Julian: Hey everyone. Thankyou so much for joining the Behind Company Lines podcast. Today we have FloCrivello, CEO and founder of Lindy, the personal AI assistant that can helpwith tasks from calendar management and email drafting to contract sending andbeyond. Flo, I'm so excited to chat with you and get to know not only yourbackground, your experience, but also you know, what you're working on atLindy.

Especially as you mentioned before thisshow, before we hopped on the recording, how much AI has really boomed and, andreally come to the forefront of a lot of different, not only companies, butalso in, in, in, popularity and, and people are using it in so many differentways. I mean, it's really.

We found it not only to not, not in areplacement sense, but we found it in a way to kind of enable and support the,the mundane tasks and other things that are, less of the, the creative kind ofside that, that a lot of people just either tend not to like to do or, or tendto take too much time outside of the other things.

But we'll go into AI and what that meansfor kind of a broader scheme of, of individuals and also how you've been ableto help people. But before we get into all, What were you doing before youstarted the company?  

Flo: Yeah. So before IndiaI was founder of a startup called Team Flow. We were building a, a built alloffice for remote teams right at the peak of the pandemic.

And before then I was an engineer andthen a product leader at Uber while I was for almost five years.  

Julian: and it was interestingto think about your product experience, being able to, to build and identifyways to really create kind of a simple process with a lot of complex solutionskind of built on top of it.

How do you go from a pro? I always liketo talk to product people cuz I. I, I think it helps all of us kind of, definehow to build, but how do you kind of think about the user experience creatingsolutions that are simple, but you know, with a lot of complex tasks and, andnot to overwhelm people, which I think is, it's kind of the balance.

How do you go and approach that?  

Flo: Yeah, I mean that'sthe art and science of product management, isn't it? And as an aside, Iactually believe that right now this is the skill that is in a way most lackingin ai. We see a lot of research labs come out with very cool models, but Ithink the next thing is we're going to see product teams actually build fullend-to-end product around these models.

So I mean, the way you do it is really,again, those are not end science, but the typical framework is that you, you doa ton of. To define does this framework that's called the double diamondframework. And so mm-hmm. Think of it as two diamonds, like two squares thatrotate 45 degrees, one after the other.

And the first one is the diamond of theproblem, and the second one is the diamond of the solution. And so the, thediamond of the, the problem, each diamond at first, you open it and then youclose it. So the diamond of the problem is you, you open wide, the up, the a.And you talk with your users about what kind of problems they're having andyou, you try to wrap your head around the problem space and then you close thediamond, which is you really decide to narrow it down on a few problems, one,two, maximum three.

And you really define these problemssuper clearly. And then you do the same thing for the solutions. Like, hey,okay, now that we've got an idea about the problem space, how do we go aboutsolving this problem? And same here, you open wire, you come up with a bunch ofideas, and then you close and you decide exactly how you're going to solve thisproblem, and you define it super clearly.

I actually think that when you'rebuilding a startup, there is a third diamond that is even before the problemdiamond, which is the user diamond. And so you actually duck to a bunch ofdifferent verticals and types of. And you try to figure out where the newtechnology that is coming about is most useful.

Mm-hmm. And then once you define yourICPs, then you narrow down the kind of problems that they're having. Yeah.  

Julian: That definitionthough, it is a little difficult if, if you see, say, a lot of problems in acertain particular sector, especially, what you're working on at Lindy, beingthat you can go and utilize AI in so many different ways along the life cycleof.

Workflow, how do you define that initialproblem? Is it simply, I have, a sample of X amount of user users that havethis particular problem versus this, set of users that have, a lesser problemin a different area. So I'm gonna focus on the, the larger kind of populationsection.

Is that kind of how you define it basedon, amount of impact in a particular group? Or is it the difficulty, thechallenge of the problem? How do you kind of define what to focus on?  

Flo: Yeah, it, there liesagain, it's, it's part art of science, but I think you need to find somethingthat is simultaneously aligned with your long-term vision and it's valuable ofall the short term.

