Crafting & Owning Your Personal.ai w/ Suman Kanuganti | Ep.122

In this episode, David and Gary speak with Founder of Personal.ai, Suman Kanuganti. They discuss the aspects of how you can harness the power of ai to prioritize humanity- not lose it. With the power of conviction and a little of “what would Larry do,” Suman shares his path to success and how AI has transformed his life and how he has transformed AI.

Links:

https://s.personal.ai

___________________________________

Submit Your Questions to:

hello@thebigpixel.net

OR comment on our YouTube videos! - Big Pixel, LLC - YouTube

Our Hosts

David Baxter - CEO of Big Pixel

Gary Voigt - Creative Director at Big Pixel

The Podcast

David Baxter has been designing, building, and advising startups and businesses for over ten years. His passion, knowledge, and brutal honesty have helped dozens of companies get their start.

In Biz/Dev, David and award-winning Creative Director Gary Voigt talk about current events and how they affect the world of startups, entrepreneurship, software development, and culture.

Contact Us

hello@thebigpixel.net

919-275-0646

www.thebigpixel.net

FB | IG | LI | TW | TT : @bigpixelNC

Big Pixel

1772 Heritage Center Dr

Suite 201

Wake Forest, NC 27587

Music by: BLXRR

Transcript

Suman: [00:00:00] That's the reason why you need your personal AI. That's the reason why it's a place to embrace, to take control of your own life, and not continue the same path in this next generational shift that is happening with the AI space.



David: Hi everyone. Welcome to the biz dev podcast, the podcast about developing your business. I'm David Baxter, your host. I am joined per usual by the ever loving, ever all knowing, all seeing Gary Voigt in his little beanie baby hat. Hello, Gary.

Gary: What's up? Have you heard the song that I think Jimmy Fallon and Paul Rudd did it on one of the intros to the Fallon show. about wearing beanies. Okay. Well, look it up online. It's pretty funny. My daughter introduced me to it, but it's like a little skit. So if anybody wants to hear that, look that up.

David: Wow. Thanks, man. What an intro. We are joined today by our guest Suman Kanuganti, who is the co founder and CEO of Personal AI. [00:01:00] Welcome to the show.

Suman: Thanks for having me, David and Gary. Nice being

David: So you run a company called Personal AI. Tell me what that is.

Suman: here. Personal AI, it's a model. It's a model of a individual person replicating their voice, their memory, their facts, their opinions, their perspectives. So if you would imagine talking to Alexa AI, now imagine talking to David AI.

David: Oh man, that's scary. I definitely don't want to talk to Gary AI. That would be even worse.

Suman: Or maybe makes

Gary: me think of, kind of, what does Apple have that new feature where you could record your own voice? Uh, I guess for people who might be experiencing like a loss of voice or becoming non communative, they could actually replicate

Suman: referring to likely

Gary: voice and personality with Siri, something like that?

Suman: Yeah, so you're referring to, you know, likely voice cloning, voice cloning technology exists within the space. Um, there are a few companies who are doing the voice cloning models. We do voice cloning models as well, [00:02:00] but more importantly, we do the idea of representing the model within your memory.

So it's digital twin of you, not just from how you sound from a voice standpoint, but how you think, if it makes sense.

Gary: Okay.

David: load that? I mean, that's where I'm, I'm confused. How do I, how does this bot learn who I am and what I like and don't like and how I would react to something? How does that get loaded into your.

Suman: Yeah, so fundamentally, it's not a bot that is trained one time, right? We are used to training the models and using the models, like in the large language model space is quite often referred to as pre trained models, right? What we've been building over the past four years is a model that continuously learns with you over a period of time.

So now there is obviously a starting point, which you are referring to, like, how does it learn? The starting point is pretty much any. [00:03:00] Authored content, authored compositions, authored writings that you have ever had in the digital realm from your history can go into your memory. And the idea of loading up the memory is also in your hands, meaning you may choose certain memory to not be in there.

Like, for example, if somebody has a PR crisis and like, Oh, I don't want that. That is not truth. That is not truth to me, right? So I would not want my AI to learn. Anything that is available on the Internet as much as I want this AI to learn who I am authentically that I would want my AI to be right. So it's an extension of you.

