task that says if this pattern comes out, alert the wearer, they could have a seizure within the next hour. And~~ I, and you know,~~ in the accuracy [00:17:00] rate of the model. Is very, very good from our partner and you know, they’re going through clinical trials right now, but it’s just a, a wonderful use case.

I loved it when I talked to the company themselves and they talked about their, their downstream customers, and they said a lot of people with epilepsy, epilepsy were afraid to leave the home because they didn’t know when they’d have a seizure. Now, if you could predict it, that allows you to get out.

Go to lunch, go for a walk, whatever, free up your whole life. So a real human impact story where people that were literally afraid to leave their home could for fear that they’d be in danger, can now do something and say, okay, I can safely get home and lay down, or I can safely go take my medication. So it’s a, I love that story.

Luke: It’s fantastic. I love it too. ‘cause I think these are the stories that should be getting a lot more airtime, you know, where you, these are things where it’s like, look like practical use cases that are like improving quality of life and in making it so that people that couldn’t do things can do things right.

Like, it’s such a, [00:18:00] there’s so many benefits that tech can bring for, you know, improvements like that. And it is great to hear that you guys are working with partners on, on making those things reality and, and that it’s possible. Right. Like I think if people are gonna have mental models of like what’s doable and what’s not.

And, and I think in some cases, like with you guys, you are way farther ahead than people actually realize. So it’s super cool to hear, hear about these things.

Sean: Well, thank you.

Luke: Yeah. So, are there any other kind of breakthrough use cases of, Akita that surprised you or your team? Or, or maybe we can go into a little bit around what is Akita to you?

If, if you don’t mind.

Sean: Well, you say you’ve got a somewhat technical audience in there, so it, so a key to. is our accelerator, right? The industry uses things terms like, you know, you have CPUs, GPUs, and then what they call an NM PU Neural Processing Unit. And basically, an NPU is, is, is a piece of silicon that runs AI models very efficiently.

So we produce an NPU. Now, we, we have some chips out there, but we also sell ip. To our customers as [00:19:00] well, who they wanna build a custom ASIC or an SOC and one MPU integrate it in there. And we’ve got some great examples. I’ll give you another partner that we announced last year that is building a chip for space deployment out Sweden called front grade Geisler Wonderful partner.

Development goes exceptionally well. And you know, you think about space, you could easily lose your communication. So you need to be able to do compute. So that kind of matters. You think about space battery matters. ‘cause once you’re out there you’re not gonna plug anything in. And so, so you know, I remember hearing this use case.

I’m like, wow, that’s a cool use case. ‘cause you talk about wonder. So yeah, every day I get these very interesting use cases that come around. But our MPU back to what we do is in there, of course, is the company produces this, which is that incredibly efficient ability to run these models. Mm-hmm. And I talked about it earlier.

We do it supporting industry standard frameworks, make it easier for people to deploy. We also are, [00:20:00] are building a really comprehensive set of models and solution sets as well. So we got, we’ve got customers who say, can you not only gimme an MPU, but can you give me a model as well that does certain things?

We do that as well because some companies don’t even have the wherewithal to, deploy these things. They know what they want to do, but they really can’t do it. They don’t have the skills in house, and I think that’s really what’s going on in the industry right now. People that can supply a lot of expertise, like a company like us, not only the underlying technology, but the expertise and the models, et cetera, really stand out from the crowd.

Luke: Oh, it makes sense too. Things get way more specialized in this, you guys’ specialty, right? Like and, and what a cool juxtaposition too when you think about it. Like, you know, whether you’re talking about sensors in glasses that can help somebody with detecting seizures to doing stuff in space, right?

Like, like it’s such a. You know, macro micro, right? Like a, a cool, yeah, a a cool, like kind of a, a picture they’re painting there of, how this technology’s getting used. I, I think it’s, it’s really, really cool. I mean, so, so we talked about kind of some of the [00:21:00] breakthroughs, you know, what are the most kind of important user concerns you hear?

and how are you guys kind of, addressing and by concerns that could mean things from, like, we talked about privacy, but also like kind of trust and safety, and how is Brain shift addressing those things? With your work?

Sean: Yeah, I, well, we talked a lot about it, right? It is certainly around privacy, so we, address that.

But it’s funny, when I hear the word user concerns, I really gravitate towards the kind of answer I was given earlier. They wanna make sure of a couple things. One, that it’s easy. To use. Mm-hmm. Right, because, so when you say users, it could be the users that are building the technology for their products or the people that consume it.

Right? So it depends on which, yeah, yeah. Which user you’re talking about. So you know, I’m talking about the users that would buy our technology. They wanna make sure it’s easy to deploy. That we support frameworks that we’re a company that says what we do, which we do. They wanna ensure that we’re a true partner and operate in the spirit of partnership.