So if you think about it as climbing theEverest, You're looking for a base camp on your way there where you can rest.And so that's what you're looking for. You're looking for like the best basecamp that is on your way to the Everest. And I think one tricky thing that onetrap that funders sometimes fall into is that they find the base camp that isnot aligned.

It's not in the Everest, it's likeanother model. And, and they go after it. I'm like, if it's not on the way tothe air, why are you going after it? You want, you wanna climb the air, right?And so, yeah, I mean there's, there's multiple ways to go about it. So one.Problem first, and the other is solution first.

And so the solution first approach is,hey this new technology that is seeing the light now, what is it uniquely goodat? And what is it that needs these things that it's uniquely good at? Forexample, Jeff Bezos did that with Amazon. He was like, Hey, everything's toole-commerce. What is it uniquely good at?

Well, there is infinite self. Okay, sowhat is it that's got infinite skews? Oh, books. There's a lot of books. Andalso they, they, they have infinite self life, so boom, it's a perfect fit forthis paradigm, right? So in our case, we were like hey, what is AI really goodat? Well, AI is really good at, really good at processing information and, andnow actually performing actions.

And so we were like, okay, who is inthe. Processing information and performing actions all the time. Mm-hmm. Andagain, in this spirit of base camps that are not on your way to the Everest, weactually identified also law and finance. So financial analysts or in thebusiness of analyzing tropes of information lawyers or in the business ofgenerating contracts and in the discovery process they have.

Thousands of pages of documents to readand to extract information from. And we actually found an like, oh wow. Likelawyers were like begging us to give them something here. And then we werelike, you know what? It's not really a wedge into our long-term vision, whichis our long-term vision is we want to automate all knowledge work, right?

And no doubt, law and finance or bigcategories of knowledge work, but we more and more realized that those won'twedges into our long-term visions. Those will entire categories in and of themselves.Yeah. In the same way that like Amazon doesn't sell houses, right. You go tozero for that, right? Yeah. So, eventually we, we, we ended up blending onexecutive assistance.

Mm-hmm. So that is where we found anopportunity where we were like, Hey, it's pretty well defined. People want helpwith the calendar, with their email, with their meeting recording and nottaking and so forth. It's very tractable and it is very aligned with thelongterm vision.  

Julian: Yeah. And, and what inparticular, when you think.

Knowledge you said, you said knowledgework. How would you define that? What is knowledge work and, and does everybodykind of have this knowledge work in some different variation? I'm curious aboutthat. I haven't heard that vocabulary before.

Flo: Yeah. Any work thatcan be done behind a computer that could be done by your remote worker isknowledge work your manipul policing information all day.

And so effectively as a human, you are afunction sitting between input and output. And your input is your screen andyour output is your keyboard and your function. And so AI, I think, willeventually become very good at approximating that function.  

Julian: Yeah, yeah. Andthinking about, just backtracking a little bit, you were in, you were at Uberand, and, and then you kind of started to build things.

Teams and, and the kind of workflowprocesses. What particularly got you interested in that being that you couldhave gone so many different directions? What was, what was the catalyst orinspiration for starting out and, and starting to build a startup when, youworked for arguably one of, one of the larger companies, at least in, in theBay area, in, in globally at the time?  

Flo: Yeah. I would actuallypose the questions the other way around. Why did I work at Uber if I wanted tobuild a startup? Because that's actually. To me building a startup was alwaysthe goal, and I actually viewed Uber as a stepping stone for stone for that. Sowhen I arrived in San Francisco, I was 20 and I had a checklist of full itemsthat I wanted before I started the company.

And it was, I wanted, number one, agreen card. I wanted money savings. I wanted skills and I wanted network. Thosewere the full items that I wanted to check. And by the way, in hindsight, Ithink I placed the bar too high. I, I, frankly, in hindsight, I wish I hadpulled the trigger earlier and started the company earlier, but eventually, soI, I joined Uber and I studied, steering my career in the direction that I knewwas aligned with this vision of being a founder.