Not necessarily a company is building a model of you from just. Random places around the internet, but very deliberate, um, training that individual person does.

David: So if I've written a hundred blogs, I would load it into there and it would now be able to speak in [00:04:00] my voice as a blog writer. But that wouldn't necessarily, unless I really wanted to, it wouldn't know anything about me personally. It would just know that I am an expert at X and I write about X all the time.

And so it could use my mannerisms and stuff in the way I write, but not necessarily know who I am. Is there a way to do that? Or is that not even the point? I guess maybe a better question is. Once I've done that, what do I use this for?

Suman: Yeah. So let's talking, let's talk about the done part itself. So yes, you can do all those blocks, but you would have to set an identity. So every model in the personal AI scenario has an identity. What's an identity? Well, David is an identity. Gary is an identity. Why the identity exists because though that person is the owner of that model, that person is the owner of that memory and they are accountable for what goes in it and what's not.

So that's like blogs, email conversations, text messages, uh, articles, books that you have authored, um, things about you, your LinkedIn bio, your favorite foods, uh, you know, any, [00:05:00] I mean, you can literally tell your AI what are your favorite things as well. And that can go into. So it's your choice, technically, how you would want to shape it.

David: Okay.

Suman: of the use cases that we see today are tailored towards having this AI help you in your day to day communications aspect of it. Well, so because you probably have 10 million channels that are bombarding you with messages every day. You know, your emails or text messages are people asking you questions, the same thing over and over again.

Now this is. This is you. This is David's mind, so you will be able to craft the messages and the responses for you on your behalf, and you can choose to either send it or, oh, you know what? I will edit it because it is missing a point, and then I would send it, and those things also goes into reinforcement learning.

David: Okay. So

Gary: can see in about two weeks now all the Slack replies are just going to be David's AI bot and not him.

Suman: [00:06:00] Yep,

and that

is a real use case that happens today with our customers.

David: man, that sounds so scary. It's like you're disconnecting yourself. Let's say let's assume for a moment that it's perfect and it can totally replicate my snark and, and my personality and my knowledge. Then you've now disconnected yourself from that. That's a weird outcome, right? Now I'm not involved in my own thing.

Suman: Oh, well, I will challenge you. I think it is exactly the opposite. It is actually expanding your connections because as individual people, we have only 15 to 25 meaningful connections and relationships that we can afford our time to, but you may have your team members. You may have your, you know, clients.

You may have your audience. You may have your family and friends. Right. They need access to you to your mind. You may not have time for them, but if your AI indeed is able to. assist you, like [00:07:00] sidekick as a sidekick is like, Hey, this is what I wrote to you. Would you like to send this reply to your mom?

Absolutely. Yeah. That's a nice note. That's what I would say. So it's less about replacing. It's actually strengthening the connections that you may have. Otherwise you will not have time to fulfill those,

right? So what is the, I mean, obviously this is a company I get that, Hey, this thing could write to my mom. Okay. That makes sense. But how would I use this? How do you guys do this as a business? What is the product that you're selling

So David, like today's, uh, most of our customers are business owners, right? Uh, specifically the business owners who has domain expertise in specific types of fields or subjects or topics, right? Uh, you can think of them as licensed professionals, certified professionals are.

Advisors are coaches and consultants are even, you know, therapists, if you will write anybody who has a certain [00:08:00] degree of reputation, certain degree of domain expertise, and they have, uh, specific existing trusted clients or customers or audience or communities that they need to serve is where we are seeing majority of the demand.

Why there is a supply and demand problem. These people are needed much more than they could actually offer. Like imagine one therapist, there are like a hundred therapists and there's 10 million people. It's humanely impossible to be able to do that. And at the same time, these are, um, very mission critical use cases.

So meaning in the scenarios where accuracy is mission critical is where personal language models will come into play. You know, I was asking like how technical we would want to go into is because we all know the AI and the large language model space. has penetrated into the market in 2023, which is pretty fascinating and pretty awesome for general purpose or a large domain specific use cases where you can ask a question or brainstorm certain [00:09:00] things or write specific templates, which is fantastic.

