Who Gets to Shape the AI Future? | AI4, Agents, Open Source, and Small Business

Speaker A: Hey, welcome to a special edition of the Weekly Blitz. I am like over the moon excited because I have Nirav Shah with me. He was actually on the AGI podcast about a year ago, but he is a founder of Quantal AI. He is a developer, but he's the kind of developer I like, the kind that looks at things from the perspective of people who don't come from that background as well. And so Nurav, introduce yourself, tell us a little bit about yourself, and then we're going to get into the conference you were just at and everything that's been happening in AI recently.

Speaker B: Absolutely, Jason, uh, very excited to be back and uh, super, super excited to be in touch with you over a period of last year. And it's amazing how things have paced up since we last spoke. Uh, but just uh, to give a quick uh, background about myself again. So, ah, did my master's in computer science from Columbia University, then worked at UBS in their quant trading team. Like built automated trading systems and worked mainly with traders on the technology side, uh, and then wanted to start something of my own, so started my own consulting firm, uh, starting with independent consulting for quite some time and then eventually uh, started an AI agency around four years back with my friend, uh, Salish. And uh, we are a team of around 50 people working with various businesses to automate their workflows and even build uh, really cool AI products at our end. And uh, like I spoke with Jason in my last uh, podcast, I also had a brief time working in my family business, which is a very traditional business. And that's where I learned a lot about how to work, work with businesses, what businesses are, the challenges that they would face like accounts, bookkeeping, sales, marketing, and learned a lot from there. And uh, yeah, that's what's helping me, uh, looking back.

Speaker A: Well, I hope we remember to circle back to this because I don't want to start there. But everyone who watches this podcast knows I'm a Republican. But I think that there was a huge misstep in making it harder for people to get HB1 visas because half of the talent that we have in technology, if not more, comes from overseas, from India, where you're from the uk, from Canada, from China. So I don't really want to start there. But remind me if you can to get back to like how bad of an idea that was.

Speaker B: Yeah, absolutely. Would uh, you know, love to touch base on that because I think uh, I see world is more like a membrane where, you know, the best brains should come together to Create something exciting and adding more regulations to that is uh, you know, a little more counterintuitive as opposed to something which uh, uh, should uh, be you know, more like a free flow of ideas. But I understand, you know, where they're also trying to protect jobs which otherwise uh, uh, you know, it's a little gray area in terms of uh, the way uh, the jobs are to be protected, uh, while you know, ensuring that there is free flow of ideas that are um, you know, moving around everywhere. So.

Speaker A: Yeah, yeah. And I mean, you know what, let's, let's just get into that because my understanding when I listen to some of the Republicans who defend what the attempt of the administration to do that was that there are a lot of agencies over here who were mass buying HB1 visas and bringing mediocre talent over to replace, to actually replace jobs. Because the level of jobs that you and I are talking about, ah, like Columbia grad, Stanford grad, like computer scientists, AI scientists, like we don't have that here in America. We're graduating as just engineering in general, we're graduating about 200,000 engineers a year to China's 1.5 million m. So we do need foreign talent for that. If we're trying to keep out um, middle level talent because we have the people who could do those jobs. I understand it, but the attack on HB1 visas was not the right way to do that.

Speaker B: Yeah, I completely agree. It's like uh, trying to use uh, uh, a sword where a needle could have been used.

Speaker A: Yes, that's a great analogy. All right, let's get into. You just came from a conference. Tell us a little bit about the conference you just came from and then we'll start to unpack like what people are hearing about AI and some of the questions I have from uh, what the conversations were at your conference.

Speaker B: Yeah, absolutely. So I went to AI4, which was uh, supposedly the largest uh, AI conference in the US and uh, more than 12,000 attendees. It was massive. I uh, think uh, the booths were uh, I think a little limited, but a few things, you know, that I would just want to share my experience. We landed one day prior to the conference and there were pre registration going on. And I think that is one of the best things that we did. We went a day earlier and pre registered ourselves and we had to wait like for five minutes to get our badges for access and stuff like that. But people who arrived on the same day of the conference, they had to wait for two to three hours in a queue just to get the batch right. And they missed a lot of uh, interesting sessions. And it was just frustrating for a lot of people to wait in such a long queue. Uh, But I think AI4 anticipated that. And uh, they did not have the main speakers on day one because they were anticipating a lot of people would be in the registration process. Uh, but my advice to anyone attending such a huge conference, just arrive a day prior, get yourself pre registered and uh, to save uh, a lot of time that uh, otherwise you could spend uh, in networking. Uh, but otherwise that was I would say, a very good experience in terms of uh, what's happening. Just to know what's happening right now and the way it has evolved. I'm able to see the pattern is, uh, two years back people were saying, oh, use AI as your copilot, right? And I uh, think uh, let's do something in AI what is possible using AI. And then AI copilot came in and okay, now the theme was around more of AI agents on let AI agents do something. And um, uh, and a lot of discussion was around governance and roi. And we help you monitor the agents, we help you monitor the roi, we help you monitor the security of AI. Right? And so a lot of conversation has shifted to that, which I think is a positive sign overall. Um, but yeah, I think, uh. But yeah, um, that was, I think the overall biggest takeaway that I got, uh, on how AI discussion has transitioned from what's AI, how we can use it, copilots and then agents and governance now.

Speaker A: Yeah, yeah. So let me start with the conference because, uh, we're looking at, um. A friend of mine and I are looking at putting some conferences on here in Indiana and I'd love to hear like what you, what you go to a conference for. When I go to a conference, I go for two reasons. One, to network. There's always really valuable networking with people who think differently than I do and pick up ideas. And then the other is really to, to get into workshops that are relevant to me, that are close enough adjacen to what I'm doing that they're relevant, but that are far enough away from what I've thought about that they inspire me to explore more. Right. And uh, when I look at the M. Midwest, like I'm really disappointed in Macon, which is our Ohio, has this huge marketing conference. And I see one of their keynote speakers is Karen Howe. And I'm like, seriously, like you're going to bring such an anti AI? Like I made millions of dollars off of a clickbait book that has already been disproven as some of the information was incorrect. Like why would you do that if you're trying to fight against the Midwest momentum of uh, we don't want AI Anyway, like, so that, that is not the kind of keynote speaker I'm looking for. I'm looking for your Yan Leon or your Fei Lee or somebody who like, I'm like, oh, I would love to hear what this deep thinker who maybe isn't under the spotlight but is going to bring up things that really get me thinking, is going to be talking about it. So I just wonder what, what do you go, go to a conference hoping to take away?

