Uizard’s Tony Beltramelli on text-to-UI & PMF + building a Notion CRM w/ Kelly Schricker | E1722

0:00Okay, everybody, we've got a great show for you today. Remember, AI is changing everything. It's moving at a pace that we've never seen. And you need to listen to every episode of This Week in Startups because I am bringing a different CEO on almost every day to talk about this AI revolution. Today, we have the co-founder and CEO of Wizard. It's spelled U-I-Z-A-R-D. It's basically Jeter of AI, you know, build stuff for you, but it does it for UI design tools. So you can take a sketch on a piece of paper and make a pretty decent website in just minutes. It's very impressive technology. this is stuff that used to take you know days weeks and now it takes minutes it's a great interview and then we have another amazing segment for you kelly on my team who co-runs founder university that's our 12-week course she's going to show you i like to do we like to do some practical tactical stuff she's going to show you how to build a crm inside of notion we did this ourselves i can tell you from experience building a customer relationship management system inside of Notion is brilliant.

1:00And this Notion CRM that Kelly is going to do has saved us, I kid you not, $100 ,000 in software costs, and it's made my team bionic. This is a very short talk. It's only 22 minutes, but it's going to save you a ton of time and money. And that's what the Founder University series we're doing is meant to do. It's a 12-week course, plus it's a podcast. Today's going to be a great show. You're going to learn a lot. You're going to get inspired. So please stick with us. This Week in Startups is brought to you by Lemon.io. Need to speed up your product development without draining your budget? Hire vetted engineers from Europe at Lemon.io.

1:38Go to Lemon.io slash twist to get 15 % off for the first four weeks. Embroker's startup insurance program helps startups secure the most important types of insurance at a lower cost and with less hassle. Save up to 20 % off of traditional insurance today at ambroker.com slash twist. While you're there, get an extra 10 % off using offer code twist and Trovada. Starting up is hard. Trovada makes managing cash easy. Start automating your cash management at trovada.io slash twist. Use code twist for 30 % off one full year of premium features like AI forecasting. All right, everybody, there are incredible ideas that have been floating around about you, an individual, talking to your computer and having the computer design and make beautiful things for you.

2:34uh we saw this in uh star trek when captain picard would say to the replicator that he wants earl grey tea right well one of the holy grails of voice interaction with a computer would be to build an app or build a design something beautiful so instead of taking out photoshop or taking out canva or taking out any adobe figma all of these great tools uh bubble comes to mind um envision what if you could instead of learning all the layers and learning how to use all the tools and the plugins could simply speak simply speak to your computer and have it build magic for you well we're going to see a demo of that today on this week in startups joining me is tony beltramelli he's an italian who grew up in france i'm a french that grew up in france you're you're a french who you're you're french but an italian last name

3:36what's the penultimate like or i should say the ultimate what's the ultimate italian dish that the french do just as well i'm always interested in these crossover dishes because the japanese do a number of french and italian dishes particularly well what would you say the french with an italian dish the french do particularly well well Well, you know what? I think I'm not going to be able to judge really well because my Italian ancestry is just so far in the past. But I would say anything with cheese, I think both countries kind of nail it. They nail it on the cheese front. I agree. I agree. All right.

4:09So you have a company called Uzard, as in Wizard. Am I pronouncing it correct? Uzard? Fantastic, Jason. Okay. So is it Wizard or Uzard? You know what? We like to say whatever people like to remember it because ultimately that's what they speak. That's what they say to their peers. So we like to call it wizard internally, but some people call it user. Some people call it UIZR. It doesn't matter. So UIZR.io is how you're going to find this app. So you've been working on this for a couple of years now. This isn't part of the chat GPT for 3.5 boom. You raised money back in 20... Well, you raised your series A in 2021, but you've been working on it for a year or two before that.