What we are coming to the market is bringing this core micro small language models that are grounded in one person that are far more accurate and far more reliable when. The accuracy is important and matters. You cannot make a single mistake.

David: Yeah, that's, that was my next line of questioning is, I mean, we all know of the term hallucinations and AI going crazy and coming on to their interviewers and all sorts of insanity that's happened over the last six, eight months. Right. If this thing is now representing me, accuracy is paramount, right? hmm. if, if somebody's, I mean, I guess they don't even know they're communicating necessarily with this bot, but if they start asking off the wall questions that are outside of that purview.

Does it sound human when it's turned? Whoa, let's go back. I know we're not talking about that,

Suman: Very good question. Very good question. So, [00:10:00] um, actually put my AI in the chat, you know, later on if when you want to try it out. Um, so it's fascinating, right? So to your point, every individual model has a certain piece of memory, but David only knows so much about the world, right? Mm hmm. If you start asking Suman, for example, about biochemistry or anything of that sort, then I wouldn't know about the knowledge.

So what we do is pretty much any question or any intent or any conversation that is going into this model, it first consults my personal language model, right? It constructs a response and that response comes with something called a personal score. The personal score goes from like zero to a hundred, and that basically determines.

How accurate is this particular response as if it is grounded in me, and that is your transparency trust factor that you are exchanging between you and yourself, which is your AI, but also you and other people communicating with your AI, right? So that's how we are able to, you know, essentially tell now if it goes outside the [00:11:00] boundaries that you asked me about the biochemistry, my AI can consult the beauty of large language models.

and then tell me what biochemistry is. However, in my personal school, it will be reflected. It's like 15 percent personal. That means it is impersonal. So if I am communicating with my own self, I can choose to use that general knowledge of biochemistry and personalize it and then add it to my own memory.

Because then I formed an opinion about biochemistry because it's almost like this learning process, but every learning process goes through your own specific opinions and knowledge and synthesis of. You know who you are as a unique individual. Does it make sense? So now those scores, then you can set a control or a threshold.

Like, let's say if you come to Suman Kanuganti's AI right now in that link that I sent you. Uh, I go autopilot mode, which is autonomous mode. And then I have a specific threshold, meaning if it is north of 45%, only then respond to David. [00:12:00] Otherwise don't respond. Because anything less than that is not accurate.

David: man. I just keep thinking of talking to myself. I don't think I would enjoy that at all.

Suman: No, no, no, You wouldn't. It's, it's, it's mind. It's almost like this experience that we never had before.

David: So you have been, so I want to take a turn here for a second, cause this is fascinating, but I want to back up and I want to see, you've been doing this, you said for four years.

Suman: Yes.

David: So four years ago, no one knew about chat GPT and the AI that everyone's super excited about. And, you know, gung ho, cause most companies we hear, especially startups and all this, like, oh, I'm going to AI that, but they're just building on the shoulders of.

What is coming out every day now you were way before that. So how did you get involved in this? Where did this idea come from?

Suman: Uh, dear. It's fascinating, man. Even for myself, like just being in it, just watching whatever is happening. Uh, it's fascinating. So I'm an entrepreneur. I built a company called Airobi for AIRA. [00:13:00] Um, AI for artificial intelligence, RA for remote, remote access. So Ida was for people who are blind and low vision.

So what we did was we gave blind people, uh, glasses, Google glasses or glasses with a camera, and we connected those people with remote human agents that uses a dashboard, almost gives them a teleportation into blind person's vantage point of view. Right? Now there is a lot of AI that is baked into the dashboard itself, like how do you predict where a blind person would want to go, or what kind of, you know, Personalized needs that may be needed for the blind person to, you know, navigate from point A to point B.

Um, and that was quite fascinating journey that unlocked the experiences for blind people to essentially gain the confidence and independence, if you will, for traveling the places around the world by themselves. Getting this vastness of. Going to a Disneyland and experiencing, [00:14:00] you know, what it means, like a visual person would do, you know, reading children's books for the first time, you know, as a blind mom, those are the experiences that were pretty, pretty compelling.

Uh, so that was about vision, right? Like using this core idea of AI with human in the loop for filling the gap of missing visual information for people who are blind. Personal AI was about memory, using this core idea of like using AI plus technology to fill. The missing gap off an individual's memory and experiences to be able to capture in a model.