Yeah. And so that's why I went frombeing an engineer to being a product manager. Yeah. Because I actually thinkthat this is what prepares you best inside the, the company to become afounder. Yeah. So I became a pm. And yeah, I, I, I do think I learned a lot.It's not exactly like being a fund, obviously, but it comes closer inside a bigcompany.

A little by little. I checked these fullboxes and I was like, all right. And at some point I felt ready. And again,frankly, I think I waited one or two years too long, but, oh, well. Andeventually when I, when I felt ready, I just, I just did it.  

Julian: Yeah. And, andthinking about, that experience at Uber, what are some things that.

You saw them, that they were doing well,that you carried over into your other companies that, that, kind of offeredmaybe some maturity to, whether it's destruction, operations, or the efficiencyor effectiveness. Being that you worked at such a hypergrowth company, anythingthat you carried over that was, useful to implement at your other companies?

Flo: Yeah, so much. Ithink,

I think there are two or three bigthings that. From Uber. The first one is the value of ownership. Yeah. Thething I learned at Uber is that no one's going to help you and no one's goingto stop you either. I've seen Uber is really good at taking 22 years old andgiving them entire business lines and telling them You go do this heels piping,and you're like, me, like, who?

Like me? And it's like, yeah, you. Andit's like, so that's another thing I learned. That works. You can actually givepeople an insane amount of ownership and people step up. Yeah. Even if you'reyoung, you're inexperienced people. It's just mind blowing what people canachieve when you give them autonomy and high expectations and you hold themaccountable.

So, The value of hassle as well. Thevalue of thoughtfulness. I think Uber is really good at stats. The value ofbeing extremely fast, moving faster than feels reasonable frankly. Yeah. Ithink the value of a culture, so some of the values that you built back then atleast were merito, crassy and toe stepping and what those stent for. And therewas another one that was truth. And so what they stand for is like when you'rein a room, it's not the person with the biggest title who wins theconversation.

It's the person with the best peers. Andyou see these people, you, you see people in meeting with Travis speak up tohim and not be afraid of speaking up to, and Travis with a fully courage thathe would, you would get into a, a debate with you. Right. Sure. And, and, and Isaw that all the time and, and it was, it was very much a courage to speak upand, and, and engage in heated debates even sometimes.

Sure. Because most important thing wasfor truth to come back. Yeah. And that's been a huge learning as well.  

Julian: Yeah, the, the valuesreally steer kind of how the, the company kind of acts or, or when you don'thave, say, an objective or, or you're looking for an objective. What are our inline values? How do we work, how do we communicate with each other?

What are we seeking out? And kind ofdetaches, the individual from the overall company or project or, or product,which, which you can see so much. A lot of companies having success by aligningin that direction. Thinking about Lindy, what in particular was exciting aboutAI and, and when you started building it and how is it, kind of, I guessevolved now that, AI is kind of almost it's household concept now that a lot ofpeople can access and, and, and engage with.

How was it when you first startedbuilding the company and, and what's changed now since all the popularity hasreally gained traction and, and put a lot of put a lot of spotlight on thistechnology.  

Flo: Yeah. I think AI hasgiven us really impressive demos for a very long time. So I really started toget in AI as many people after the image net moment, which I believe was like10 years ago or something like that.

And like for the first time, AI becamereally good at image recognition. And, and like, that's when I started payingattention. I was like, wow, like something is happening there. And then Iactually came close to starting an AI startup after DPT two came out actually.And back then, my idea was I wanted to start something like an enterprise SEsolution based on DPP two.

And I played with it actually quite abit. And I, I ended up concluding that the, the technology was not quite thereyet, but it was always in the corner of my eye. I have many friends in ai. I'vealways like stayed very close to it. And then GPT three. And I think franklytook everyone best surprise what or lhf could do to these models, like the kindof capabilities that it could unlock.