4:53Is that correct? That's correct. we started in early 2018 so it's been okay you've been at it for a while and so when did you first become aware that ai could help with voice to product voice to product interface or did you just start free ai with voice to ui interface we actually so this this company actually started as a research project um so you know i have a computer science background i've always been frustrated by the way we were designing and building software and so all the way back it was not voice like or text to design it was like sketches or screenshots to design and the idea was can we use machine learning to kind of like automatically generate design from my dirty napkin sketches or from you know a screenshot from the app that we pushed last quarter and kind of like automated design process.

5:50And I was, you know, way before text to design was even supposed to be practical for productive use cases. And so let's just do a quick demo of the product because I think once people see this, they'll understand how far you've come from the initial idea. So maybe you can share your screen and remember a lot of people are listening. Not everybody's at youtube.com slash this weekend, watching this episode but uh i'll help you sportscast it so here we are we're on a simple screen and maybe you could start walking us through this yeah yeah so you basically select the device you you want to design for mobile tablet web um okay and then you basically describe your project so we have a form and the form is asking you what do you want to design for mobile tablet or web which one did you select there i selected web um well except if you have a good idea but i thought it would be fun to generate a website for the all-in summit great Let's do it.

6:46All right. So then you described your project, a website for the Olin Summit, which is a tech conference from the hosts of the Olin podcast. And then you described the style you want to go after. Oh, okay. So how about cyberpunk? Cyberpunk. I like this one. Cyberpunk. I guess also techie. techie loss uh classic hollywood and uh modern post-modern los angeles i'm just throwing out wild stuff here let's go let's go for it um and then you generate you go get your second coffee of the day or whatever that is um and what is happening here because we just filled out a simple form with i think four fields or so three or four fields and then the ui wizard is doing something right now what is it actually doing so we've built some of the building blocks you know from 2018 and we have an engine to understand the concept of layouts within user interfaces we have a model that can understand styling and then all of these kind of like work together um to basically generate the entire thing and then we use a large language model to just understand the text and generate text so now we have our all-in website all-in summit the future of tech um buy tickets i don't know this is not going to be accurate location san francisco i guess this is miami um but here we go but he gave you some text like chat gpt might are using chat gpt for the natural language model or a different one for the content the text in the content is actually a build with the gpt3 correct yeah got it okay um so you have access to the gpt3 uh api but you need to get access to gpt4's api shortly we have access we haven't integrated yet in the tech you can see right now okay got it and so you'll you'll this will get dramatically better um as it keeps going from there but it gives us a bunch of different website designs um and these are classic website designs with a contact form and uh just various amounts of copies so this is going to start and it has here one of them looks like a more like a an itunes or a spotify kind of interface so you're just throwing different interfaces at us with different styles yeah yeah exactly and so once you're there you you're there you can basically modify everything customize it to your liking the goal of wizard is to make design accessible to everyone so the entire editor will thought of you know thought as it needs to be easy to use if you can use powerpoint you can use wizard and so the the goal of other designers really to overcome the blank page syndrome that so many of our customers are facing every day i want to get something to start from and instead of the blank screen for like you know an hour and it actually pulled in illustrations where did these so we're looking at these illustrations and it has like a nice stock library i would say images of people walking around with their laptops and talking to each other um and uh you know some you know people like sort of collaborating around a desk it looks like these illustrations were those made in some stable diffusion or something are they unique images or are these pulled from the web from clipart libraries?

10:18Where did the images come from? Really good question. So if the model actually also make his own prediction, whether to use our stock imagery or whether to actually generate its own. And so these illustration we see here, were completely generated by the model to kind of fit the conference theme. So these are actually completely generated. We we've basically fine tune on stable diffusion to do that. And then you can see others that were pulled from Unsplash or you know, some of these online libraries. Got it. Now, it seems quite random here. As a starting point, it doesn't feel like exactly what I asked for.

10:54I don't see any cyberpunk, for example. So I have to give it a second round, a third round and kind of do some reinforcement learning here in my mind. Is that built into the product yet? Could I pick one of these and say, Okay, now, let's work with this one. But let's add some cyberpunk elements to it, including the city of Los Angeles in a as the background. Can I just talk to it like that? and give it more prompts and edit the screen? Or am I now just left with, okay, now it's time for me to start typing and moving stuff around? No, so what you can do, you can't like type a new prompt directly in the editor just yet, but you can use all the other AI features that we've built over the years.