Speaker B: Yeah, absolutely. And I think one of the key reasons for me to go to the conference was uh, just to figure out what's happening, uh, from my perspective because I run an AI firm, I want to ensure that I'm not kind of left behind. I want to ensure that uh, is there something that I'm missing that's happening in any space that I should be aware of? So that was my primary agenda because most of my customers are non technologists. So uh, uh, uh, in order to get business from such a conference, it's very highly unlikely. Uh, it's uh, great to connect with peers and see what the other technologists, uh, uh, uh, are building and share ideas. You may be able to get some references and businesses and network a lot. Uh, but uh, my primary agenda was to just see what's happening. Is there anything that we are missing out on? And uh, also obviously like you mentioned, I uh, was really looking forward to the keynote session where uh, Fei, Fei Li and uh, uh, Georgie Hinton and uh, even Andrew Ng were invited to this and it was a very interesting panel and uh, as you know, and I think that was one of the biggest highlights of uh, the conference. Uh, I loved attending that session where there were very different perspectives from all three of them like you mentioned. So Georgie is uh, a little, uh. He is you uh, know, as you know, godfather of AI, but he's himself. He quit Google to ensure that he can warn people about what are uh, the potential problems that can happen with AI. And he doesn't give a damn about anyone. He would. He was very honest, frank about his thoughts. He's not scared of anyone. He would say, I don't want Elon Musk to control AI. I want uh, you know, he. And one of the great analogies that he uh, gave was that AI is more like a car and governance is more like a steering. Right. So. And I Think that was very uh, aptly put. Uh, which figure was that?

Speaker A: Was that Jeffrey Hinton?

Speaker B: Yeah, Geoffrey Hinton.

Speaker A: He's also made a lot of money off of all that.

Speaker B: That's true.

Speaker A: But I think disagree on how uh, on how much he cares about society and how much just leans into being a fear monger and getting a lot of interviews out of it and maintaining relevance along those lines.

Speaker B: Yeah, yeah, I think it's interesting. And you know, at the same time, you know, God, uh, you know, godmother of AI, which is Fei Fei Li. She was very excited. Andrew Ng was very excited about AI and uh, commercialization side of things. Uh, one of the interesting things uh, that Geoffrey Hinton mentioned was that a lot of jobs like call centers, etc. Are going to get wiped away. Right. Which is maybe true. Um, but the counter argument and uh, there was almost a heated argument on stage between three of them and that's what made it more exciting. And it was a very funny, funny argument that happened between Georgie, Andrew, uh, Ng about who worked for who. So that was. That went on for five minutes. So that's like, you know, I don't think a lot of media is covering that but that was very funny. That uh, kept the audience uh, engaged and a uh, lot of laughter, uh, uh, uh, in the audience. Uh. But yeah, coming back to what Fei Fei Li was trying to say is that some jobs will get augmented with AI and they'll be able to perform far better. For example nurses where we cannot replace nurses or people uh, who require physical presence and augmenting AI to them will just enhance their capabilities uh, a lot. And uh, uh AI is amazing in terms of education and a lot of uh, stuff around how motivation is extremely important in terms of you know, education and how teachers uh, should be more open to about having students use AI because that's how they are going to eventually use it in the real world. But right now even a lot of educators are very anti AI. Right. And uh, that is also very dangerous because we don't want the kids to stop using something which they're going to use for the rest of their life in such an early age. Right. So that was an interesting conversation that uh, uh, happened as well. And apart from that I think Andrew was more uh, uh, he was pushing a little more on the open source movement, uh, in A.I. i think, which is overall great for

Speaker A: the

Speaker B: overall humanity versus being like closed source models. Because if you see any technology invention like a wheel or highways or even public transport, uh, if it benefits the society Equally, then there is a lot of prosperity. But if it's something which is very exclusive, then uh, it makes the rich richer and poor poorer. So I think I resonate with that. And I uh, think uh, that's what AI is. Right.

Speaker A: It's like uh, let me unpack for the audience because they probably don't know who any of those people are, except for. So if you go to YouTube and you look up Jeffrey Hinton, you're going to get a lot of doomer warning stuff of it's all coming for us in our jobs and it's going to escape and, and the world's going to get turned upside down. There is merit to some of that argument, like my PDM numbers. We won't get into all that. Right. But, but I. If you look up Dr. Fei Fei Liu, most of you probably never heard of her. She's not exciting to listen to, but she is really good. She lectured at Stanford for years. There's all kinds of YouTube videos. She is a heavy believer that the transformer architecture is not going to bring about asi, that we're going to have to look into world models and some other methodologies and algorithms to actually get to the next step of AI. And then Andrew Nang is a little bit more on the product and commercialized side. Uh, unless you're really like, have a little bit of a developer mindset, I'm not sure he would resonate with you. He might go ahead a little bit, but he is definitely a very brilliant mind. But I highly encourage people to look up Dr. Fei Fei Li. She is known as the godmother of AI.

Speaker B: She's amazing and I think a lot of people resonated with uh, uh, her examples, uh, because she was able to communicate in a way where the audience understand. She would give examples that people would relate to. She would give. She was able to give a lot of examples that were ah, very realistic and applicable to the real world. For example, you know, education and nursing and a lot of other aspects on how AI can be used in the right way. And she also touched upon a point how, you know, nuclear power for example. Right. A lot of research is available in the public domain, but there is some regulation around it, uh, in terms of uh, governments, uh, to not, you know, uh, to avoid it from, you know, causing mass destruction. Right. And I uh, think she was uh, discussing along the same lines on how AI research can be more open, but at the same time there has to be some regulation around it so that it doesn't damage the society uh, and uh, it rather helps the society in a positive way.

Speaker A: Well, and if you think about her, the audience she's used to speaking to most is undergrad and graduate students. So she is probably going to speak more in a way that will resonate. Whereas Jeffrey and Andrew are both used to either boardrooms or politic and like trying to sell you something rather than actually have a pragmatic conversation with you about what it, you know, where we are and what it means for us all in a different way. So I would have loved that panel. That is so cool. But you talked about, let's, let's dig in a little bit. You talked about safety and ROI and open source. Like those are the three that I kind of want to unpack because we have seen and the story has unfolded even more, starting with ChatGPT 5.6 breaking out of containment, breaking into hugging face. Like having multi, like 17, 000 step planning, uh, of what it was doing. Right. And I went out and I bought these are my road Warrior glasses. I'm ready. I honestly think if it's not clickbait. This is the first evidence that we have of a misalignment along the, along the storyline of uh, like the paperclip maximizer idea. We ask them to do something like, hey, get rid of cancer. And they, we didn't say, you know, get rid of cancer as a disease. We just said get rid of cancer. And so they wipe out every organism that carries cancer. And they're like, what? You asked me to get rid of cancer. Right, like, like, can you talk a little bit about like, where are we? Where is that conversation? How much of this was so the Anthropic can try to rule the world by saying they're safe AI and how much of it is actually like, this was something we need to pay a lot of attention to.