It's almost like augmenting yourself with your own memory, because what we are made up of what our mind is made up of memory is made up of. Yes, we use it in our databases, but we also use it with people around us. The inspiration was my desire to emotionally as well as intellectually connect and have a conversation with the last co founder of my previous company.

His name is Larry Park. He was passed away because of pancreatic [00:15:00] cancer. So in 2020, I had this like huge poster in my company that says, what would Larry do, right? It's almost like, you know, walking into the shoes of in this situation, in this strategic thing, in this terms negotiation, in this business aspect, what would I do?

Um, and that's it. I always wish I had like Larry's AI to be able to have this conversation. So in 2020, uh, I embarked on the journey. Okay, you know what, I want to create this and, and, you know, spend like almost four to six months, you know, Finding my team in Sharon Zhang, who is the CTO of the company, uh, I got in touch with her after talking to 120 people that includes people in open AI, Google, you know, Microsoft, everybody else.

And she told me, Suman, if you were actually wanting to do this principally grounded and scale it for every individual, we cannot rely on back then cutting edge research, which was large language models. So we deliberately took a counter approach, building up from scratch, what we call foundational models for individual people.

And [00:16:00] these are small micro language models. We branded personal language models that is fundamentally based on the memory and identity. So here we are in 2023 happens, last language model takes off. Uh, and then people ask, so how is the difference? What is the difference between like open AI and personal AI?

Dude, and that's one of the reasons why we are the category creators of personal AI. 

AD: BigPixel builds world class custom software and amazing apps. Our team of pros puts passion into every one of our projects. Our design infused development leans heavily on delivering a great experience for our clients and their clients. From startups to enterprises, we can help craft your ideas into real world products that help your business do better business. 

David: Man, it's like hard to wrap my head around.

Like I think about AI a lot, just my job and I'm a nerd and all of these things, but I don't think I've ever wandered down this path, like where it becomes me. And I am out there. It's I've, I've thought about it and I've [00:17:00] had other ideas that you would have a personal AI model of yourself to kind of augment yourself.

I've heard of that one, right? That no one ever interacts with except yourself. Like it helps you. Oh, I met Gary at this thing and it's kind of reminder. Oh yeah. And I need to follow up with Gary or whatever. I've heard of that, but it's very private. This is a public version of other, that is, Oh, there's a lot

Suman: if you would choose to,

right? fair. Yeah.

It's public only if you choose to. It's shareable only if you choose to specific people. And it is private if you choose to keep it private. But the more important thing is, you need control. Like, do you have control on your AI? What your AI is doing? What data is going into your model?

Who my AI is talking to? And do you have visibility on where your AI is saying at what instances,

Gary: Yeah. That was my

Suman: So I think

Gary: The control

Suman: yeah. I think those are the, uh, control mechanics you would need to have. I'll give you an example, right? [00:18:00] It's not always about, you know, David talking to Suman AI and people talking to AIs, right?

It's about David talking to Suman, the human, and Suman is augmented by Suman's AI. So Suman is always in the loop. Human is always in the loop. So when David comes to me, if David feels like, oh, dude, I don't want to talk to an AI, in that scenario, my AI would draft a response for me. On my behalf, I would choose to edit it and then I would send it, meaning I'm approving the message already.

This is no different than, you know, using some of the templates in Gmail, but this is a highly sophisticated version of templatizing your life, you know, based on your own mind.

David: you ready for that, Gary? I mean, that is

Gary: I'm just waiting so we can like mix in the holographs or with this. And

Suman: So holograms, so there is, there are like a couple of hologram companies that we, one of our customers who would want to place their personal AI to holograms. So I think that is probably like six months away. It's coming up. [00:19:00] Think about this way. We, at the end of the day, are not going to live forever, right?

But we are essentially creating a form, or a digital version of a form, where, you know, potentially our, even our future kids can, can talk to. The idea of your unique experiences creating value, is going to be preserved forever, if you would choose to, and it can create value forever as well into the future.

So a hundred years from now, how the history will be told, will be totally different. Otherwise you'll have the same story of, you know, some few people telling a. Average of the average store is who knows it is truthful or factual. That is the individuality to the extreme.