11:31So you can also say, for example, we could actually go and pull out the All In Summit existing website. I use that as a prompt. And use this as an example. So All In Summit i guess you already have one for last year dot com i think it should be always yeah um and you can take take a screenshot of this and then go in with it and say you know what i want to create a new screen from the screenshot and then basically start from again something else than just a blank screen and this enable our customers to just you know not reinvent the wheel when you already have something to work for it with oh and port it as a mobile screen let me just do this again we want to build this for the web um fascinating so i could pick five websites i like on the web i could pick five other people's websites and say hey these are ones that are inspirational to me let's start with theirs so and that wouldn't be necessarily stealing it would be a starting point maybe and you'd have to make it your own starting point yeah exactly and so you know here the ai is doing the work in the background and then once it's done we'll basically have all the elements from the screenshot, editable, customizable.

12:39You can just change the color, change the style, change the copy. And so if you are a marketing team or product team, it's just so much easier to just, you know, kind of like modify what you already have, tweak it to your own needs and then, you know, ship faster basically. So you're going to make this so that each element I could do a prompt for, right? That's the next shoe to drop. So instead of doing like a starting prompt, I could say, okay, let's go to the contact page. It opens the contact page. Okay, let's make the contact page include your Twitter handle and your LinkedIn handle and boom, it adds those elements.

13:10And I could just, you know, rework one page at a time with more prompts, right? Or generate one screen like, okay, no, we have build this flow, I would like to have the signup flow, you know, improve with more, you know, attractive color, for example, and it will just generate that one screen for you. And so you can see here, we've imported the screenshot from the all in summit like 2022. And we can go ahead and just modify everything the text, the imagery. and you know it would have taken some a long time to recreate this in the design tool if you don't have access to the to the to the source file but now we can actually go ahead and change absolutely everything and we have done this in under like you know 10 seconds

13:52imagine this you've got the greatest idea ever for a tech startup and it's going to change the world but you got a problem you don't have the engineers you need to make this a reality why It's hard to find engineers, right? Everybody's in competition for those great engineers. And you've got to manage your burn rate. You don't have unlimited resources like those big, slow incumbents. You've got to be efficient. So now imagine you had a partner who could provide you with more than 1 ,000 on-demand engineers. And these devs were vetted, experienced, result-oriented, and passionate about helping you grow.

14:26And they charged competitive rates. Sound too good to be true? Well, you need to head to Lemon.io right now. It is not too good to be true. Startups choose Lemon.io because they only offer handpicked developers with three or more years of experience and really strong portfolios. Only 1 % of candidates who apply get into Lemon.io. A couple of great launch founders have worked with Lemon.io and they had great experiences. So here is your call to action. To learn more, go to Lemon.io slash twist and find your perfect developer or tech team in 48 hours or less. and twist listeners get 15 off the first four weeks stop burning money hire developers smarter faster visit lemon.io slash twist fantastic this is very powerful people paying for this yet who's buying this are you and which market are you going after because it does seem to me that this is pretty early if i was a design snob and did this for a living i would be offended that you even built this product so some high-end person who is you know doing can't not doing canvas a high-end person doing figma doing this for a living is like this is going to be more work than me just coming up with my own ideas and then i could see people who are really early they might just prefer using squarespace picking a template making something beautiful and being done with it so where does this land who is this product for because it seems like you're maybe too advanced for people who just want a simple website for their portfolio or restaurant or project?

15:58And then maybe you're not advanced enough yet for the Figma design head to want to give up Figma or InVision. That's a great question. So our customers are basically the folks that don't have the skillset to use Figma or InVision or Adobe. These are, imagine like product leader, head of product, they're maintaining Slack and just want to have an easy way to just put a few ideas together, communicate to the team and then ship. Or a marketing team, you are spinning out new landing pages, new ads, and you basically want to have a quick and easy way to do that. And what we see every single day that these folks have a need to do design, but they don't have any idea on how to use Adobe.

