This Week in Startups - Summarize Anything & Discover Connections with Recall’s Sankari Nair | E2010

Episode Date: September 18, 2024

This Week in Startups is brought to you by… Squarespace - Turn your idea into a new website! Go to https://www.Squarespace.com/TWIST for a free trial. When you’re ready to launch, use offer code TWIST to save 10% off your first purchase of a website or domain. Lemon.io - Hire pre-vetted remote developers, get 15% off your first 4 weeks of developer time at https://www.Lemon.io/twist CommandAI. Seamlessly integrate an AI-powered guide into your software, making navigation intuitive and interactive. Visit https://www.command.ai/twist to get a custom live demo. * Timestamps: (0:00) Recall’s Sankari joins Jason (4:15) Explanation of automatic taxonomy creation and adding content to Recall (8:11) Recall’s augmented browser feature and web content indexing (10:26) Squarespace - Use offer code TWIST to save 10% off your first purchase of a website or domain at https://www.Squarespace.com/TWIST (14:39) B2B applications, multiplayer mode, and capturing institutional knowledge (19:15) Lemon.io - Get 15% off your first 4 weeks of developer time at https://www.Lemon.io/twist (20:40) AI costs for summarization and model preferences (22:08) Personalized learning features and Recall pricing strategy (26:10) The concept of luxury software and jamming on potential features (31:00) CommandAI - Visit https://www.command.ai/twist to get a custom live demo. (32:34) LinkedIn integration and visual knowledge graph representation (35:26) Jason and Sanks play with the quiz section of Recall (40:43) Final thoughts on startup journey and investment opportunities * Subscribe to the TWiST500 newsletter: https://ticker.thisweekinstartups.com/ Check out the TWIST500: twist500.com Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcp Founders! Apply to LAUNCH: https://www.launch.co/apply * Check out Recall: https://www.getrecall.ai/ * Follow Sankari and Recall: X: X: https://x.com/getRecallAI LinkedIn: https://www.linkedin.com/in/sankari-nair-13a549a7 * Follow Jason: X: https://twitter.com/Jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis * Thank you to our partners: (10:26) Squarespace - Use offer code TWIST to save 10% off your first purchase of a website or domain at https://www.Squarespace.com/TWIST (19:15) Lemon.io - Get 15% off your first 4 weeks of developer time at https://www.Lemon.io/twist (31:00) CommandAI - Visit https://www.command.ai/twist to get a custom live demo. * Great TWIST interviews: Will Guidara, Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland * Check out Jason’s suite of newsletters: https://substack.com/@calacanis * Follow TWiST: Twitter: https://twitter.com/TWiStartups YouTube: https://www.youtube.com/thisweekin Instagram: https://www.instagram.com/thisweekinstartups TikTok: https://www.tiktok.com/@thisweekinstartups Substack: https://twistartups.substack.com * Subscribe to the Founder University Podcast: https://www.youtube.com/@founderuniversity1916

Jason, I've been talking about this knowledge graph, right? What is a knowledge graph? It's just a type of database. It's behind the scenes connecting everything. But what we also launched was just the visual representation of it. When you look at your entire knowledge graph, you know, it looks a little bit nuts. I've been using recall for ages. So this is just a full view of my entire graph.

It's nuts, you know? So beautiful. It's basically a combination of ingested content, but also user-led. So it's this perfect hybrid. Wow. This Weekend Startups is brought to you by Squarespace. Turn your idea into a new website. Go to Squarespace.com slash Twist for a free trial. When you're ready to launch, use offer code Twist to save 10% off your first purchase of a website or domain.

Lemon.io. Hire pre-vetted remote developers. Get 15% off your first four weeks of developer time at lemon.com. And Command AI. Seamlessly integrate an AI powered guide into your software, making navigation intuitive and interactive. Visit command.a.ai slash twist to get a custom live demo. All right, everybody, welcome back to this week in Startups. We have a great guest today. She is doing a really interesting company called Recall that is using AI to help you recall things that you've seen

were consumed in your browser and then build a language model around that. Her name is Sankari Naya. But you like to go buy Sanks. You went to our accelerator and you have a great demo of the product you're building. And we got very excited about what you're building because, gosh, I'll be honest, I love it. And I love the idea of collecting information while you're going around the web and then having that available to you in what we used to call a mind map or etc. So why don't we show the audience what you're building. If you're listening, we'll sportscast this.