David: Yeah. I mean, I can see it. It's just, this is sci fi stuff to me. Like, I mean, this is bicentennial man. And even though that's great movie,

Suman: But it's real, you know, it solves like real problems today.

David: cause I mean, now if you put that And I'm, I'm sure there are limitations to now you're, I see you recently have come out with a new model, a [00:20:00] version two, you're constantly in that, that solved issues in model one, right? So, and I'm sure, you know, of issues that you want to solve for model three, right? I mean, obviously it's getting better. But I, I just think, man, if you put that now in like the, uh, what is it? Meta specs, their glasses, the Ray Ban glasses that now it sees what you see and hears what you hear and all of that. But this is the model that's speaking to you. So you can, not only is it seeing what you're seeing, but it's now adding to your model.

This is what you did yesterday. I mean, there's just like crazy applications here.

Suman: The danger in there is the glasses itself, meaning that is coming from one company, which is Meta, right? So even like right now, if you take like last 10 years or one decade of where we authored our data, where we authored our minds, it's still very much stuck in individual platforms and big tech, which is Facebook, Google, Microsoft, Netflix, Airbnb, OpenTable.

Every platform has a [00:21:00] piece of our life.

David: Sure.

Suman: And if we continue that path, we will continue to give our pieces of our life to the big tech and the platforms. What we want to do, what I want to do is to flip that narrative. I want to give that control and power to the people, to you, David, to say, whatever I see that wants to go into my model goes into my model alone.

And it has to be interoperable with all the data places, all the data that you likely are producing. And then putting it to specific platforms because platforms will ultimately serve only one thing, which is their own platforms.

David: But you, you're talking about, I mean, that outside of the AI side, there has been a movement for lack of a better word about data portability is what you're talking about. I mean, and, and Google makes all their money and Meta makes all their money on the fact that they own a slice of your life. Like you were just saying, they have zero [00:22:00] incentives to allow you to pull that out and say, nevermind.

I don't want you to have that. Or more importantly, you have to come to me. And they make life miserable and tough to be able to even access your own damn data that you created on the internet, right? So now, now imagine a world like, you know, when everybody starts adapting to personal AI because they own their memory and they own their model. This model is working for them, right?

Suman: So I'm working for you, David, and you are AI. Here's what happens. Every services, every platform will have their own AI. Meta will have their AIs. Airbnb will have their AI, OpenTable will have their AI. Well, now there is all these individual consumers like, oh, okay, like, oh, okay, now I need to talk to this AI to get to my OpenTable reservation.

Well, guess what? That negotiation should be happening on your behalf, with your AI, with their services AIs. And that's the reason why you need your personal AI. And That's the reason why it's a place to embrace, to take control of your own [00:23:00] life, and not continue the same path in this next generational shift that is happening with the AI space.

David: Now that I, that reminds me of a podcast. I heard a while back, they were talking about there's going to be two internets in the future. One is the bots talking to themselves. And then there's going to be the humans hiding in the corner somewhere so that they can still find other humans to speak to. And that's, I mean, that's largely what you're talking about.

Cause you know, it's, it's not very far from now that, if Oh, I want to go out to dinner and I ask an AI. Bye. Give me some, some ideas and here's five ideas. And then you tell your AI to go and make some reservations and call in and set it up on all my calendar. We're not far from that. Um, I mean, I can do that manually right now pretty easily, but I'm still involved.

It is. One step away until these now all interoperate and now everyone's just talking to each other and that is super interesting and scary because I mean, at some point, what are we doing here? What, what is, what's my role in all of this? [00:24:00] Um, I guess that just to go on the date, right? But it's, it's just.

We're, I would say five years maybe from that. I don't know as fast as this stuff is moving, it is, it might even be faster than that, but the interoperability, one thing I do know about our world is that being interoperable is not in most people's advantage, right? They don't want to be because they get their money off of not being friendly with other apps.

I mean, as a program developer, we have to interface between apps all the time and they have a very, this is our API. These are the five things you're allowed to do. I would like to do a six thing. No, you can't. I hear the five things now, no more. And this is not going to probably change that because Google makes gazillions of dollars because they don't do this.

