How This Entrepreneur Built A $13 Million Company To Address A Market Need For Billion-Dollar Brands

Forbes

When she worked for Apple, Kalyani Ramadurgam’s job was to block anyone on terrorism watch lists from using Apple Pay. Even at one of the most sophisticated technology companies on the planet, she was shocked that the compliance process—making sure that the company was obeying financial laws and regulations—was still analog and tedious. “It meant reading through not just hundreds, but thousands of pages of documentation,” she says. “Organizations were just throwing bodies at the problem.” In 2023, Ramadurgam founded Kobalt Labs to bring compliance into the machine-learning age. Kobalt’s AI models sort through mountains of documents to help banks vet their business partners, ensuring they’re following the rules, like halting money transfers from sanctioned countries and promptly disclosing security breaches. Her startup has raised $13 million and has more than 20 customers, including fast-growing fintech company Bilt and Celtic Bank.  Read the full story on Forbes:   Subscribe to FORBES: https://www.youtube.com/user/Forbes?sub_confirmation=1  Fuel your success with Forbes. Gain unlimited access to premium journalism, including breaking news, groundbreaking in-depth reported stories, daily digests and more. Plus, members get a front-row seat at members-only events with leading thinkers and doers, access to premium video that can help you get ahead, an ad-light experience, early access to select products including NFT drops and more:  https://account.forbes.com/membership/?utm_source=youtube&utm_medium=display&utm_campaign=growth_non-sub_paid_subscribe_ytdescript  Stay Connected Forbes newsletters: https://newsletters.editorial.forbes.com Forbes on Facebook: http://fb.com/forbes Forbes Video on Twitter: http://www.twitter.com/forbes Forbes Video on Instagram: http://instagram.com/forbes More From Forbes:  http://forbes.com  Forbes covers the intersection of entrepreneurship, wealth, technology, business and lifestyle with a focus on people and success.

Transcript

00:00I sometimes take a little bit of issue with people that say, oh, don't use AI because it

00:03takes up electricity. You could make the same argument for a Google search. And people did

00:07at the time.

00:17Kalyani, thank you so much for being here with us.

00:20Thank you. I'm so excited to be here.

00:21So tell me what's happening at Kobold right now and do it like I'm a 10-year-old child

00:26because you guys are doing AI, financial regulation, very heady stuff. Tell me what's

00:31going on.

00:32So Kobold is basically an AI-powered platform that banks, fintechs, credit unions, folks

00:37in financial institutions are using to automate a lot of their critical risk and compliance

00:42and legal manual work. And so we basically synchronize with all of the regulations. We

00:46process hundreds of thousands of documents all at once and just really cut out a lot

00:51of the grunt work around document processing and the stuff that nobody

00:56really wants to do anymore.

00:58Documents are like my least favorite thing in the world. Tell me about what was the

01:02status quo like before Kobold came along in AI? What was the old way of doing things?

01:06Yeah, the old way of doing things, and honestly, in a lot of ways, still the current way of

01:10doing things is really just a world of paper pushing where there are people who, you know,

01:15in the world of risk and compliance, there's the paper pushing aspect of the job. And then

01:20there's the really critical strategic aspect of the job. And the paper pushing is just

01:24growing and growing. And what it means in real life is that people are sitting down and imagine

01:30reading through like a thousand page document and trying to identify where any of the potential

01:35issues or risks are. And then filling out like a 600 question spreadsheet, identifying where

01:40all the things you need to look for are. Like that kind of work is something that AI can just so

01:44powerfully, like completely accelerate. And then you can just spend your time really focusing on

01:49what the day I find. And then how can I actually use my human expertise to act on that information?

01:55Very cool. And tell me like the kinds of customers, because you have some very big customers,

01:59the kind of customers who have worked with you and kind of how that all, you know, the partnership

02:03goes. Yeah, absolutely. So we work with kind of fintechs and financial institutions of kind of all

02:07sizes. So maybe a fun fintech that a lot of folks might have heard of is Built Rewards, the credit

02:11card where you get points on rent. So they were one of our earliest partners. And they actually now use

02:16COBOL to actually automate a lot of this work behind the scenes. That's so cool. How did you get into

02:21this? Because, you know, getting into the world of finance and regulatory stuff is very wonky and

02:26very specific. Yes. And I know you've never worked in finance, really. How did this come to be?