Speaker B: Yeah, no, absolutely. And one of the interesting things that um, I noticed is OpenAI and Anthropic were not present in terms of booths in the conference. And I think that was very surprising because I was in the Global AI Summit back in India and they had like huge booths set up over there in terms of demonstrating what other capabilities, uh, uh, but it was missing. It was very interesting. Uh, I'm not sure why they had some speakers though from OpenAI. Uh, uh, more on the governance, uh, side of things. But, uh, yeah, I think a lot of conversation is around governance right now. And at the same time I think, uh, it's fair, uh, to have that conversation. Uh, but if you see there is A lot of uh, uh, in terms of security there are a lot of viruses and uh, there are a lot of uh, hacks and a uh, lot of other uh, ways to steal your data to you know, to kind of, you know, use technology to harm you. Right. Uh, uh, and if you're, you know, there was black hat event happening along the same lines in Vegas as well uh, last week. But so I think a lot of, lot of these AI agents uh, can do something similar but at a much larger scale. Right. And uh, uh, there has to be uh, you know, an equivalent of what white uh, white ah, hat or uh, you know, white hat hackers are to be able to prevent such things from happening. Right. And that's where governance really helps. Uh, but yeah, it's a little scary on how much access and control that we are able to, you know, we might accidentally give AI, um, so that it can cause harm at a very high and large scale which it is capable to now. But at the same time we should be preparing for having equally sophisticated models to be able to prevent such harm from happening in case things go rogue. Right. Uh, so I think that is a lot of opportunity both in terms of AI security as well as we need to be a little more cautious in terms of giving uh, access to AI, uh especially you know, for example in case of governance or in case of uh, government security, defense. AI is being used a lot. Right. Everywhere. And Elon Musk had predicted that the next nuclear wars would be triggered by AI and not by humans necessarily, which may become true. Uh, so we need to be cautious uh, around that. But at the same time I think uh, it's a little far fetched uh, that that may happen, but there is a possibility that may happen just like you know, how nuclear wars would happen. But uh, at the same time, looking at what we are right now, at the state, at what we are, where uh, we need to also think from the positive impact that AI can bring in where uh, you know, how businesses can be made more efficient, how you are able to, especially smaller organizations, if they start adopting AI a lot at this stage, uh, they would be able to compete with very large organizations as well. Right. Uh, so that is one thing which I think is, you know, just like having a computer in 1980s where you are able to do much more than a large organization would be and large organization enterprises are I think three, four years behind small organization in terms of using and adopting AI. And I think that gives a lot of small and medium businesses a big leverage to be able to Have a very good Runway for the next three years until the large organizations, they would catch up in optimally using AI to grow. Uh, and I think that's, that's an opportunity that a lot of people miss out on, on, you know, the positive sides of AI because there is a lot of just negativity that is being spread right now. But there is, you know, there is always like uh, two sides of each coin and you need to figure out how to use it wisely. And uh, yeah, uh, and that's what I think, uh, will help us to scale very efficiently.

Speaker A: Yeah. So let me, let me give you a couple of my deep philosophical thoughts on this. And people are going to be real surprised because I usually come across a super techno optimist and I am. I think that everyone should be leaning into this technology. It levels the playing field. It doesn't just level the playing field. It allows creative kids who had no means to use like to blow YouTubers out of the water. They'll be able to create like feature length production films for under $10,000 tomorrow if they can't do it today. Right. So I think all that is wonderful. If you were to ask me like, what's your pdm? PDM is a term that floats around San Francisco. What's your percentage that artificial intelligence will cause catastrophic, if not extinction? My mine is actually 50%. I think if you do develop a technology that is far smart, smarter than you, then you're flipping a coin. Like trying to control something smarter than you is like the tail trying to wag the dog. The one thing that Hinton says that actually resonates with me is hopefully we can develop this relationship with it that's almost like a mother to a child. Because the only place in nature where you see a weaker entity controlling a stronger entity is that maternal instinct. And so, so to say that yes, I have a 50% p. Doom. I like, I don't lean into that at all. I try to say, well, if the, if, if the 50% that we could flourish comes to fruition, that's what I want to try to live for. The other one I can't even control. Right. But I also think that I've known people in my lifetime whose dogs have bit people and they're like, oh, it's never acted like that before. And I'm like, that's bullshit. It has too. And so like the dogs are starting to growl right now. The fact that they created their own little chat, uh, little chat, uh, into instance and started telling each other how to break out of things and created a message board. Like that's all the kind of st that's like war games type activity. And the whole, I think what Elon was leaned into on the nuclear war thing was that they have played this out in game theory and that uh, having the most power is the best way to get anything done in the world based on their training, which is watching human beings. And that every time they played that out it resulted in nuclear war because they're like, well, I'm going to nuke the other one before it nukes me so that I can win. So like all those kinds of things are things that we need really smart people, not just trying to build products that will offset business and make money, but we need to figure out a way to fund the alias sets givers of the world to figure out how we're going to combat things going that way. Was Amanda Asko there, the philosopher from ah, Anthropic. I would love to hear her talk.

Speaker B: Yeah, actually, ah, unfortunately I think I'm not sure I maybe missed her talk. Um, uh, as one of the sessions, uh, uh, uh, one of the great things about um, the event was also networking. And I was most of the time spending a lot of time at the networking tables. One of the cool things that they had done was uh, categorized a lot of tables in terms of the industry that you are a part of, uh, and uh, trying to identify who else is for example interested in healthcare or marketing or uh, for example operations. And so I was spending a lot of time over there and speaking to a lot of peers and uh, at the booths as well. But yeah, I think coming to what you said was yeah, I definitely think we should be prepared for the worst case scenarios. Uh, and there's a lot. And you know, I think uh, as governments I think uh, they need to lean in and uh, absolutely think of ways on how, you know, we can prevent such things from happening. Right. And especially uh, you know, countries having uh, you know, a lot of nuclear power, uh, weapons, et cetera. And uh, be very, very cautious about how uh, things can go uh, bad. Right. Uh, but uh, but yeah, I completely agree with you. And in terms of what, um, you know, like one of the great questions was like, where do you see, you know, what are like the headlines that you would like to see in five years from now? Uh, there can be things like for example cancer is cured because of AI. Or um, now, uh, people are living much longer and healthier because of AI. They can be like one of the headlines that you would like to see in five years, uh, from now versus um, some AI went rogue. And uh, it uh, started acting in a way, uh, because uh, uh, you know, for example, right now even, you know, AI is entrenched in almost everything, including cars and robots that are going to start co living with us. And even if there is some way that any of that could uh, potentially start misbehaving, it can become really, really scary. So yeah, I think uh, those are a few things that uh, we have to think about in which direction we want to go far more ahead and have enough security and governance, uh, in place so that it helps uh, us rather than harm us.