16:37They don't have any idea on how to use Sketch. They don't know how to use Figma. And we basically provide them with a better tool to kind of like turn their vision into some shareable IDs that ultimately can be built by the product team. Fantastic. So all of the business people who are not part of the design team and the wireframing team can now start annoying the people who will eventually do the work by coming to them with a bunch of designs that are 70 or 80%, let's say, of what the designers will eventually get to. And so in the way, this is me as a CEO, sketching stuff on a napkin and then handing it to the design team and saying, hey, why don't you guys run with this from here?

17:16Yeah, pretty much. I mean, we see two things like some design teams are really frustrated to just get some really rough, really weird requirements and would rather get like a design file that's kind of like workable so they can see where where are you going with this vision. But we also see some team that are like, OK, our design team is already way too overwork. They don't have time for anything new. So I'm going to go create this in wizard and then I will just give it directly to the developer who is going to implement it using their library of component and basically save time to everyone. How do you make money?

17:49Subscription, good old SaaS. People pay, you know, seats per month, depending on the volume of usage. First tier is at$12 a month, and then the business tier is at$39 a month. Got it. So if the 10 people on a product team want to use it, $400 a month,$5 ,000 a year, and everybody can make reasonable designs and go faster. So have you seen anybody be able to, let's say, reduce their workforce or maybe fire their outside design firm because this was good enough for the non-designers to make product? Have you had those stories yet start? And then what are your thoughts on, you know, the elimination of, you know, people having to have this as an in-house job function?

18:42Are you there yet? That's a good question. So what we've seen is startups that would basically not have to hire an expensive agency to build their first MVP. This has happened a ton. So it's not someone getting fired. It's someone not getting the job. Got it. So this reminds me of logos. Like in the old days, you would hire a logo firm. They would charge a minimum of$1 ,500, probably on average$3 ,000, maybe in high-end cases$7 ,500 to$25 ,000. a series a a series a firm would typically take of their three or four million they might spend one percent of it 30 or 40k building a whole design a logo so here instead of and now we know everybody just makes logos on their own to start or they hire somebody for 500 bucks uh and they're done exactly yeah and it's done and at least you get something to start from and then you can iterate on it pretty cheaply as well got it so normally what would people would a startup spend making this sort of mock-up design of their prototype.

19:43And then, yeah, we can compare that to what you're charging, which for a startup would be, I don't know, five people times 40, 200 bucks a year, 200 bucks a month, 2 ,500 bucks a year. Yeah, roughly. And we've seen everything from, you know, like agencies charging a startup like 10K to, you know, even more just to build like five screens MVP for a SaaS or like a mobile app, which is pretty insane when you, you know, you don't even know exactly what you're building yet. You just want to explore ideas, get the first few users or paid customers. so it's yeah i like that i think you figured out who your customer is this is this was the big question i had coming into our interview was who is the customer here i think actually that makes a lot of sense because there are design firms right now saying two thousand dollars for a screen we're going to spend three or four days on that screen we're going to spend a month on your project we're going to put you know we're charging you ten thousand dollars but we're going to put a hundred hours into it we put two hundred hours into it it's going to be like 50 bucks an hour for us that's barely a living that's a hundred thousand dollar salary like that's what a great designer or at least a very good designer should get but what they don't realize i think in this calculation is now there is an easier way and then your customers just like the logo if the logo is 80 of what it will eventually be that's fine for now that 500 bucks is fine for now the 2500 bucks they're going to spend with you is fine for now this is a very smart strategy and it's not how did you come to figure out that this was your customer?

21:04Maybe let's talk about that. And you also used Raul Vora from Superhuman previously of Reportive, an amazing founder I've invested in twice and LP in his fund. He had a really great product market fit machine blog post he did, you read that. So maybe you could talk about how you found product market fit, how you made sure it was strong product market fit and found this group of customers. So actually, the product market fit engine, although Rahul has been the one kind of like turning into a machine, the core question of that survey actually was done by Sean Ellis, which is like asking your customer, how would you feel if you could no longer use the product?