And that's when, for us, it was like anholy cow moment. It was really like, this is technological revolution of ourlifetime. It's just, it's just extremely obvious and you, you get sucked intoit almost, you, you almost don't wanna get sucked in and you get sucked intoit. Yeah. It's just, it's just, it's, it's hard not to get, it's just thepossibilities are endless.

Julian: Yeah. And, anddescribe to the audience to give a little bit more context what Lindy does todescribe the product and, and what the user experience is. If I'm somebody whouses Lindy, what does that look like in my day-to-day functioning? And at whatpositive or, or I guess, productive outcomes should I be able to kind ofmeasure or receive or acknowledge when working with a tool versus not?

Flo: Right. So what Lindydoes is she's a, a personal AI assistant that takes care of your email, yourcalendar, and your meetings. At first. Those are the things she does very well.Now, she can do a lot more than that, but those are the things she does very,very well right now. When you use Lindy, the idea is that it puts your work onautopilot so it takes care of all the menial tasks in your work such thateverything just.

Seamlessly. You don't have to worryabout any of the logistical work anymore. People are much too clever to bespending their time spending calendar invites, right? For example, of takingnotes during meetings. You should keep the high capabilities of your brain todo what it's uniquely good at. So for example, when you use wind, you wake upin the morning, you open your Gmail inbox, for example.

And responses to your email are predrafted in your inbox. And they are pre drafted using what Lindy knows aboutyou and what she has learned about your voice. And she actually tailors hervoice based on how you speak to each recipient. Yeah. So hopefully you don'tspeak the same way to an investor as you do to your partner.

And so it gives me draft emails withthat in mind. She's going to triage your emails to let you know which one'salmost important to pay attention to right now. She's going to join yourmeetings and take notes during your meetings. Yeah, so at the end of yourmeetings you can send a summary email or if you say during your meeting, like,Hey, let's capture this as an action item.

She's going to send that afterwards. Oryou can say, Hey, let's meet tomorrow again at three. She's going to send thecalendar invite, and so on and so forth. So today, it just makes your, your,your life and your work way easier in, in a myriad. Our very long-term vision,again, is to automate knowledge work.

We want Lindy to be like an otheremployee in your company, except she works at Superhuman Speeds and sits onevery meeting in your company, has write every email, every document, and canjust perform an insane amount of work for you and, and your company.  

Julian: It's so fascinating tothink about the amount of, of things that it can handle and, and, and thinkingabout the amount of freedom that you have after that free time and, and, and,do other tasks that are a little bit, maybe more high level or a little bitmore involved with other moving parts.

But describe one thing I'm interested inis how do you create a platform that communicates with other pla with other,software, whether you work on different calendars or you scheduled meetings toa different, resource. How, what is the challenge around building a platformthat integrates all these other softwares and are, is there any software.

Unfortunately it gets left out becauseof certain restrictions or, or, or certain, maybe lack of capabilities. Maybethere's an API that plugs into the model. How do you create something thatYeah, I guess works with everything else?

Flo: Yeah. I mean todayLindy integrates with pretty much anything that has an api and soon she will beable to integrate with things that don't have an API too, because we're alsoworking on making her be able to manipulate user interfaces.

Yeah. So, building these integrations isactually nuts. The. It's actually pretty awesome because we use Lindy to buildthese integrations, so she builds her own integrations based only on thedocumentation of these APIs. Yeah, which actually leads to some surprises fromtime to time. Like for example, recently we asked her to send a positivemessage.

To the general channel on Slack. And shedid that. She was like, Hey everybody, today's a new day with infinitepossibilities. And she changed her profound picture on Slack. And not only hadwe not asked her to do that, but she, we actually didn't even know that therewas an API endpoint to do that. So she actually has capability that we don'teven know about.

So, the, the, the challenge is not asmuch in buildings integrations as much as it is in, in teaching Lindy how touse Disintegrations and in what order. Yeah. So that's, that's, that's what we.And, and the way Lindy does that today is when you ask something to Lindy, shewrites a piece of code to perform that task for you.