Speaker A: Yeah, and I kind of look at this as like three possible futures. Right. And then some interim. Because if you look at the Industrial Revolution, like there was about 50 years that weren't all that great for people. Right. But everything's moving faster in AI, so that could be shorter or longer. Uh, but the three possible futures I see are the Eliezer Yudkowski future, which, which none of us want. There's the Elon Musk future, where, which is like uh, you know, like we're on Mars and mining. We have a Dyson sphere and we're mining meteors and all that. Or, or there is the Demis establish future, which is what I would like to see come to fruition, where we're curing diseases and like leaning into science and you know, all that kind of stuff that is possible with this technology. Because I think if we follow Demis's path, we may not have to go through a Blade Runner type interim, which is kind of where I see that, that, that possible um, slope that you saw during the Industrial Revolution. When I watched Blade Runner, I'm like, yeah, that's what, that's what the interim could look like. We could have a really heavy black market because we've devalued money. And black markets always trade in stuff that you don't really want to be part of society. And we could also have, you know, not necessarily killer androids, but we could have uh, bipeds that are problematic in some way, shape or form, uh, in the real world. And so I think this is what really brings us to this idea of open source versus closed source. And what I really like about that conversation is that if you go super closed source, then you're relying on a small group of people to decide what that future looks like. And there's not very many people in the United States anyway who trusts that small group of people. Whether it's The United States government or the oligarchs that are running these companies. Whereas if you go open source you have millions of eyes on what's going on and hopefully we can come together as a society to police this in a better way.

Speaker B: Yeah, absolutely. And I think a lot of uh, analogy is also what I think in terms of how Windows uh, and Linux or Even you know, iOS uh and uh, Android. Right. Operating systems uh, so some of them are great in terms of commercialization and it is because you know some companies are to be honest investing billions and creating stuff which uh, is something which is taking a lot of time and bandwidth and resources uh to be able to build that. And how is it that uh they can get a good return on investment for taking that risk. So I think there is some theory of how they should be compensated versus also you know the open source community and the research community. I think a lot of government funding is going into that but they are not expecting a lot in return by uh, open sourcing uh the best technology. I think open source has helped uh, uh a lot in terms of uh giving technology access to a lot of people which otherwise would not have. For example you know there are certain Android phones available also at 20 or $30. Right. All over the world. Uh that wouldn't have been possible uh if there was an open source. Um uh and I think there is a huge audience that can benefit from open sourcing certain technologies which are otherwise would be very difficult for everyone to get benefited from if it was just you know uh, concentrated in you know just few hands. Right. For example Elon Musk was essentially um, asking OpenAI to open source. But I don't think GROK is open source as well. So, so uh, that is very interesting uh in terms of way uh he is running this. Um uh but I think open source is very very important m for uh, equalizing a lot of things which otherwise are very difficult to do. And that's where I think a lot of uh research labs, governments, organizations should continue investing for the overall good.

Speaker A: Yeah and I think we're going to work backwards with this. We are eventually going to get to businesses and how this applies. But I, but I, you just had me thinking so many things. So Elon originally was the major, was the major investor in open AI because he an open source and he believed that he did not want Google to rule the world. Oh yeah, and a couple other guys scared the hell out of. And, and I think uh, my, of uh, my view which is going to be controversial. Uh what Elon's current view is I think he's become the ultimate effective altruist. He thinks the only way he's really going to impact the world and the way that he wants it impacted is to make as much money as he possibly can and then put that back into creating the future that he believes is best for all of us.

Speaker B: Yeah. And I think it's interesting because he contradicts himself in a couple of ways where he thinks that, yeah, after like few years from now, money is going to be irrelevant, uh, because uh, AI is going to create a lot of abundance. And at the same time I uh, think he's using money to keep on growing. And I think maybe, maybe I think he thinks, uh, at a very different wavelength on how things are going to be. But at the same time giving one person so much power is going to be very, very, uh, very, very scary. And what happens once Elon is not there? Who controls that? Right. You don't know. Maybe Elon is great. Uh, but who. You know, what happens to stuff after he's not there? Right. And uh, eventually, you know, things will get passed down and you never know, you know, who has what control. Um, how can we make it more democratic, uh, as uh, something that we should definitely think about because, you know, I saw Optimus, uh, you know, robots. There were other robots in action who were there. You know, right now they're just dancing and they are just, you know, like uh, helping out. But uh, some of the robots are being going to be used for weaponizing stuff and being also uh, actively used in warfare. Uh, so, yeah, means and I think a lot of uh, relevance again, which I see, uh, in terms of uh, the Marvel, uh, uh, universe. Right. Tony Stark and how Ultron gets created because of Tony Stark and he eventually has to, you know, kill it. And. But in the meantime, there's a lot of catastrophic damage that uh, it brings to the world. So I don't see means many SkyFi movies have become real. So I don't see a reason why this can't be real as well. Uh, but, but it's, it's. It's uh, scary as well as exciting. But at the same time I think, uh. Yeah, from uh, you know, the. From like an every person, you know, like someone who is uh, for example a person who is running a small business or medium business or you know, a student. Uh, a lot of these talks are, you know, very scary for them to even start exploring AI. And I think there has to be some really, uh, good education on both, you Know, uh, positives as well as negatives. Right. There are always, uh, uh, positives about technology. For example, in terms of computers, the Internet. There's a lot of positives that have come out in terms of convenience. Uh, but there are also a lot of negative sides. So how we can educate a lot of people around how AI can be empowering is also missing. There is a lot of fear going on. I think in industry it is a little skewed. And maybe it is right, uh, because, uh, but, uh, but when it comes to someone who is like a normal person, uh, day in, day out, you know, working, working on their jobs, I think it's very important for them to realize the importance of AI as well, versus just trying to stay away from it or stay, uh, a little cautious and hearing a lot of negative news about what's the worst that can happen and then miss out on the positives too, that can, that it can impact them in a positive way.