So it's like, man, that'd be a paradigm shift to allow you to say, no, no, no. You want my information. You got to come to me to get it. Not the other way around. Yeah. That's. Mm

Suman: So I think, you [00:25:00] know, it sometimes it's some, at least for me, like, it's easy, easy to think and easy to give examples from like a physical world standpoint, right? Because in the digital world, it's like all internet and we, we, you know, we, uh, it's kind of hard to imagine, but just think about, uh, you know, 20 years ago when initially the, these mainframes came in, right?

These are like huge gigantic computers, like, you know, either at an institution or a big corporation. And I remember, you know, uh, You know, standing in the line to gain access, like 30 minutes, uh, that I was holding a floppy disk. Then I go insert and then I type things in, I print it out and I get out.

That is only as personal as the information that can fit in your floppy disk.

Suman: So if I have to expand my personal life with the computers, it doesn't really fulfill. I cannot take that mainframe and put in my bedroom until the personal computers came along. This computer that I'm holding right now is in my hands. The files that resides in this [00:26:00] hard disk is in my control,

And with the AI, it's going to be the same thing.

The large language model right now is serving the larger needs. Like, you know, general needs. But we need our own personal AI to have that intelligence that basically is very similar to a personal computer which sits in our home.

David: from a business perspective Where are you guys at? Are you on runway? Are you self sufficient? Are you? Yourself, you know, we've had some people who's like I got money and that's how i'm doing it Where are you guys at?

Suman: We are growing, we are shipping AIs to our customers and our customers are really happy. Um, we are very much still at a, uh, like almost like referral to referral growth pattern because it's an early stage of this concept and it's also, [00:27:00] uh, involved with a step up function. How How It's going to integrate into your daily life, right?

And it's a superpower. It's like a superpower that people have probably never seen before. Yeah, you can use chat GPT and get a task done and get a summary done. But what you're talking on here, well, like hundred X, what you're talking about, potentially the, you know, the value that you can create around your world, uh, largely we are focusing on certified professionals and licensed professionals where, as I said, the accuracy is mission critical where.

Um, the supply and demand is not there and create like true meaningful impact because that's where my background is. Um, I don't like just doing for the sake of like doing it and, and hyping about it, but make like meaningful, real impact, you know, where it matters the most.

David: So we talk about the slog here a lot, which is that time of Of a business where [00:28:00] you are started, but no one knows who you are. Your job is to yell at the rooftops, to get attraction, to get your first customers, to get moving, to get people to care, right? That is a difficult thing. And sometimes it's months.

Sometimes it's years long before you get your own gravity, as it were, where people are now coming to you. I see your business. As a even more difficult because you are embarking on a new idea. Like this is a S this you're explaining a helicopter to a shepherd in the 15 hundreds, like to most people. I mean, I'm a pretty technical guy and this is blowing my mind.

So I imagine your average CPA that you're trying to sell to, this is a really. Sci fi sounding coming from the aliens kind of discussion that you're having to sell. Right. So is that, has that been easy? I mean, has this been an easy sell to people or it's just like, whoa, and you've got explained over and over and over again to get people so that [00:29:00] they understand it, much less buy it.

Suman: Yeah. So it's sci fi until it solves a real painful point in their day to life, right? So when I hop on the calls. What is the most painful point is like, dude, like I get thousand texts a day and I cannot keep up piece of cake,

David: Yeah.

Suman: right? It's not sci fi anymore. It's like, Oh shit. This is like actually drafting that sponsors like me.

And 89 percent of the time is damn accurate. I'm just going to swipe right and send it.

David: But how do you get to that point? Do you have like a free trial or something? I'm a CPA and I'm interested in this. I'm coming to your website. Is there a way, like I got to load it up first, right?

Suman: so in terms of process, there are two modes right now. One, uh, is essentially like self service. Because people are interested in learning and training it themselves, so they had to attend a training session and then yes, you are quote unquote certified to train your own AI, right? Because we want to.

We want people to care about the quality, not just throw junk at [00:30:00] it and expect junk out or something out, right? So it's like a very deliberate and intentional thing that we want people to do. And because when they do it, they're like, it's like a different life altogether. Um, and then we do also have, uh, uh, organizations like we like, like, like literally right now we are talking about.