21:45And if at least 40 % of them will say, I will be very disappointed if I could no longer use the product, then you have a measurable sign of product market fit. And early on, we've built an alpha, we've built a beta, and we've invited a ton of people and we've measured that very number. How many people answer very disappointed to how would you feel if you could no longer use Wizard? And then we use this as a way to basically segment customers, ask the right question and basically iterate. When we started this company, we thought we were building a tool for designers and developers. And then we realized through this journey that actually people benefiting the most from the AI automation we were building were the non-designers.

22:26basically everybody else in the team so it's been kind of like an interesting journey building for the wrong people and ultimately finding that there is a market for the product we've built um but no although this was a great way to product to measure early sign of product market fit or like rhino or ultimate product market fit check or validation is basically long-term retention how many people keep using the product over and over again after 12 months after 18 months after 24 months um but this engine was amazing to start from when you don't have a lot of data and you kind of don't really know how where to go amazing yeah it's it's such a great question to ask and to just be absolutely brutal and honest with yourself if you've hit it yet um how depressing is it to deploy this inside of a company talk about how the emotional wreckage the reality of strong product market fit is in in terms of like us as a company trying to find product market fit you mean yeah because it's one of these things where it's really hard is it not when you don't have it and it takes years to get to it so maybe you could talk about you as the leader of the company with your co-founder keeping your energy level high in the face of day week and month after month not having strong product market fit not having a product that people would care if it went away yeah and then eventually do you feel like you've gotten there that people actually really care uh that's something i was going to say like what's interesting about product market fit is that it's really hard to know when you know if you have it once you have it you know right you can see it in the data i just it just works and so when did you know you had it we knew we had it when we opened the beta in february 2021 and we saw the adoption and then we thought we and then you know you think you have it growth is is on the right direction and then you you hit another infliction point you're like oh no that's how product market looked like we thought we had it but it was actually just the early sign um so it's actually interesting how it goes in like phases um so that first wave was just people signing up downloading and then the second phase was this group of people who were just using it a lot was it the that turned it over for you in the second phase people using it a lot people telling us that this is too cheap for what the value they get oh really that's an interesting sign so people telling you you're idiots before your pricing model yeah that was really interesting to hear as an engineering like background founder you know you're always too scared to charge too much you know i'm a developer i don't pay for anything i'm i'm spoiled by the market giving me free stuff yeah so realizing that oh people are willing to pay more for what we offer that was a good like inflection point like okay we found product market fit did you ask a follow-up question to those people who said it was too cheap like why do you think it's too cheap what would you pay for it what was their reasoning what what did they say i'm curious because this is something i have been working with the founders we invest in we invest in 50 to 100 new teams every year between founder university and the accelerator and direct investments and one of the great lifts we have as a firm is having candid discussions about pricing and man sometimes we will see somebody increase their price and usage goes up and adoption goes up and more people are interested in it because they were starting at ten dollars a month and people didn't see value but when it was 199 a month people put more effort in to using the product learning how to use the product they were more committed to it so maybe you could talk a little bit about your price discovery that's something that was the most surprising to me is how much psychology is at play when you discuss pricing and actually just we when we started we just had free and pro and just by adding a more expensive business tier had an impact on how many people convert to pro because they felt like okay there is a more expensive package i'll just sit in the middle just this trick helped also convert more people to pay and then to answer your first question when we you know got into a conversation with a customer and they were like this is too cheap of course we'll ask you know what would you pay for it and and then try to figure out exactly where would be the falling apart like how much would be too expensive how much would be too cheap what do you think would be the right price and then when you do this for you know over a thousand of customer you start kind of have a ballpark of what the price should be but you know full disclosure our current price is still lower than what those discovery highlighted we should charge for because you know it's always a trade-off between charging and optimizing for revenue versus slowing your growth so it's kind of like i mean pricing needs to change on a regular basis but it's always hard easier to say than done is you know at the price you're at it should be sustainable for your company you have a company you told me of 45 people and there's definitely you know 10 20 30 000 seats that are out there waiting for this product at 40 bucks a month and if you can get but 10 000 of them uh you know 10 000 times four dollars is 40 000 and at a zero if it's 40 bucks that's 400 000 and that puts you right in the break-even space with 45 employees in my experience so somewhere between 10 and 20 000 paid subs your company's break even you're playing with the house's money you raised a bunch of money uh so you'll be in just absolutely great shape so i would just say the march to 10 000 and then you get the data right i mean that's the other thing if you charge too much you have less users you might be more profitable but then you get less data yeah and you get less use cases so it's one of the things that has influenced me a lot is watching what happened in skiing over my lifetime it used to be 60 then 80 then 100 for a lift ticket i know in europe it's much different right they way cheaper in europe you guys are insane bucks a day pretty much like and let's pull that would be a good resort like ski skiing is so expensive in the u.s it got so crazy if you walk up to a lake tahoe or a colorado resort now it's you know 175 250 dollars per day but at the same time they came out with the epic pass and the icon pass which are for just a local area just if you did colorado local or lake tahoe local might be four five six hundred dollars for the entire year in other words on day three you're break even and then there are like no blackout dates and it's seven hundred dollars let's say if you ski 10 days a year 20 days a year um with even one of those passes you're way into the break even and it's just extraordinary how it's changed skiing now there's so many people at the mountain um and they get the money in advance so they've discovered their own sort of sass model where you have to buy these before the season starts which gives them they don't have to take loans or money to maintain the mountains they get the money in advance from the user base which was elon's big thing with taking pre-orders for cars right so it's all kinds of great learnings here i think i think the same with conferences that always sell the tickets you know uh on on the year before the actual one year out yeah yeah exactly yeah it's a great way to do it it works yeah yeah listen we work with super early stage companies at my investment firm launch pre-series A, maybe you got a couple of thousand dollars a month in revenue.