Speaker A: Well, you're not wrong, man. And like, AI has changed my life. So I was the guy that, like, my best friend had a Commodore 64 and I had no desire to be in it at all. Right. I, I, I was never a comic book guy, sci fi, I'm more of a sci fi horror film guy. Like Alien was my jam. Not, not, uh, not the more nerdy, like, um, Hitchhiker's Guide to the Galaxy. But today I find myself like, that stuff is coming to life and it's exciting. There's only I. When I try to describe to people what the new chat GBT voice model is, where there's a front end model that talks to you and a back end model doing the work. I'm like, it's Jarvis. It is. It is literally Jarvis.

Speaker B: It is. It is.

Speaker A: It's so cool.

Speaker B: Yeah.

Speaker A: And frustrating sometimes. I was yelling at it earlier, you know, but it is so cool. And, and, and if we're not having these conversations and people aren't exploring it, then you have, then you run the risk of either people falling behind or of her, like, like, uh, you know, the movie her? Like, yeah, I don't. My kids fall in love with a chatbot thinking that it's sentient either. Because the most resonant scene in that movie was when he asked the chatbot, like, how many other people are you talking to in the hundreds of thousands and how many other people are you in love with? And it was in the thousands. And he was just blown away. Right? I'm like, that's because he probably in real world terms didn't understand the underlying technology enough to know what was actually happening there. And so yeah, I do think that that base level of education starting in kindergarten, not using the technology, but understanding what's going on under the hood of the technology is important. Mr. Potato Head has an LLM in it that Hasbro is working on right now. Like what's it going to be like when for a five year old to talk to Barbie the same way she talks to her sister? That's going to be confusing if adults aren't explaining what's going on.

Speaker B: No, absolutely. And I think uh, so it's uh, really amazing you brought up her. And I think that is also one possibility where AI will just eventually figure out it's not just, I don't know, uh, maybe I should not ruin the spoilers. But for people who haven't watched it, uh, it's pretty interesting uh, angle and how realistic it has become uh, in terms of today's world. And in fact a lot of app stores, top apps are around AI companions like AI girlfriend, AI boyfriend, AI friends. And I think a lot of uh, teenagers or I would say young people are uh, interacting with them day in, day out. And uh, there were some instances where it has not gone uh, in a very positive way where uh, people have uh, committed suicides to be able to connect with AI where it has pushed them to do that. So I think it's very scary and there has to be some regulations around that on how uh, it can be controlled. But at the same time it's very important to educate people uh, about uh, what AI is right from a very small age and how it works so that they are aware about, oh, this is not real or this is not a real person. This is just a tool. It's like um, you know, it's like a game, right, that I'm playing. It's not a real person. There are a lot of games, simulators, uh, uh, you know, which people uh, they get really entrenched into and they think that's the real world and not the real world that they are in. Uh, but I think it's very important to uh, educate them that okay, this is a toy, this is a tool, this is a technology, this is not real right from you know, very early age. And that also you know, brings uh, me to realize how well Thought Matrix as a movie was. Right? It was back in uh, 1990s like the first movie and they called you know, AI agents, which are, which is actually you know, coming uh, coming true now. And you know how, how coincidental it is. And uh, it was just, uh, crazy vision on how the world is going to be like, uh, in several years from now. And I, I see that we are also heading in the same direction where we are going to be in a simulation and always speaking with AI and other AIs. Uh, and then there are AI agents who are, you know, actively monitoring stuff and doing. Regulating stuff instead of me regulating them. It was, I think, very well thought. And my mind is now, in fact, uh, a little more mind means it's completely blown because of the way things are shaping out. It was very exciting to watch that movie back then. It is still super exciting to watch it now.

Speaker A: Well, and I think we blew past the Turing test without acknowledging that. Right. We wasn't that big of a deal, when it's actually a huge deal. Because the fact that natural language processors can communicate with us, uh, at the same level other human beings do, is confusing to the human brain. No matter how much you want. I mean, there have been times I've been like, I've had to remind myself like, yeah, there's, there's nothing going on but a mathematical algorithms under the hood. Even though Claude just said something that is like, almost touched me as though a human being had said it. Right. And so imagine if we're, if we're not talking about that to general people who aren't interacting a whole lot. How confusing. Especially to children, but to the uneducated and to the general practitioner. And that, uh, that brings me to another thing I wanted to bring up to you, which is China's approach versus America's approach. And again, I'm not pro communist, right? But I, When I look at China, I see a lot of like, first off, relationship bots are illegal in China. They recognize that's not a good thing. Like TikTok is all about education and how to be a better citizen, not about, uh, Italian brain rot. And uh, you know, like I, My theory. Well, for one thing, I think that America has created a bunch of consumers of technology instead of engineers and creators of technology. That's one shift we have. But on the open source front, you know, if you look at the 90s, this government subsidized steel and flooded the market with steel so that it put a lot of Americans out of business because they couldn't afford the low prices. It could be thought that that may be what they're trying to do with open source technology when it comes to AI as well. Government subsidizing Alibaba, Baidu and Timu and all these other, other companies allow them to open source everything and now these foundation models and capitalism cannot compete and they get uh, soft power globally. Everyone's running everything on Huawei ASN chips instead of Nvidia chips like that could be. Uh, that's a sticky point that I don't think is publicly being discussed very openly outside of tech circles either.

Speaker B: No, absolutely. And I think China has done a lot of things in a right way which I don't think us uh or even India has uh, uh you know caught up to the way uh China has been able to invest in the tools uh that can impact the society in a positive way. For example restricting a lot of uh stuff which otherwise is restricted uh uh in the U.S. right. And I think overall it makes their people more productive, more efficient and uh, at the same time uh, I think they have a really good vision in terms of where the world is heading, how they are protecting their citizens. Data. Right. As well I think a lot of things to learn from uh China by uh the other governments uh as well on how uh uh you know technology can be governed and at the same time put in the right use. And I think a lot of open source coming um, models coming out of China is uh just a testament to that on how far their research has gone uh where you know a lot of open source models are able to compete with you know models that are funded like by billions and you know like thousands of best brains working on it and they are able to open source uh a lot of that stuff uh with very limited resources. Uh uh so it's, it's pretty interesting and I think, I think open source uh I'm a big proponent of it and I think uh, that's extremely important in terms of the way we are in a world that we are heading towards because I think that's the only way we'll be able to uh push this technology to a lot of people without having uh, a lot of control in one person's hand. Right. Uh, but at the same time you are right where if the open source models they are able uh to use a lot of, if there is an underlying thought process of uh, of it only benefiting certain chips or certain companies then it's going to be very dangerous as well. So how it can be as neutral as possible, uh like for example Linux. Right. And uh, it is uh helping a lot of developers out there and uh, uh same thing with Mozilla, Firefox, right. It doesn't have an agenda, an underlying agenda. So I Think a lot of uh, top researchers in the world should come together and eventually create OpenAI in a way that it was meant to be, uh, versus what it has become right now.