Yeah. Which actually think is, isendlessly interesting because effectively what's going on is that you get a miniapp. That is dedicated just to the one session that you have right now. Yeah.And when you close your tab, effectively, that mini app gets deleted, which Ialways compare that to the industrial revolution because before the IndustrialRevolution good, were very expensive.

Right. If you think about it, uh uh, capwas very expensive. If Penn was very expensive, these things will be $500 orsomething. After the Industrial Evolution, they got so cheap that we gotdisposable versions of these items. Right? So like the solo cap, the big. Andso forth. Yeah, I think the same thing is happening to code right now.

So code, right now, an averageapplication costs say 10 $100,000. We are rapidly approaching a world wherethese application, actually these applications actually cost 1 cent or 10cents. Yeah. They're actually so cheap that you can get single useapplications. Just for the one session that you are using right now.

So for example, I could be like, HeyLindy, show me my calendar events and she's gonna show me a list of calendarevents, and I'm like, show them to me on the map and she's going to render orGoogle map with my calendar events so I can see where my meetings are for thatday. Yeah. First of all, this is something that I don't know of a singleapplication that does that.

So I'm getting an application just fordecision that is still just for my specific use case right now, which I thinkis just.

Julian: Wow, it's sofascinating to think about how it's training. To integrate with these applicationsand utilize and almost act like, it's like a small dev, team or full stackdeveloper being able to quickly iterate and, and integrate with all thesetechnologies.

And one thing as I think about it, isthe model building and, and how that's advanced in a lot of different ways totrain these AI systems to be able to be more sophisticated and on target withthe request that you're, you're submitting to it, right? Or, or generating or,or, asking of it.

How are those models, how are thosemodels advanced? And, and where are those models coming from? Are they comingfrom, research teams? Are they coming from other industries that are doing alot of investment into, the, these types of models? Where, where does thatinformation come from and how, how more quickly and rapidly are they advancinginto the understanding of, of how to complete task and cooperate and work withhumans as we, go through our day to day?

Flo: Yeah, these models arecoming from a lot of different places, so some of them are being open sourcedby very big risk. Village Labs, the biggest ones right now, the biggestresource labs are going to be Deep Mind Fair from Facebook or Meta Open ai.Obviously though they are not open sourcing, they're not language models, so,these big labs are releasing models and then the open source communities as awhole is also releasing models and training models like stable diffusion and.

Well, so, what, what we do, we spend alot of time keeping abreast of the latest models and constantly evaluatingthem. So we've built our own, what's called an evaluating, so a way to evaluatethese different models based on our given use case. And so we spend a lot oftime and we've built a lot of tooling and infrastructure to constantly testthese models and see how they're performing.

And then we take these models as a baseand we fine tune them for our use case. So we also spend a lot of time and, andmoney and effort bringing together a data set. So we have a very large data. Oftasks and how to perform them best from a lot of different sources. And we usethis data set to fine tune and train these models.

And so, not only that, but we also havea little bit of a hybrid system where we, we actually have different modelsthat are specialized in different tasks like entering a dialogue with the.Answering factual questions, browsing the web writing codes. So we havedifferent models for these different things and we've built a sort of communityof models, if you will, that work with each other on models who perform thetask that is given to them by the user.

Julian: Yeah. And what's thechallenge? And, thinking about all these, a lot of companies are using AI and,and one way, shape or form, we have, Jasper kind of, Jasper AI is, is doing alot around content and, and helping. Disrupt kind of the, the writer's blockand all that kind of and then chat GBT is kind of an all around tool whereyou're able to ask questions to generate.

We have Lindy who's able to, take a hugeload of the workload that is just unnecessary for people to do and, andautomate that and process that's seamless and, and in your own voice. What'sgonna be the challenge in, whether it's this year in the next couple years?Differentiating but utilizing AI in similar AI systems for, for companies beingthat now we're seeing a huge kind of boom and a almost like a red ocean ofcompetition with a lot of different companies and, and how they utilize ai.