29:35You've raised a couple of hundred thousand dollars, maybe a million dollars, right? That's the early days, year one or two of a startup. And I'll be honest, a lot of times startups, they don't have their insurance. They haven't set that up yet. They haven't set up their accounting properly. They're getting things cleaned up. In fact, I was recently had a great startup, but they didn't have D &O insurance. That basically protects your directors and officers. That's the D, directors, people on the board, officers, the people who work at the company, right? Directors and officers, insurance is super important.

30:05So what do we do? We send them right to InBroker. Our friends over at InBroker are a business insurance company that's built specifically for startups. You just fill out a simple application, right? And then startups get four quotes for four lines of coverage in 15 minutes. Four quotes, four lines of coverage, 15 minutes. Easy, breezy, lemon, squeezy. That's right. They connect you with one of their expert brokers for unmatched service that goes beyond your policy. And listen, you might think, oh, it's too early to have insurance. It's not that expensive. It's not that complicated because Embroker makes it easy.

30:38So here's what I want you to do. Try Embroker today with the code TWIST and you'll get 10 % off their startup package at Embroker.com slash TWIST. That's Embroker.com slash TWIST. We love Embroker. I use Embroker. They're an amazing team. They do a great job for startups. whether you're in year one or year five go use in broker.com slash twist so what do you think about competition now that you have um you you've already started to see adobe has their own stock image um generative ai product canva lets you replace images and do some ai on them figma will obviously release things it's part of adobe adobe's got a big push on ai so how do you think about competition um when the world now is on to ai and applying it to every vertical and certainly applying it to design because generated design really is a function of topography copy images illustration and layout and not much more right so you know all of those things logos stock images those things are all being iterated on so how do you survive and carve a niche for yourself when all the big lumbering elephants are right behind you and they're going to release products i think the issue i have with products that are plugging an ai element on top when it's not built in or like you know where the product is not designed for being ai first you create this really weird user experience shift where the user has some really easy to use feature, the AI powered feature, but then I have to cope with all the complexity of the existing products.