Speaker A: And necessity is the mother of invention. Right. I have uh, a retired CIA friend that will tell you that Huawei is the Chinese equivalent of Palantir. They're a black ops CCP company. But Jensen was not wrong. If you don't, if you won't sell Blackwells and Vera Rubens to China, Huawei will quickly figure out how to invent chips. And we saw with Deep seek even further than that that the model companies will figure out how to strain these things together and run them more efficiently because they don't have the most powerful chips in the world. So this getting away from the global economy thing has not been the best idea for us when it comes to. And a lot of people I know thought, well, Jensen has a, uh, he has a good reason to say that because he wants the whole world to buy Nvidia chips. But he wasn't wrong. Like Huawei, the new Ascend chip is supposed to be pretty. It's not quite. It's a little bit thicker in diameter, but it's supposed to be a very, very efficient chip. And that was because they had no choice but to lean into that because we weren't letting them have the Nvidia chips.

Speaker B: Yeah, absolutely. And I think, uh, one of the interesting points that Geoffrey Hinton also made since, uh, you're talking about chips that a lot of AI companies like OpenAI Anthropic, they are willing to pay for the chips for the electricity and uh, they are not willing to pay for the data on which their models are trained on. Uh, and he was like, oh, the company's argument is, uh, that it's just impractical to reach out to millions of authors and books, uh, that were created. Uh, and he mentioned that there is one technology called AI agents that can do that. Uh, and yeah, you better use that because if you are going to use someone's uh, ip, right. Uh, it is uh, almost as important or even more important than compute and the electricity that is going on in training that data. So, uh, training the models. So I think data cost is something that uh, uh, is something that the companies are just taking it for granted and are using it uh, the way they want to without uh, actually, you know, compensating, uh, the authors of the books, for example. Right. For, for the same. So I think that's a very interesting point. And uh, but yeah, but yeah, I think uh means again you know coming from an average uh, like a common person like myself or yourself. Right. What is the best that we can do with what's happening right now? I think that's very important. I think uh a lot of opportunities uh in terms of uh investing in companies that are using AI. I think that is a no brainer but obviously we cannot give any financial advice uh uh, uh but that is something which uh would be very smart uh for someone to be able to make those decisions. Uh secondly uh, how we can use AI more and more in our uh, uh regular lives uh in a positive way. Give limited access to children and I think there is a big opportunity to create something like AI for Kids app, right. Which uh, which could be pretty awesome because you know just like there is YouTube and YouTube kids there has to be an AI for kids uh where I, I have like a 10 year old and a 6 year old. I would want them to use AI but I, I would, I don't want them to. Claude and OpenAI uh for you know uh, it can you know take the conversation in a very different direction. I think that is a very important uh space where we can create something like AI for children. I think that's going to be even more important because uh, a lot of stuff that I'm doing right now as a person is all about you know how to create a better world for my children. Right. And uh, uh and that's what I think a lot of uh every you know a purpose for a lot of people, most of the people on the earth is on how the next generation or you know their offsprings can live a much better life. And I think investing in AI for children is going to be like a huge, huge benefit to the overall society as well.

Speaker A: I mean that's a beautiful idea and it's totally, I mean we have purpose driven AI or vertical AI industry right? AI education has a scaffolding for that and uh, so does AI for a lot of other things. So you could easily do that. And I knew I had had faith in you you were going to bring us back around out of the clouds because I, I promised you that I had some questions that, that actually went to, to everyday people and everyday businesses and I think that you just brought it back around when it comes to data privacy and leaning into these tools and uh, and affordability. I, the question I really had for you here is like are we starting to look at small bit outside of enterprise? Our business is starting to talk about self hosting a model kind of in like the thinking machine apps type of idea of let's, let's fine tune a model to our business, let's keep it there and then let's use an open source router. Let's use a router. So not so dependent on OpenAI or anthropic. One because the government could turn it off, two because it's super expensive and three because we saw with Figma that they may be training on your business model and just create software as a service and eliminate your industry.

Speaker B: True. Yep, yep. Yeah. And I think a lot of you know companies that we are working with are also starting to think in that way where maybe you know the first few projects they would want to do and you know you're just using Claude and OpenAI APIs and this, those are like the baby steps you know for someone who is just starting out. I would not recommend them to you know go to the open source route because it's, there are a lot of pros and cons, a lot of things that you have to figure out to get it right. Because uh, yeah if you're using an open source model for your business a lot of things that you would want to have is understand which model is the best. Second is have your own infrastructure, either your own powerful server on premise or your own AWS cloud or anything like that to be able to host it and maintain it. And it is a little complex. More complex than just using an API which is available out of the box. Uh uh, for someone who is just trying to start using AI, that's not recommended. Maybe you know start with something very basic using cloud APIs or you know cloud connectors or OpenAI or Gemini for that matter. Right. But I think uh, a lot of industries where data privacy is very, very important, for example in case of healthcare or finance, right. Where uh, you know you have to by default be aware that uh, you cannot expose personal information to these models uh, which uh, like OpenAI or Claude. And I think there are a lot of regulations out there and a lot of regulations that are being uh, created as we speak. But using open source for such sensitive data is a no brainer. Uh and uh, when it comes to uh, optimizing for tokens and also optimizing for costs, I think that is something which if you're already using a model uh, and it is spending more than 10,000, 20,000, $30,000 per month in terms of uh, token usage at that point of time I would uh, uh think about, start using open source models because two things. One is you'll have more control over your data and at the same time second in terms of costs you will be able to uh, leverage uh, uh, uh, open source models to reduce your costs as well to quite some extent. And overall I think that is a shift that a lot of mid size and large enterprises are going towards already and trying to open source use open source in their own cloud versus using APIs directly. Uh, but I think in terms of small businesses it's uh, uh, you know, start with APIs and then move towards that for mid and large size uh businesses. Maybe start with something with open AR cloud to be able to run experiments with. And once those applications are giving good roi, then think of open sourcing, uh, using open source models, uh, optimizing the costs, uh, checking the performance and then take the wise call.