02:31Yes, it was a bit of a topsy-turvy way to get there, I'd say. So my personal background is I was

02:37an engineer. I studied computer science at Stanford and I was an AI researcher there for a number of years

02:42and throughout grad school. And my research was actually on like, how do you use AI? At the time

02:47they were like, you know, there was no chat GPT. It was like, how do you use these transformer models

02:51to actually, you know, conduct really efficient work on highly sensitive data? So we were working

02:56with like government data, financial healthcare data. And then around the same time, I had a brief

03:01stint at Apple Pay where I was actually on their fraud team. So that was me kind of touching financial

03:06services. You know, I was helping, how do we automate, you know, checking whether somebody's on a

03:10terrorism watch list so they can't use Apple Pay. Some of these are really big problems.

03:15And then around the same time, one of my very close friends from Stanford, Ashi, who's my co-founder and

03:19CTO, she was also working at a fintech, a firm, where she was building a lot of their like scalable

03:24infrastructure as an amazing engineer. And so we kind of teamed up and we just realized that this world

03:30that we came from was just, it was unfathomable how ridiculously archaic a lot of these things were

03:38being done. And that was just in the world of fintech. And now we've, we work with some of

03:42these larger banks that are really using technology that's like over 25 years old. And it's just so

03:48like ripe for disruption. So cool. I saw, I saw in another interview that you did that you and your

03:53co-founder were talking about starting a new company and you applied to Y Combinator and you got in

04:00and you both quit your jobs, but you didn't really know what your company was going to do.

04:05Yeah.

04:05That sounds very terrifying to me, but obviously not to you. Tell me a, how you got into Y Combinator

04:10without having like a clear, I guess, mission and what it's like to kind of take that leap when

04:15you have like, you're working at Apple, very good job. You know, she's at a firm, hot startup, you know,

04:19fintech, billion dollar company. And just to make that plunge without really knowing where you're

04:23going.

04:24Yeah, totally. I think that kind of is the ethos behind why she and I get along together so well,

04:29because we're both the type of person that got excited by that. So the context is I, I was never

04:34kind of completely satisfied working as a software engineer and kind of working at, I guess, a regular

04:40job. I think I was, I would get bored really easily. And I always knew that I wanted to do a

04:46little bit of everyone's job. And that's the only way that my mind is, it feels like it's reaching

04:50its fullest potential. And so I was actually, I'd left Apple and I was kind of an early engineer at

04:55this company called Cenesys, which is a healthcare company that uses AI. So not in financial services,

05:01but faced a lot of the similar problems that I see in my world today. And so Ashi and I both just

05:06knew that we, we wanted to take the plunge. And so we decided to apply to Y Combinator. We,

05:12we didn't have an idea per se, but you have to apply with an idea these days. So we were,

05:16we spent a couple months just after work every single day, we would like make random side projects

05:21and brainstorm and talk to people. And then we eventually were accepted under the almost

05:27agreement that we were still trying to figure out what we wanted to work on.

05:30Wow. So you got past that. I don't have an idea. And give me an example of some of the ideas you

05:36were thinking of that didn't make the-

05:37Yeah. All over the place. Like all the way from, we were working on initially like this AI security

05:43idea, which is a little bit tangential to what we're doing now, but maybe the most out of their

05:48one, which was the primary one that we applied to Y Combinator with was we were building a, what,

05:53what I would call almost like a sub stack, but for authors that are writing like long form novels,

05:58it's like, how can you actually, how can you actually almost like a Netflix for books,

06:02we were calling it like where authors could actually release chapter by chapter. You could

06:06subscribe to them. Everybody would get the chapter drop on Friday.

06:09A serialized monologue.

06:10Exactly. And so we talked to a lot of folks that were old in the serial fiction world,

06:14and we got some authors involved and we, we actually built the product and we, it was a,

06:19it was a very humbling journey understanding like the pros and cons of that business model.

06:24Actually around the same time, sub stack had tried to do the same thing and we both ended up

06:28shutting that down at the same time. I think everybody realized like the pitfalls with that

06:32business model. Let's go back to with, so you, you studied AI at Stanford, but you said it's

06:37interesting because there, I feel like when chat GDP came around, it was like an inflection point.