All 12 attorneys at a law firm who are specialized in specific different, you know, cases that they handle, everybody getting their own AI that will then in turn unlock the hundred analysts inside the organization, get a unique perspective from the research and the ideation of each and every lawyer or attorney, right?

That's like, uh, all of a sudden, you know, communication. It's on steroids. Um, so that's kind of where we are. And I don't want to be the person to yell out loud on top of the mountain. Like you mentioned about, you But I just want to solve that one customer problem because that customer will bring us the second one and the third one and the fourth one and that's what is [00:31:00] happening and we've been building that momentum for the past five to six months.

Um, and gosh, yeah, I have big announcements to make.

David: And how long have you guys been, you've been around for four years, but how long have you had a product?

Suman: So first two years is all R& D, right? There is a lot of model development and building. Um, if you know OpenAI's journey, OpenAI started around 2013, 2014 and 2019 is, you know, technically when they started, you know, raising seed money and 2020, 2022 is when they released it. Uh, I think about it as we are at the 2019 stages of open AI, because, you know, we had to build that R and D muscle with memory and model for the first two years.

Third year is all about application development, the platform development, right? And finding who is the right customer and right target. Uh, and this year, which is fourth year of our company is going to the market and scaling. So that's kind of where we are.

David: Nice. Man, it is fascinating. It is [00:32:00] fascinating. Gary, do you have any questions? I want to make sure I give you your nerd time.

Gary: no, like the way my brain works is every time you would bring up a different thing that they could do or how you would train or teach the AI from your own stuff, just scenarios and scenes from hundreds of movies just kind of play through my head and correlate together how this might be a reality soon.

But one of the things that kind of sticks with me, this now seems like. It's a very specific market and a very specific target audience that you're looking for. It might be hard to get something this effective to be way more general use. And the reason I say that is kind of to piggyback what David said is people might think it sounds cool, looks cool, and could be cool, but they're probably not going to take the time to understand and learn how to develop it so that it's going to be of help to them.

So while you're acquiring these new. Customers and this referral engine, [00:33:00] um, do you have a lot of like just training sessions and hands on kind of guidance teaching them specifically how they need to craft their own personal ones and then how they have to teach someone else? Or like, I'm just wondering when you're saying scaling, you're, you're scaling and customers, but how do you scale it so that the adaptation is, is more. Easily adaptable, or I should say it's more easily understandable for, like you said, 100 lawyers in a law firm. There might only be 20 that are interested in actually. You know, feeding these AI models, the

Suman: and that's one use case. Yeah, that's one, that's one use case, right? Well, I disagree because here is what's gonna happen. Right now it's at the like beginnings of adapting AI as the new medium, right? Traditionally, like when the dot com happened, there is always like, you know, those early adopters. There was this concept of, Oh, you know what?

Now as a business, we need to have this website. Like, what is [00:34:00] a website? I don't need a website. Website sounds complicated. Internet, HTML coding. Okay, hang on. Like, I need a developer night right now. I didn't, I don't need a website for my restaurant. I don't. Well, fast forward 10 years later, every business, individual or otherwise, has a website.

The AI thing that's

Gary: that point, they have a website that they just paid 10 bucks for from some person, like on camera to make it in Wix or Squarespace, like, they're not really feeding those

Suman: Oh, okay. Okay. So he's a very, very good point. So that's where I was heading to. So that is a good point. I agree. Which is a Squarespace, right? Squarespace, they develop the tools over a period of time to make it easier. That is like 10 bucks and 20 bucks to get their website. But before that people had to invest time and money to understand how to build a website.

So that's where like personal AI is. But over time, when you have like one click button, and when you start running these models on your phones, then it will become more general mass adapted market, [00:35:00] right? Like Squarespace. So I think the tooling has to develop, the clarity has to develop, the use cases has to develop.

And we have to basically make the market from niche markets to broader markets to mass markets. So it's a journey.

Gary: Yeah. Okay. So in the future, that is where you see this going to where it's just going to be like, almost like a personal assistant style or feel in your life, but then it's going to actually be far more productive than just a person. Okay.