Those are the conversations we have. And one of the things that [00:22:00] I, when I came to this company, like all good tech companies, we’ve got incredible technology, but I really want this company. And we have, we are, and we have become better at it. Is incredibly customer focused. Our customer success is our success.

So when, when you engage with Brain Chipp, you get the full power of the entire company, all the engineers, all the scientists. So we really address that Now, downstream, their customers, their users, their concerns primarily around privacy and things like that. We feel with the technology we have serves them incredibly well.

Luke: Yeah. That’s awesome. And it’s a good distinction too, I think you, especially if if things are so specialized, right? Like, and, you gotta be looking at it from kind of first principles, right? Like through the whole, the whole stream. And, and seeing that like, cool, like, you know, the, whether that’s like, working directly with the partner or thinking on that, that.

Customer. ‘cause people do put a, a huge amount of trust, right? Like into the device that they’re using. And there’s a lot of pieces to that and, and a lot of relationships that kind of go with that. So it’s, it’s a good distinction. You break out there called energy efficiency, kind of the hidden [00:23:00] frontier of ai.

You know, why is efficiency such a critical piece of the puzzle?

Sean: Sure. Well, I think, you know, literally a day doesn’t go by if you watch any kind of news or eat anything on social media, you recognize the power situation around the world that you know, you know, with AI going up like this every single day and the consumption of it, we just can’t build enough power sources.

We just can’t. And so you’ve got some choices. You either slow down the innovation, which is most unlikely. Or you drive efficiency, right? And so not an option, it’s an imperative. It will be done. It has to be done. You know? And then just as being good, good global citizens, we really don’t want to be doing this with power.

I mean, it makes no sense, right? Right. To be wasteful with these power. So, you know, it’s ironic when you talk, we started this by talking about the way we approach compute in a very efficient way. And people say, well, that’s interesting. You know, I never would’ve thought it that way. I actually, when I wake up in the morning, I’m like, why would you ever think about it the old way?

I mean, do you get up in the morning and say, how much water can I [00:24:00] actually try to waste today, or Right. You know? You know, it’s just because we got used to this as society. We grew up with this, but the right way to look at it, the way we look at it, which. Do excellent job with the lowest power. It’s good for the world.

It’s good for the environment. Advances technology,

Luke: and it’s doable now. Right. It’s, I think that’s another big part, right? Yeah. That’s, it’s doable. That’s the thing. I love people to take away from this conversation, you know, because we covered so many different things and, you know, a lot of. Really awesome stuff’s happening now and is possible now.

And it’s super exciting. I, I think, is there um, from your point of view too, like along this vein, like, what’s the conversation that we aren’t having enough of innovation space from your point of view? Or what would you like to see us all talking about a little bit?

Sean: You know, I, the conversation that I think we’re not having is why not, we should always ask why not?

And push, right? Again, I, I mentioned earlier the things that we’re doing now we’re almost inconceivable when I joined, and so I’ve stopped thinking about limitations and start thinking about why not [00:25:00] and, and what use cases can we, we, can we actually close? Or what can we actually drive and how can we change the world?

There, you know, this technology really has the way, you know, we gave some examples about humanity. We could change the world here. We can change people’s lives for the better. We can make the world more secure. We, you know, we can make ’em healthier. So I always say, why not put a challenge in front of a company like us?

We’ve got the resources and the brains and curiosity to solve it. So that’s how I would answer that question. Start asking why not ‘cause anything’s possible.

Luke: I love it. I love it. Simple, easy. A good, strong, powerful question to be asking a lot more often. I, I, I love it. I really appreciate you’ve been really gracious with your time and, and covered some really interesting stuff that you guys are doing.

Do you mind letting our audience know if they wanna follow along with what brain shift’s doing or, or anything that you’re doing, you know, out in public? where, where should people go to check you guys out?

Sean: Yeah, absolutely. Please go to brain chip.com and check us out. We’d love, we’d love to hear from you.

Luke: Awesome. Sean, I really appreciate your time today. It’s been a fascinating discussion and [00:26:00] love to have you back too, to kind of check back in on things and, and we’ll be definitely tuning in on for those announcements towards the end of the year.

Sean: Yeah, sounds great. We’d love to do it again, Luke Enjoyed it very much.

Alright,

Luke: thanks very much. Thanks for listening to the Brave Technologist Podcast. To never miss an episode, make sure you hit follow in your podcast app. If you haven’t already made the switch to the Brave Browser, you can download it for free today@brave.com and start using Brave Search, which enables you to search the web privately.

Brave also shields you from the ads trackers and other creepy stuff following you across the web.