06:41Like suddenly everyone was talking about AI. You were doing it before it was like the hottest thing

06:46going on. Right. What got you interested in it beforehand, before like the, the gender of AI,

06:52the chatbots, that sort of thing. Yeah. Yeah. That's a good question. I mean,

06:54it's funny that you say that it was this inflection boom, I think from a consumer standpoint,

06:58that's totally sure, but we were, but we were all you, like everybody who doesn't realize they were

07:04using it has been using it for like over a decade now, actually, you know, but I think in terms of what

07:09brought me excitement to it, I mean, I was, you know, in my freshman year at Stanford and even

07:14before that, I think I've always just been super excited by technology. I have my parents are both

07:18scientists and like, they're very like innovative people. And so I've always been interested in like

07:23the new thing and like understanding how things are going to change. And so around that time, a lot

07:28of around like the 2016, 2017 era was another big, almost like improvement in a lot of like AI models

07:36and a lot of the research going on. And so you could do, you could do crazy things then. And I

07:40was super excited by it. Like you could start manipulating images way, way worse than you can

07:44today, but all that stuff was starting to really be possible. And so I just like, even in high school,

07:50I was doing research on that stuff, like doing little projects, like applying to Stanford with that

07:54stuff. And so I was always just super intrigued by it. Our high school experiences were very

07:58different. Um, in turn, I wanted to go back to starting Cobalt because you said you applied to

08:05Y Combinator with a promise that you'd figure out an idea eventually. How did you land on Cobalt?

08:11Yeah, this was a very, very long process. So it was not within the confined three months of

08:16Y Combinator. We applied to YC with the serial fiction idea, actually. That was what our interview was

08:21about. And that's what our, what our partners talked to us about. And they were like, Hey,

08:25we think you're probably going to change your idea, but we think you're smart and you're just going to

08:28figure it out. So that was kind of the premise there. And so then we kind of realized that for

08:33a lot of reasons, you know, that idea had flaws and we wanted to work on something that we felt

08:36more conviction in. And so we kind of thought about, okay, what were problems that we see in our jobs,

08:41in our, in our daily life? Like, how can we have like empathy for a problem we want to solve?

08:45And so at my previous company, we had some issues trying to use AI models on really sensitive healthcare data.

08:50And so we thought, okay, there must be something we can do to unlock the ability to use like open AI

08:56or other kind of foundational models on like sensitive data. So we built the product to help

09:00companies with that. And we worked on that for three months during Y Combinator, like we got revenue,

09:05we raised a seed run off a bit. And then after all of that was kind of done, we again realized that

09:11we had built a product without doing the proper like diligence and like really understanding

09:16deeply that it was a problem people really needed us to solve. So then we spent another

09:20almost four to six months, all the way going back to the drawing board, we scrapped everything,

09:26revenue went back to zero, we told all our investors that we were pivoting.

09:29Okay.

09:29And I just spent, Ashi and I both spent almost six months from like,

09:34yeah, almost six months all the way until the spring, really just talking to folks and digging

09:39deeper and deeper and deeper into what were the tinges of interest in our initial idea and how

09:44can we pull the thread on what people actually needed. So for the initial idea, we had a couple

09:49Finta companies that we had been talking to who told us, hey, you know, this is kind of useful,

09:54but we have bigger fish to fry. And I'd say, hey, will you talk to me about what are the real

09:59problems you're facing? Just give me 20 minutes of your time. Let me ask you anything, pick your brain.

10:03And people were nice enough to say yes. And so after maybe a thousand of those conversations,

10:09almost like a game of hot and cold, this thing emerged where we realized like this fundamental

10:14problem that was foundational to everyone we were talking to. And then by the time, you know,

10:20all these things came together. We were like, okay, we have this idea. Everybody has the same

10:25problem. And we think there's like one solution that can solve it for everyone. That's a business.

10:29I love that. It's just like a myth that people like a founders have like a lightning bolt idea

10:33or they think of something in it. Yes. No, I was looking for the lightning.

10:36You were reporters, you were, you were, you were storm chasing, storm chasing the lightning bolt,

10:40thousands. Yeah. 10 hours a day, just talking to people and talking to people and slowly iterating

10:45the idea until all of a sudden people stopped asking like, when is this conversation done?

10:50And they started asking how much does it cost? I was like, oh, why do you even care how much it

10:54costs? That means there's something there, you know? Wow. Tell me about the mindset required to

10:59build a product, get investors, and then just scrap it and go back to the drawing board and look at

11:04something else. Yes. I think it was a very scary time because we had a little bit, you know,

11:08there were people that somewhat bullied in what we were building. I had been able to convince

11:12investors that this was something worth investing in. So it was very scary initially to just say,

11:17hey, by the way, this thing that I told you with like full passion last week, I believed,

11:21like, I don't believe this anymore. But I think that's also the magic of how smart

11:26like our investors were and even the like Y Combinator community. You know, if something

11:31is eventually not going to be the right thing to do, then there's no point in delaying the inevitable.