32:17And I created this really weird, you know, dissonance in the use experience. And so the way we believe we have an advantage is that our assumption and our vision has always been to build, you know, an easy to use tool. And then the AI element just becomes natural as an extension to that. So I guess if those companies you just mentioned need to really want to enable AI inside of the product, they're also going to have to rethink the rest of the user experience because ultimately, who's the user of this AI? You mentioned it earlier. If you're a professional designer, do you want to just type text or would it be faster if we just create a screen from your existing library of components?

32:54And so it seems to me that there is a dependence between the user of this AI enabled features and the rest of the product and who that serves. Actually, Canva is a good example. I think they are the one that can nail it because they've always been and easy to use software. And we are often called the Canva for product design. We do the same that Canva did for marketing design, for product design. So they're probably the one that I think will just execute the best on this front. So you obsess over your user base. They will probably find Canva too simple and they already think Figma is too complicated, so they probably want to have this AI first.

33:37And the great thing is you can really cater to them and their needs in a way that the other products maybe don't. And I guess something that's also interesting that if you are Adobe, you sell to designers. But if you make design easy, don't you replace or get your own customers out of job? You know what I mean? So there might be this weird like you cannibalize your own customers if you are... It is an innovative dilemma. You're absolutely right. Wow, what a great profound insight, Tony. if you're adobe the more you incorporate ai the less you need designers and the less designers there are the less or the the two arguments would be the more like what you're saying there'll be less designers because or less dedicated design positions because everybody can do design just like now you don't need to have copy editors everybody can do copy with chat gpt4 i'm pretty convinced between four and six the idea that you would need to hire a blogger or a copywriter or somebody to write a press release that's out the window i think anybody in the organization will be able to write a serviceable press release just like everybody could write something without spelling or grammar errors thanks to grammarly and you know just the grammar tools built into word processor so this is like a brave new world um and then adobe's customer base would have to shift from designers to everybody that's that's kind of a hard thing to do i think right while Yeah, time will tell, but that's going to be tricky to solve.

35:01Man, it is going to be a brave new world and super excited to see what you come up with when chat GPT-4 gets incorporated into the product and plugins. So have you started to play with the plugins that are available for chat GPT-4? And do you have any favorite ones or crazy ideas that have inspired you since you started playing with the plugins? Actually, we went really deep into the plugin API because we wanted to figure out how to build our own. But I must admit, I haven't used the plugin myself much. I'm using ChatGPT every day, but not the plugin just yet. Do you think ChatGPT is going to replace search?

35:36What do you think the impact will be on search? Have you found yourself using Google, replacing a Google search with the ChatGPT4 chat session? Oh, that's an interesting one. I find myself oscillating between the two. I mean, for programming, I use ChatGPT. I mean, I need to tweak things, but I will need to tweak things from a Google search either way. But I find the level, the communication experience is so much better. Like, hey, tweak this. That line was wrong. Maybe you should actually, you know, I can respecify my requirements, then code changes. When I'm going for very specific information, I'll still use Google.

36:13Got it. But I think what I find interesting as an entrepreneur is that, of course, SEO and content is a big part of our acquisition play in the long run. and no the big question is like do we keep optimizing for ranking high on google search how do we start optimizing for ranking high you know in a database that ultimately will be used for training chat gpt it's kind of interesting as well that the the entire game of seo is also changing for most companies um well and what would the what's the point of seo if people are going to stop searching and just get answers and then the answers don't have links in them so oh yeah it it means you know and i think that's going to be the big rub that's coming soon which is uh i think chat gpt4 is going to be forced to put in citations to where they got information from as part of their reasonable uh negotiation with content holders because they just served up the entire web from what i understand um but they're not going to be allowed to use quora reddit youtube transcripts tweets facebook craigslist amazon reviews all of those are going to be siloed out and you're going to have to license that data so chat gpt's performance could go down because the model will have to remove that training i believe the lawsuits will be such that they'll say to chat gpt4 we know you learned a lot now you have to go back to 3.5 and redo your model but you got to take out reddit you got to take out quora you got to take out wikipedia you got to take out this open source project and you got to redo chat gpt4 in chat gpt5 with citations and with a licensing deal.