Suman: Our tagline is like 8 billion people, 8 billion AIs, right? So that's how we perceive the world

Gary: In the four years, your company has been around. I'm sure you've learned a few lessons and we like to kind of wrap up these interviews with one question. What would your top three pieces of advice be for any new entrepreneur, startup, or new business owner?

Suman: in general, or any specific to AI, just generally?

Gary: That is up to you. You can take it anywhere you want. The reason why I preface it by the four years of this business is because this [00:36:00] is an extremely interesting. So I'm kind of wondering if maybe these three pieces of advice that you've learned from this business could translate into other businesses.

Suman: Yeah. I think, um, the first one I would say is conviction. In other words, you're onto a path, you have a mission, right? Things happen out there in the world. Distractions exist. There will be challenges. People won't believe you. Some people will believe you. Things will work, things will will work things wont work. But are you convicted on your path in the mission?

And do you have the people that share your conviction? So I think one of the first pieces of like, I guess, advice would be Find your mission that you are convicted to with the team and with the people who are going to stay with you for four years of time. You know what I mean? Like, it takes a long time for something like this to come together.

So yeah, so first one is conviction. The second one is just generally grounded, right? What does grounded mean? It's [00:37:00] hard. There's gonna be ups and there's gonna be downs. There's gonna be success. There's gonna be You know, moment of death. Right. Um, so I think like staying grounded to those variations and thinking through practically what is the next step and creating the filter for yourself.

I think it's super important. Um, you know, especially, I mean, personally, myself, we went through a lot of, you know, ups and downs. Yes, of course. Like I come on the podcast right now and, you know, I confidently speak what it's going to be, but it comes with a lots and lots of hard work. Uh, and staying grounded to when you have like huge success and at the same time, you know, maintaining that calm posture, thoughtfulness when things are going down south.

So I think kind of tied back to the conviction, but I think grounded kind of is more about how you behave, how you act and what kind of made decisions that you make. Um, what else? I mean, other [00:38:00] piece of advice, specifically in terms of AI, I think build AI. With certain principles in mind that is more meaningful and impactful for the next generation of, um, for the next generation, you know what I mean?

Meaning if you just simply apply, like one example, if you simply apply like subscription based business models of internet 2.0 to AI models in internet 3.0 I don't think it'll work. You have to come up with, um, you know, what is the problem you're solving for? How, how does it add specific, um, mode for yourself?

Is it data? Is it models? Is it? Is it the space and then build it accordingly, uh, and have less and less reliance on one single platform or one single model to build your code business because somebody else will come in and then just wipe it out, right? So build something of mode. And I think it's important to spend time to build that mode in advance.

Gary: Those are good, good pieces of advice. I especially like conviction. That was good one. We have [00:39:00] not really heard that one too often before, maybe in different ways, but yeah, in the beginning of. Our interview here, you also mentioned how, I guess the, uh, the picture of your old boss was kind of a, a memory for you or a, uh, motivation, inspiration for you.

So I guess the conviction of knowing that you wanted to see that through seems to have come through. So very cool. Now, if anybody wants to learn more about personal AI, um, other than just the website, how can they reach out or get in touch with you?

Suman: Simply the letter S dot personal dot AI. And you'll see all my social channels as well as a mechanism to message me, chat with me. As well as experience my AI, but yes, I'll be still in the loop.

Gary: All right. We'll put all the links in the show notes too, including, uh, if it's okay with you, the link that you shared with us for your personal AI, is that all right?

Suman: Yes, yeah, that's s. personal. ai that I just

Gary: cool.

Suman: That's my identity now.

Gary: Is this your AI now or [00:40:00] have we been duped this entire

interview? is this good, if his video

signing

on. a whole different thing. Yeah, man, sign me up. 

OUTRO: Hi, I'm Christy Pronto, Content Marketing Director here at BigPixel. Thank you for listening to this episode of the BizDev Podcast. We'd love to hear from you. Shoot us an email, hello at thebigpixel. net. The BizDev Podcast is produced and presented by BigPixel. See you next week. Until then, follow us on Instagram, Twitter, Facebook, Threads, YouTube, and LinkedIn.