11:35So we just said, hey, like we, it's like, it's a tough call because we were excited about this.

11:40But I just know that if I don't say this now, we're going to burn a year's worth of company

11:45money and realize it then. So we might as well cut our losses and like do something that we know can

11:50be like a home run. And it was very scary. But basically, all of our investors said, like,

11:56we're actually really happy you do this. Even in Y Combinator, the majority of the companies I know

12:00from my cohort have pivoted. And it's just a part of the process. Yeah. You know, you're like,

12:04it's like if you're lost and you find that you're walking the wrong direction, like,

12:06you want to turn around as fast as you can. Exactly. The further you go, the more pain

12:10it's going to be. Yeah. Talk to me about the business model and how does Cobalt make money?

12:16Yeah. So Cobalt makes money because banks, fintechs and credit unions pay for licenses to use the

12:20platform. So it's like a SaaS platform that's cloud-based and underneath the hood is like all

12:25of our AI infrastructure. But at the end of the day, it's as simple as, you know, if you're a kind of,

12:30if you work on the risk compliance team at a bank that that's working with Cobalt, you know,

12:34you start your day, you log into Cobalt and you just start doing your work in the platform.

12:38And so we have annual and multi-year contracts. It's super simple.

12:42How much money have you raised?

12:43Uh, so we had raised as part of our seed round with this kind of AI security idea I mentioned,

12:49we'd raised about $1.6 million. And then actually during the summer, we closed our Series A.

12:54So we raised an additional $11 million. So in total, 12.6.

12:58What is your one tip for fundraising effectively?

13:03My one tip is believe what you're saying. I think when we were fundraising for the first idea

13:09that I mentioned we pivoted from, we were able to fundraise. But the questions that people were

13:14asking me to poke holes in the idea, I almost realized that I had to bullshit just a little bit

13:19to figure out how to wrap my head around it and justify the fact that that wasn't going to be a

13:24problem. And so I think we could have realized that it was a bad idea even earlier if I had listened

13:29to my gut and said, Hey, I don't even completely believe what I'm saying. And like, this is going to be

13:34a problem we face. And so now this time, when we fundraise this time, and we're talking to investors,

13:38I knew that I actually believed the answers to the questions that we're asking. And I believed in

13:46how we were going to overcome a lot of the obvious ways you can poke holes in the problem. And I think

13:52the conviction just kind of follows through. It's like you build a good company and you actually

13:56think it's going to work, then people can be convinced that it also will too.

13:59I love that. I never heard that before. That's really cool. Talk to me about how you guys got

14:04started at first. Talk to me how you guys started, because you mentioned your AI, it's ingesting super

14:10complex information documents. And if it makes a mistake, you could be letting criminals get off

14:14the hook. They could be, you could be helping people finance bad things. Totally. Like I,

14:19the starting point of a company always blows my mind. Like, how did you get started? Because this is

14:24complex and high stakes. Yeah, totally. Which I think is also why I like this space. Because I kind of

14:28like, I like the high stakes environment. I kind of thrive under pressure. And I like when things

14:33feel really serious and that it's very important to do things right. I think we honestly, we relied

14:39a lot on some of our early partners that really just said, Hey, we know we're using a product that,

14:43you know, we're one of your early customers and partners. We're going to use it every day. We're

14:46going to know in the early days, we can't rely on it. And we're going to give you very frank feedback

14:52and help you make it better. I think we would be nothing without those initial partners actually

14:56telling us where the product was bad. We said, tell me what's bad and I will make it better.

15:00And then over time, you know, as the last two years have gone by, we've come to a place where,

15:06you know, when, when banks are benchmarking Cobalt, or even when some regulators are testing the

15:10platform, we are the most accurate AI tool in the market for this. Like they, they compare us against

15:15other AI tools. They compare us against even some human, human reviewers that are just getting bored

15:22reviewing this hundred thousand page document. And we've come out on top almost every single time.