What do you kind of predict for, notonly Lindy, but for the overall industry? In this next evolution of, AI and,and tools that, that come along with it.

Flo: Yeah. I, I, I thinkit's Na Friedman who recently tweeted something like, and I actually agree withhim, that tech community has become overly focused on modes.

Yeah. I, I, I see too many foundersspend too much time talking about strategy instead of tactics. I feel likeproducts live in tactics land. You are in the business of surviving for thenext three months. Yeah. And look at the end of the day, the reality is that asa startup who just traced a seat round and you have five or 10 people buildingsomething, you don't have a mode.

Because if you, if you, if you had amode, then so did the next guy next to you, because you don't have anystructural advantage at that point. And so that means that you don't have amode. And so all that really means is that you are in an, in an exhibitionplay. Your job is to build the most insanely awesome product that you can anddo that faster and better than the next guy.

Yeah. Yeah, good enough that it beatsthe competition. This often talk about 10 x, it's gotta be 10 x better than thecompetition to overcome the habit of using the competition. Yeah. In this case,there isn't much of an incumbent yet. Perhaps. Open AI is what comes closest,but the field is so new, so you are in an execution game, you just gonna naildown the execution.

So in that perspective, I don't thinkbuilding an AI is very different from building in any other. Perhaps buildingan AI is different in so far as you owe operating at the, at the bleeding edgeof technology. Sure. Yeah. And so there isn't as much of a playbook. Everythingis changing very, very rapidly.

The models are changing. The best waysto build are changing the paradigm. The day of framework is changing all thetime. And so I think that a lot of things will seem obvious to people five or10 years from now that today are absolute mysteries. Just in the same way thattoday things like cloud, right?

This idea of you have code running onthe server and it hits up a front end and you've got based on the back end andjust rip on the front end, like some code that gets executed in front parts ofthe side today. For us it, it's very obvious it took the industry decades todiscover that. Yeah, so I think it's going to be the same thing with ai, exceptthe playbook is still being figured out and what goes on at what parts of thestack is.

Figured out. And, and it is a verychallenging tech to build with, right? So that is the add level of difficultyhere. But otherwise again, it's, you're basically in an execution play, whichis the same game as any startup.

Julian: Yeah. And tell us alittle bit about the traction that Lindy has so far. What are you particularlyexcited about, your initial launch and then what are you particularly excitedabout the future and, and the different goals that you have set for this yearand, and beyond?  

Flo: Yeah, the launch thatwe did a few weeks ago went super well.

So we've accumulated more than 10,000sang up to the waiting list. When people sang up to the waiting list, we ask,we ask them for their willingness to pay. And obviously omni only subset ofthose are going to. On and so forth. But right now at the winning list, we havemany millions of dollars of, of willingness to pay.

And I personally have never seen thisamount of market pool except perhaps at Uber where I have gotten dozens ofpeople reach out to me, begging me to let them through the waiting, the winninglist. I've actually gotten friends who received all the reach outs becausepeople knew that they knew me.

Yeah. And they were like, Hey, let methrough the wait list. I've never received it. I've never seen anything likethat before. It's just like an insane amount of market pool. Only, I think alot of that is also a function of the current AI moment that we're goingthrough. What I'm excited about is just to see this product become real.

Yeah. It's just, it's, it's only verygood. Like I use it all the time. It's, it's, it's really good and I think it'sgoing to become even better over the next 3, 6, 12 months. I, I, I'm justexcited for this kind of product to exist, frankly. It's going to be soinsanely good and it's going to make such a difference in people's lives.

Julian: Yeah. Yeah. It's, itis incredible. The. Honestly, the biggest change is the velocity that thingsare being done at, whether it's on the individual level or by companies, as youmentioned, it, it things are just being done so rapidly and so quickly that we,we've just never seen this level of progress or the speed of progress in solong.

Let alone, I, I don't think ever, Imean, I, I can't imagine. I mean, how, I mean, what's impacted in the last few.That speed is just so different than I think anything historically. Thinkingabout just AI and, and whether it's internal or external, and thinking aboutLindy, what are some of the biggest risks that you think the company facestoday?