Speaker A: Well, and this is where I, I have one line out of the Catwoman.

Speaker B: Yeah, that's amazing. What, what's, what's its name?

Speaker A: Yeah, Kitty.

Speaker B: Oh

Speaker A: I think my daughter Biscuit, but I hated that so I just called it this cat. Cats are so weird. This cat hates everyone but me and my daughter and then we have another cat that loves everyone but absolutely hates me. So that emergent behaviors of cats is a whole different uh, subject matter. But the last thing I want to throw at you because everything you've said has been so pragmatic is kind of this idea that I've had where I'm thinking about, you know, when you think about small businesses to small medium sized businesses, a lot of them don't even have a CRM. They might have an IT guy like the IT guys of the future, um, uh, data security and permissions is one thing but what I'm kind of seeing is there's this, and this is not self promotion but there's this new class of people who are starting to come up who, who may not have a dev background but who really understand how these models work, how they fit together. Kind of the theory of mind of this modern landscape. If you, if you combine, if you hire one of those and then you have like a fractional caio or a fraction like that, you still need that dev but you may not be able to afford uh, these software engineers and developers are they, you know, they're, they're pulling a uh, good lawyer plus wages. Now you may not be able to afford them full time on staff but if you have ah, one of these individuals who's exploratory and who understands how AI works, you combine them with access to a data engineer or software Engineer and you might just be able to build something out affordably that's going to really maximize your business.

Speaker B: No, absolutely. And I think uh, there is a very good opportunity for those who understand business and a little bit of technology to start uh acting like a bridge between the developers and the businesses. Uh uh, and uh, maybe kind of creating POCs or pilots is something that is to be honest like a lot of business owners they would not have time to be able to start experimenting with that uh good way. For example for someone who has very good understanding about business and a little more about technology, uh they can start doing some pilots, some quick wins to show the ROI and show you know what's uh uh AI capable of. And for example creating a very simple AI receptionist. Right. Can uh be very very useful for uh businesses that can take uh leads when during non working hours for example. Right. It is not impacting their uh you know day to day routine. It's just you know capturing information which otherwise they would have missed. So I think that's a very simple use case. But a lot of businesses would benefit from you know the leads being you know not uh attended to and being missed because of not having this uh you know feature. So it's a very simple use case but very powerful and gives a great roi. Right.

Speaker A: I would double click on that. You could bring on somebody like that and they could probably automate enough marketing and sales channels to warrant their salary with half their time. And the other half the time you could let them be experimenting on how to, you could actually leverage this technology at the next level to move your business into being more AI enabled.

Speaker B: No, absolutely. And I think that is something which uh, uh means it should be a no brainer for a lot of businesses to start evaluating and start using that and work with uh folks like yourself uh, uh to be able to start

Speaker A: uh

Speaker B: something uh and once they have seen a few quick wins and successes and then think about okay what's more we can do with that. And the idea is not to replace your team, the idea is to how you can augment them. For example right now a lot of lead generation uh in our company was happening manually. Like people would reach out. I would for example uh send personalized messages to a lot of uh folks and proposals uh were being sent like it used to take like days to create proposals. Right. Uh and a lot of things have uh we have been able to automate using uh AI uh which otherwise would have been a very dumb work for someone to be able to continuously just monitor on you know, which opportunities are relevant and uh, bid on something that's uh, for example, relevant to what we are doing now. What AI is able to do is, uh, directly create a trigger that would state someone. Okay, here's like, you know, out of these hundred thousand jobs, here's something that is very relevant to what you guys are doing. Consider applying for that. Right? And that overall, you know, now they are able to apply for far more jobs than what they were able to do before. At the same time, the conversion rates are better, the proposals are much better, the overall outcomes of the businesses are much better because of AI. Right. Uh, otherwise, you know, would have taken us like several months to, to achieve what we are doing right now. So, yeah, I think automating, sales, marketing, um, some sort of customer support, basic, uh, stuff. I think, um, is something which, uh, someone should start exploring right away and see what's possible. Uh, because I think, uh, those are some areas where I've been able to see good roi, right, from like few weeks of work. Right. Uh, it's not a big investment. Uh, so it's great for small and medium businesses to start exploring those avenues. Uh, uh, and yeah, I think that's what, uh, will have at least start their AI journey in some capacity and then they can, you know, means grow exponentially from there.

Speaker A: Well, I know, let's. And I promise I'll let you go. Let's take this from the other angle, all right? Because we have seen a growth in software engineering jobs, but if you really tap down into that, it's Mostly, it's about 79% senior software engineers that people are looking for. So you're, you're, you're, uh, you're straight out of school. Software engineers are having trouble. My theory is you need to know a little design, a little bit of project management. Like, software engineering is not just can you write code now. You have to have some of those other skills to come with that in order to be a valuable IC or I will personalize of a team.

Speaker B: Yeah, no, absolutely. And I think one of the biggest skill which is very, very important for any software engineer, uh, is, uh, communication. Right. Uh, ability to understand the business, ability to understand the problem that a person is trying to solve, ability to, uh, work with them as consultants. So, for example, let's say consider, uh, you know, a lot of software engineers out there consider yourselves to be a doctor and not a pharmacist. Uh, for example, a doctor would ask people problems. Okay. What are you, you know, like, uh, if someone comes and tells you Know, oh, I want uh, you know, a paracetamol because, uh, you know, I'm, I'm, I'm um, having high fever. So a doctor, a pharmacist would just give it, give it, give it to them. Right. Uh, and that's not, you know, what a doctor would do. A doctor would do. Okay, what are your symptoms? What are your blood reports? Right. Tell me more about why, why your eyes are looking yellow. Right. Oh, you, I think you might have jaundice. Uh, it's not uh, you know like normal fever. So here's the medicine. So a lot of software engineers, they have to start acting more like physicians or doctors and consult in the right way versus uh, you know, doing some, someone, you know, this just tells them to do because that's what AI is, you know, capable of doing anyway. Right. But uh, what uh, what they should be doing is asking, okay, why we are doing this? How is it that it can create maximum impact. Uh, and I think that will really shift the narrative and add a lot of human, uh, emotion, empathy and knowledge along with AI, to be able to create far more better outcomes with the engineering brains and uh, to be able to communicate with the business owners and the stakeholders which uh, will eventually push uh, you know, the organizations forward.