15:26And the reason why is because I think we invested a lot of engineering effort into making it like

15:32transparent, citable, doing a lot of work in like showing the explainability around where we're

15:38getting our results from. Like I think without any of that, you can get a bank to trust you because

15:42it's too important to mess up. That's so cool. Did you, like the software, does it run on a bigger

15:47LLM? Is it, does it run on like a Gemini or run on an open AI system or did you like,

15:52how does that work? Yes, we did not train our own foundation models, didn't have the cash for that.

15:56So under the hood, we're using a couple of different foundation models. Like

15:59some of our document processing is powered by Anthropics Cloud and open AI's GPT model.

16:03Those alone wouldn't be able to get us to what we needed to do. So then there's a lot of

16:07kind of custom infrastructure that our engineering team has built in place to

16:11kind of ensure the end-to-end workflow. And most importantly, make sure that it

16:16is actually accurate and work. Yeah. What is the secret sauce for Cobalt? Like

16:19again, what is, because you know, a bank could just maybe go right to the LLMs and do it,

16:24you know, instead to kind of bypass you. What do you, how do you add value? Like what does your

16:28engineering and software add value to those, you know, multi-billion dollar LLMs that are already

16:33out there? Yeah, totally. And I think this goes down to like, what is an AI product at the end of the

16:37day? Because if your product is just something where you upload something and you get results in,

16:41that's the same work style as an LLM that eventually one day will just get better and

16:47better and overcome its pitfalls. Like that's not enough to build a sustainable business.

16:51And so for us, Cobalt is more than just the AI tool. It's an AI enabled like complete platform,

16:56where for example, if we take an example of a workflow that we support, which is

17:00kind of due diligence on third parties, like vendors or fintech partners, et cetera, you know,

17:04you can upload some of that due diligence documentation into an LLM and get some results. But Cobalt is going to

17:10do all of your risk assessments. It's going to reach out, communicate with the right people

17:14using our AI agents. It's going to do other research online, pull all of that back in.

17:18It's going to generate your final reporting and it's going to send those reports to the right

17:22people and manage all the approvals. So it's not just this one-off thing, but it's actually like a

17:27ecosystem of AI agents that are orchestrated to do the full work. So that's where like the product

17:32comes in. I honestly think though, the secret sauce to working in financial services,

17:37having the best and most accurate product is maybe only 50% of the way there. I think like

17:43currency here is honestly trust. And so I think our secret sauce is really the fact that the banks

17:49and fintechs and credit unions we work with, like very publicly trust us. There's regulators that are

17:54piloting us. And I think once a lot of those building blocks are in place, then you just know that,

17:59hey, if I need to look for a tool in this space, am I going to look for one that I have no sense of

18:05their references or how well it's working? Or am I going to look for one that is currently being

18:09battle tested by like the institutions with like the highest, you know, the highest quality that they

18:15need? How did you build that trust? And also how did you kind of make relationships or connections

18:19inside that world? Because you're a new company, you're a young founder, you know, obviously you're

18:25working with older companies with maybe older people. Like how did you get in the door and then how

18:30did you win that trust? Yes. I would say the absolute, the absolute hardest part about any

18:35of this was initially getting in the door. And to some extent, I'm still knocking on doors like

18:39every single day. Part of it is really just relentless, you know, outbound, hitting up people,

18:44going to places. I would sneak into conferences and say, we don't even have the cash for a badge

18:49right now. I would stand outside like, will you talk to me for 10 minutes? Let me pitch you my ideas.

18:54I love that grit. Yeah. So we did some crazy things. And now we're like,

18:58hey, I'm speaking at this conference, but you don't need any of that, to be honest. Like,

19:01you just need to like show up like a thousand times over. And then if you show up a thousand

19:06times, 10% of those times will lead to something. And then you have a hundred opportunities to do

19:10something interesting. So it really was just like knocking on doors. I heard so many no's,

19:16but then you hear one or two yeses. That person takes a chance on you. They give you feedback.

19:21They introduce you to someone else they know. And then it just like blossoms from there.

19:24And I love how your first client was built because we know Encore Jane very well,

19:28the founder, and he's a former under 30 and now a Forbes billionaire. So I love the

19:33under 30 connection. Yeah, I know. That's great. So they were not our first,

19:36they were one of the earliest ones, but yeah, they were one of our like first five clients.

19:39So they've been amazing and they, they took a big chance on us. Like they've given us so much

19:42feedback. And then even beyond that, I think cracking into fintechs was one step. And then going

19:47from that to getting a bank to want to use you is a completely other challenge in and of its own.