37:46And I think that's going to be the really interesting part. Have you played with Bard yet? Do you have any early thoughts on bard.google.com? I love Google product, but Bard is not one of them, I must admit. It's so shocking to see that Google was the one building the transformer architecture and then they got completely outpaced by, yeah, by OpenAI. It's kind of sad to see, I must admit. Especially with all those employees. All those employees and they can't build a better product yet. would you say bard is where 3.0 is of chat gpt or 3.5 obviously it's not four it's hard to tell it yeah well as a as a you know as a machine learning researchers is the fact that well google might actually be very secretive on their research now because of this and that's kind of like penalizing everyone so yeah that's kind of like i think the side effect that people don't talk enough about like are these companies still being willing to share the research with the open an AR community if they are scared that someone will use it to basically kill their business.

38:49It's a back to innovators dilemma. We had the innovators dilemma with Adobe with designers and more they innovate, the less designers will be necessary in the world. Possibly is one argument. And then here, the argument is yeah, if Google does a better job with Bard, then people do less searching and then less ads get clicked on. Because if you get the answer, you don't need to click on an ad. except if they figure a way to just embed the ad in the in the in the in the results that's what i think they're going to do i think they're going to put them right in the results so when you do something hey i want to design a world-class logo it's going to be like from this book yeah world-class logos and this book and this article and those are all going to be paid links and then there'll be other you know somebody who bought the term logo design their ads and will say for more resources here's this this and this and it'll explain what that page is in a sentence so it'll give you 10 bullet points at the end of search that'll be like a really thoughtful description of the page that the ai will write with the link and those people will pay per click why wouldn't they pay per click to be at the bottom of your discussion right this chat is brought to you by these five related resources all right listen tony great job continued success and thank you so much if anybody wants to check out the product the url again is uizard.com all right europe-based company you're scattered all over europe from what we talked about and are you hiring right now any positions that you need to fill asap not at all we believe we have all the right folks until series b perfect all right i like it a little austerity we got the right team we just need to put our heads down and build it's not about the size of the team.

40:33It's about the impact the team has. All right, continued success. And we'll see you all next time on this week in startups. Bye bye. Trovata is a cash management platform that helps keep tabs on your runway, all of your financial data, and it prepares you to answer investor questions, right? And this increases your credibility when you're a startup. It's basically the cash command center for your startup. And it's going to let you analyze, report and forecast your cash position like a pro. And here's the thing. Many startups are using multiple bank accounts, sweep accounts, you know, to increase the FDIC insurance.

41:10Well, Travata makes it easier than ever to manage multi-bank data with a single source of truth. And Travata can scale with you from your seed round to series AB and onto your IPO. Fast scaling startups should not be managing cash in a spreadsheet. Cash is your lifeblood, so let's treat it that way. With investors like Wells Fargo and JP Morgan, Travata has pioneered a massive library of corporate bank APIs, and companies like Carta and Fanatics trust Travata to gain visibility into their multi-bank data. So here's what I want you to do. Go try it out, travata.io slash twist, to get started for free.

41:48And use the code TWIST for 30 % off premium features for one year like AI forecasting and reconciliation. Start managing your cash like a pro with Travata. That's T-R-O-V-A-T-A dot IO slash twist to get started for free. Okay, everybody, Tony did a great job on the show. You got a lot to think about. But I want to get into some tactical work now. Kelly from my founder university team is going to show you a nice tactical talk on how to build a CRM, a database of people so you can manage those relationships inside of notion. And by the way, if you enjoy these tactical demos, you're going to love the Founder University podcast, I'm going to put some Founder University podcasts at the end of this week in startups so you get familiar with them.

42:31But all those tactical talks are available on YouTube or your favorite podcasting app, just type Founder University, go subscribe. If you want to rate it and write a review, that would be super helpful to me. I'd really appreciate that. Or if you go into the descriptions here at this week in startups, you'll find all those links tucked away for you. Take some notes, pay attention, stick with us. This is a great tactical talk. Hey, y 'all, I'm Kelly. And in this episode of Founder University, the podcast, we're talking all about setting up a customer relationship management tool or a CRM for short.