Flo: I think the biggestrisks are gonna be internal. I think, can we execute? Can we build this? Yeah.Can we build this good enough, fast enough, better than the competition? That'swhat I'm worried about. Yeah. To your point about things are moving very fastright now. Yes. It is really mind blowing. Like, I think AI is going to change.

Everything. Yeah, I think it is a muchbigger deal than the iPhone. A much bigger deal than even the internet. I, Ithink it's Larry Samuel who said it's all similar to the invention of the wheelof fire. And I actually, I think it's, it's a transformative invention. Now thefunny thing is that when you look back at history if you look for example, atthe time window between 1,919 30, I'll give a short list of inventions thathappened between this still the years.

The car, the radio, the television,penicillin, the vaccine plastics the plane right, that was in still the years,right. Plastics, I'll, I'll just pass on this for a minute. Electricity, thesethings are huge deal. There was still the yields, right? What happened in thelast 30 years between 1990 and and 2020.

Right. So that's the whole Peter Tealgreat technician theory and I, I wholly subscribe to it. Yeah. Like theinternet and smartphones are great. Like the hybrid or mark here was Uber, likethe biggest change we had to physical world was Uber, which is awesome, butit's, at the end of the day, it's people driving people in their cars.

And that's the best we've had in thelast 30 years before my, my late grandmother passed away and, and she wassurprisingly up to speed on the latest. To call advances. Like she was all fromAmazon and all of that stuff. And I was like smugly asking her, Hey grandma,what do you think of the internet and the smart And I was expecting, to belike, oh, it's amazing progress so fast.

And she was actually saying, Flo, like Igrew up without running water or electricity. I had no fridge, I had none ofthat stuff. Like my toilet was a, a hole in the garden. And so I remember whenwe got off those fridge, I remember when we got running water. I remember allthat stuff. So the internet is great.

So anyway, I think my point is thathuman memory is very short, and so yes, we are living through a historicaltransition right now, but it's hardly the first one. And I actually think,surprisingly, perhaps that has been the norm through human history is tothrough this kind of rapid changes. Yeah. And I'm excited to join the rest ofhumanity in living through one of these changes myself.

Julian: Yeah. Yeah. Ifeverything goes well, what's the long term vision for Lindy?  

Flo: I think it's going toautomate knowledge work. Yeah. I think in an immense amount of prosperity. AndI think it can change the way people work and businesses work and the waycapitalism operates.  

Julian: Yeah. Yeah. I alwayslike this next section, I call it my founder at faq. So I'm gonna hit you withsome rapid fire questions and then we'll see where we get. So first question Iwould like to open it up with what's particularly hard about your job?.

Flo: The surface area ishuge, so you get punched in the gut by employees, investors, customers,competitors, whether past, present of future ones. So that right here, there'slike a 12, like 12 different categories of things that punch in the gut. And soyou get punched in the gut, pretty much day, multiple times a day.

And sometimes you have like five ofthose back to back to back to back. And it just sucks. You get used to it, butit's, it's not very pleasant, so, yeah.  

Julian: Yeah. Yeah. What's onething you, now as a founder that you wish you learned earlier on?

Flo: Premature scaling isthe root of all evil. Yeah. For your religious skill. Yeah.  

Julian: And when do you knowyou're ready?  

Flo: You'd know it if you,if you got a,  

Julian: yeah. Yeah. What'ssomething that you spend too much time on that you would like to spend less?And what was something that you spend little time on that you would like tospend more?

Flo: If I knew I would, Iwould do it. I don't know. That's a question I constantly reprioritize my timeto spend time, less time on things I shouldn't spend time on and more time on,more time on things I spend time on. So, I spend an immense amount, amount oftime these days hiring and building the product with the team.

And I think those are the right thingsto spend time on.  