19:52So that kind of got us in the door. And then our first bank that took a chance on us and

19:57vouched for us for other banks. Like we couldn't have done any of it without their, their like

20:01testimonials and like, and trust. I love that. Obviously you studied,

20:05you're an AI expert. You studied computer science at Stanford for AI. Most people have it. What advice

20:10do you have for a founder or CEO that is not technically trained, but they know that they have

20:15to have an important AI component in their business? Totally. I actually would flip my advice

20:20here because I think I see a lot of people who are trying to say, Hey, my goal is to bring an AI

20:26somehow. When I think bringing an AI for the sake of AI, eventually just looks a little bit sloppy

20:31in a product. Cause you can just see that it didn't, it didn't need to be there. You just added this

20:36random chat bot and you're like, Oh, what am I supposed to do with this? Like I'm sure all of us

20:39have had that, have had that experience. And it's like, okay, great. You added this chat bot,

20:42but like, I'm not even using it cause I need to do this thing. And so I think the real way that we've

20:47realized even people who are or aren't technical can use AI properly is stop just like aggressively

20:53looking to slap in AI everywhere and actually figure out like, what are the problems that

20:57we need to solve? And some strong fraction of them can probably be solved very well by AI.

21:03Some of them may or may not be able to, but I think starting from the need is how even without

21:08the technical background, you can make sure that it doesn't look like something you just threw in.

21:12So that'd be my number one thing. And then the other thing is just like hire people who really

21:16know their stuff. I think there's also a lot of AI companies I see who don't have any expertise in

21:20house. And then you can also see that things work and work and work. And then the second you hit

21:26a huge amount of data or any kind of like scaling issue or quality issue there, you're just stuck

21:32there. And the only way you can get unstuck is if you actually like know the technical chops.

21:37And if you don't know it, then bring in someone who does.

21:39Gotcha. Is there, what's the one piece of

21:42business philosophy that you live by and how has it made a difference?

21:47Um, I think the one piece of business philosophy is really just to be delusionally optimistic.

21:53Because there were times like when we kind of shut down that first idea and we were in that six

21:58month of pivot hell, as they call it, we had literally nothing like Ashi and I had nothing but each

22:03other and literally nothing except this belief that like, Hey, I'm not going to go dead without

22:08a fight. We're going to make something happen here. Even though at that point we had, we had nothing

22:13all over again. And so I think there's like, you know, successes, you know, we randomly get things

22:19happen to us randomly. Things can get taken away. Like you can't count on anything, but I think you're

22:24just ability to say like, I'm going to succeed because I'm going to force it to happen. You know what I

22:29mean? Yeah. Yeah. Wonderful. And I want to talk one more thing about AI and energy. Obviously

22:33AI, everyone's talking about hum, everyone's building power plants, Meta's building nuclear

22:37power plants to power all this AI infrastructure. The amount of electricity is insane. How do you

22:42think about sustainability when you're building your company? Yeah. I've thought a lot about that.

22:46I'm seeing it pop up in a lot of places. I think the right way to think of it is like twofold.

22:52On the first piece, it's weighing the total pros and cons of doing something and figuring

22:57out like what really can, what is the bigger picture here we're trying to solve? So for example,

23:03automating anything is going to take electricity, but are there like second and third order

23:08consequences that ultimately make everything more productive and save people's time, save people's

23:15energy, like save more energy in the long run? I think there's, I sometimes take a little bit of

23:19issue with people that say, oh, don't use AI because it takes up electricity. You could make the same

23:23argument for a Google search and people did at the time, you know, or you could make the same

23:28argument for using any computer at all. So like things that progress innovation are going to take

23:34energy. And I think the way to do it is to invest very heavily in finding ways to make it efficient,

23:39but not to put the burden on necessarily the consumer to act against their best interests when

23:44this is something that is like bringing a lot of abundance. So that's how I think about it. I think

23:48I saw the statistic that people like doing a single chat GPT search consumes significantly,

23:55significantly less energy than like eating a cheeseburger. So I think like none of us are

23:59going to be perfect. So you just have to choose like, where am I going to do things that I decide,

24:05you know, weigh those pros and cons properly? Because you can't live, like you can balance everything

24:10perfectly. If you do, then we would all be hypocrites. Well, Kalyani, that was great. Thanks for joining us.

24:15Yeah, thank you so much. It's great.

Category

Transcript

Be the first to comment

Recommended