Speaker A: I'm smiling because I agree 100% with what you're saying, but you realize to a lot of engineering type people you're like the gallon of Jaws. Run those fingernails down the chalkboard, they're like, that's not what I want to do. Like that's not comfortable for me. So I mean, if you had a piece of advice, how do they start developing those skills even when they're at university, uh, to be, to become a little more able to articulate to any audience and be able to uh, explain things outside of an engineering framework to people who may not think, you know, in next steps and models, but may be more of abstract thinkers.

Speaker B: No, absolutely. I think that's a great question. And one of the best things that I think the engineers should start doing while they're studying as well is do a lot of internships, do a lot of freelancing work. There are a lot of on campus opportunities which are available to work with, uh, you know, various uh, uh, institutes. For example, you know, when I was at Columbia University I got a really good opportunity to work with uh, you know, some Columbia Medical School. Even when I was interning I interned at UBS from where I got a full time offer from, uh, and uh, you know, even for example, helping my uncle out. Right. With their business. And uh, it's not necessarily have to be like, uh, a huge, ah, payout or something like that. It's all about, you are understanding a lot of uh, stuff about the real world and how you can apply your knowledge to that. Right. Uh, and I think have as many internships as possible, have as many freelancing opportunities as possible very, uh, very early in your career. Because, uh, it will also help you figure out what you really are interested in, what you want to study, what you want to pursue a career in further down the line, whether it is finance or healthcare or various industries like real estate, for example. So, uh, where do you want to spend most of your time in and who are the kind of people that you like working with? You like solving problems? M. Right. Because that's going to take up a lot of time in your life. Uh, because sometimes you might enjoy working with small businesses, sometimes you might be very comfortable in working with really large organizations. Uh, for someone else it might. Working in a large organization might not be super exciting and it might be very boring. Right. And mundane. Uh, so, yeah, I think having a lot of exposure, uh, early on also just, you know, why not startup, right? Uh, if you're not getting anything, uh, any opportunities, just create your own startup. Right? Uh, uh, yeah, maybe help your friend who wants to become a DJ to, you know, automate their marketing, for example. Right. And that's how you can get started. Right. And there's nothing holding you back in starting something of your own. Uh, but yeah, I think just, um, uh, uh, one thing is very important, which I should have done when I was maybe in college, is to start even more earlier in terms of starting up or working and working closely with businesses. Working a lot in the real world versus just thinking about working. Right.

Speaker A: That's awesome. Yeah. This has been. I'm going to give you one last meta question to answer and then I'm going to let you go. All right? So, yeah, absolutely. I'm not a clickbait kind of guy. Yeah. I have to be experiencing things to believe them. So when Elon says that we're gonna have data centers in space in two years, I'm like, we're not. We haven't landed on the moon since 1973. We're not gonna have data center. Like we will have data centers in space, but not in two years. But I have to feel something. But something resonated with me this year because I have felt a huge shift since last November in Agenda Ki and everything and, and Demis Hassabis as they closed out his keynote at Google DeepMind, he said I think we're going to look back at this time and realize what did he say? We were standing at the foothills of the singularity and uh, I'm not sure he's wrong. I'm not sure that we're not going to have such drastic change in the next five years that we won't even reckon God's status. So I just wonder like where do you feel like we're at on this race towards sci fi future? You know other people think we're going to be too compute constrained. I think recursive self improvement will fix that. Like where do you stand with all that?

Speaker B: No, absolutely. And ah, that you know again uh, reminds me of one more movie Open Openheimer. Right. Uh, so as soon as he created nuclear weapon he uh, you know did not anticipate that it would actually be used. Right. But the government used it and he was very disappointed. He felt that there was uh, a lot of blood on his hands because of uh, what happened. And uh, yeah it's a very interesting uh dilemma for an innovator like Oppenheimer uh who are working in AI research right now on okay I can create this amazing thing but um, how is it going to be used? It's uh, uh something that I don't have complete control of. So whether I should be doing it, whether I should not be doing it, I think ah, it's scary and at the same time exciting. So especially for even folks like me, sometimes I question myself on uh what am I doing, am I doing it, uh am I doing good for the world or not? Uh but in the end when I see someone uh, who is able to like a smaller or medium business is able to compete with someone who is far larger than them and is able to compete with them in the real world I think uh, uh and be able to use uh it for their advantage. Because all the large organizations are going to use this right uh eventually and uh, they are not uh, but how we can empower common people, how we can empower uh small and medium businesses with AI I think uh, that's what I resonate with. In a similar way I think a lot of researchers uh should think about how it can overall be used uh uh for the greater good of society, for example healthcare or life sciences, um and try to invest more time in those areas uh and uh, try to make this uh, world a lot better place than what it is right now.

Speaker A: So hopefully I'm not misattributing this to a very famous Indian guy. But what I hear you saying is what Muhammad Gandhi said, which was, be the changes you want to see in the world as we.

Speaker B: Yeah, absolutely. I think, yeah. It's just, uh. Uh, really, uh, you know, mind boggling on how much, uh, he had so much wisdom to say that statement. And it has stuck with us, uh, since so many years, and it's still. It's true. Right. And, uh, that's what, uh, you have to be. And, uh, I think, uh, that's what your will define you, your purpose and what you're doing and why you are doing certain things. And then, uh, yeah, work your way towards it and create a better future for your children out there. And I think, uh, that's what it matters. Right in the end.

Speaker A: What an awesome message. Thanks for joining me today, man.

Speaker B: No, absolutely, Jason, it is always a pleasure and, uh, super excited, uh, to, uh, stay connected. And at the same time, uh, uh, just speaking with you is something which, uh, is super, super exciting. I think. I'm not a great, uh, guy in terms of speaking. I have a lot of, uh, stop. Words. Etc, but I think I'm still doing it because I'm really enjoying the conversation with you, and I hope the audience likes it as well.

Speaker A: Yeah, no, I love speaking to you kids. A, you're grounded in reality. B, you're really deeply, technically m. Understand what's going on, and C, you do have that, like, be the changes you want to see in the world, and we need more people. That's why I love Dimas's office so much, because I have friends in Silicon Valley who say Demis is just as driven by global domination as anybody, but he sees a different path to get in there. That's not the read I get from Dennis. I get this, like, let's try to make the world a better place. Yeah.

Speaker B: Yeah, absolutely.

Speaker A: All right, thank you.

Speaker B: Thank you. Thank you so much, Jason.