Julian: Yeah. Being a founder,especially within, today's climate. A lot going on, right? We think aboutfinancial infrastructure. We have a lot of cryptocurrency that's kind of, and alot of generative ai and so many different things that we haven't had, or wehaven't had in the past to consider.

As, as founders, how do you kind of goabout not only learning and, and, and, and kind of obtaining information andapplying it to your. What's that process like and how do you kind of balanceout building plus learning the new information that's out there to, to makesure that you're up to date and up to speed on what other companies and yourclients are dealing with. Which is a challenge kind of, it's like two differentpositions in, in, in, in one, but what, what in particular do you do to kind ofmanage that?  

Flo: Yeah, that's a goodquestion. I think you've gotta be deliberate about how you manage yourinformation diet. Mm-hmm. And I think that you cannot totally isolate yourselfas a father.

Yeah. It is sometimes tempting to go inwhat I call the cave, and you just go head down and you close everything andyou just build your own product. And then at some point, 12 or 24 months later,you broke up your head and, and the world has changed. Then you had no idea. Soyou can't do that's that. The way I do it is I am addicted to Twitter perhapstoo much.

I spent perhaps, sorry, your previous.Least perhaps one thing I spent too much time. So it's, it's Twitter and I'm,I'm deliberate at who I follow on Twitter and, and right now's it's another ai.And I think Twitter has this nice property where it's, it is addicting. And soif you, you, you can somehow make learning fun because you, you follow peoplewho are tweeting about things that you want to learn about.

Julian: Yeah. Yeah. I alwayslike to ask this question because I love how founders kind of extract knowledgefrom anything that they ingest, but whether it was early in your career or now,Books or people have influenced you the most?

Flo: It's gonna soundcliche, but it's very cliche, but it's Steve Jobs, Charlie Munger, Elon Musk.

It's not like such a, such a stereotyperight now, but I think, I think I learned, I learned from Steve Jobs theimportance of vision. Mm-hmm. And creation and like the existential importanceof, of these things. It's like, yeah, gonna make money. And even if you don'tpick money, money is not important enough to sacrifice your creative potentialas a human.

Yeah, I, I learned that from, from fromMilan. I'm, I'm learning to think big. Mm-hmm. And for yourself and fromCharlie Munger. I, I think I've learned the importance of, I think it talksabout, it is more important to remember the obvious that, than to figure outthe arcane. This is why here I. Like you're asking how do you differentiateyourself as a funder and I'm like, just drop the mba.

There's no, just build a great, awesomeproduct, right? Obsess, custom mill, and build an awesome product. There isn'tIble to instance that. Yeah. It's just gotta like the most important thing isto keep that in mind, not to figure out the arcane.

Julian: Yeah. Yeah. I lovethat. And I know we're coming to the close, the episode and I always like togive people a chance to give us your plugs and, and give our audience an ideaof where to support you. But before we do that, I wanna make sure we didn'tleave anything on the table. Is there any question I didn't ask you that Ishould have or that you would've liked to answer? Anything come to mind?  

Flo: No, this is great. Igreatly enjoyed the conversation.  

Julian: I appreciate it, Flo.And last little bit is where can we find you?

Where can we find Lindy? Give us yourwebsites, your LinkedIns, where can we support you in the business? And alsoTwitters, give it all to us so we can come and support you.  

Flo: Yeah. The website islindy.ai. The Twitter is getlindy, and my personal Twitter is @Altimor, A L T IM O R.  

Julian: Flo, it's been such apleasure not only chatting about your early experience in your career, how youview product and, and building, but also currently LIndy in kind of the, thenew generation of this whole AI movement and what it means, to everyone.

I think, and it, it's not necessarilyimpacting one individual or one type of profile. It is everyone and, andeveryone kind of. The opportunities use advanced tools like this to, to maketheir, their job either, less, less mundane, or more sophisticated in a lot ofdifferent ways. So it's been such a pleasure chatting with you. I hope youenjoyed yourself on Behind, Company, Lines, and thank you again for joining ustoday.  

Flo: Thank you so much,Julian.  

Julian: Of course.

‍