But show us what recall is. Oh, yeah, absolutely. So we like to refer to Recall as an AI-powered encyclopedia, but what you're looking at right now is actually my personal recall knowledge base. And recall consists of what we call summary cards or recall cards. And so if I just click on one, Jason, basically what you can see is you can see a link to the full form content. In this case, it was a two-hour-long podcast by Huberman. You can see that it's been automatically categorized. and that category is based on the topic of the content, but it's also dynamic.

So if I have something else related to skincare for the listeners, I'm looking at a podcast on skin and health, I've got a tag called health and skincare, it'll be dynamic so that your taxonomy is always staying the same. Everything's fully edited. And if you scroll all the way down, you can see these what we call entities that have been extracted. And entities are keywords with a high relevant score. And these numbers next to it means that it's actually been mentioned several times

in my knowledge base. So you can see Sourcraft has been mentioned in this episode on Skin and it also relates to all of my recipes on Sourcrow. So if I just come back here, yeah. So you bookmarked that episode of Huberman talking about skin health.

It then wrote a summary of it. It then tagged it. And then you don't have to do all of that. So back in the day, we'd have something like delicious, and we would bookmark a bunch of websites and then keep them in collections. And it was very manual because you had to tag them yourself. Here, it's tagging it.

He mentioned sourcrow. So you have that tag. But if you click on the sourcrout tag, it goes into your entire knowledge base and it found sauerkraut in recipes as well. So it's automatically connecting all that knowledge, like your brain might. A hundred percent.

And that's actually what it's based on. So like a knowledge ground, off, it literally works the way your brain works. It's also self-expanding, which is why it's really great when you're researching a topic or you're looking to grow your knowledge on a particular topic. And Jason, you mentioned bookmarks, but actually one of the great things is you can actually just ingest all of those books marks directly into recall. It's great for the cold start, and those bookmarks get turned into recall summary cards. So you can already take what was historic and bring it into recall. Okay. So now we've got this beautiful taxonomy on the left here. I see you have

articles, books, business, business technology, cocktails, apparently. Somebody likes a cocktail now and again. Psychology recipes. So you can go find your best cocktail, talk about business, and then go into the psychology of it in productivity. So this is really fun. Now, did you make those top level categories on the left or were those automatically created?

Those were all automatic. So I haven't created any of them. They've been automatically created, automatically ingested. And so that's really actually we want to take care of. It's sort of just happening for you behind the scenes. You don't have to put a lot of effort coming in, curating, tagging. It's done. It's self-organizing.

Okay. So to put one of these cards in, if we were on a website, how do you get the card into it? Do you cut and paste the URL? Do you use a toolbar? Show the audience how you get that podcast in there because we're using a desktop here, obviously. For sure. So there's multiple ways. We've got a web app and a mobile app.

When you're on the web app, you can just copy the URL of any online content. So let's say, I'm looking at this blog. I hit the plus sign, I paste in a URL, and you can see that a summary has been created in seconds. I've got a tag, you know, this was business marketing product-led growth, it's fully editable, and I've got those entities. If I'm on my mobile device, I can actually just share. So I'm out and about, I'm on the go, I'm digesting content, I can share it.

And one of the most popular ways is with our browser extension. So if I go to, let's say, this podcast, it's around two hours long before I want to commit to listening to two hours of content, I can just hit this browser extension and you can see that it just generates a summary in seconds. The speed is also what I think is interesting. It's dependent on if someone's generated a summary for it before. And so we cash that summary. And so there's this great network effect where if you generate it, the next person's actually going to benefit from the speed. Fantastic. Okay. So once you have all this knowledge into this recall database, it's stored locally, it's stored in the cloud. Where does the data exist?

all these summaries. So this is stored in the cloud if you save it. If you save it, it's stored in the cloud. But I actually have a really exciting feature that I'll show you in a minute. I know Jason you were talking about, I don't want to hit save. I just want it to happen for me. And so we have I'll do a demo for that in a bit, but this is where it's just crawling and it's just going to be saving as you're going. You don't have to do anything. It's one of the connections for you. Let's show right now. Yeah, this is a great idea. Let's do it. So you and I had a conversation, just so people know, you went to our accelerator. I don't know. Did you go to found a university the pre-accelerator too or no?

No, it was just the accelerator. Just the accelerator. So we met you. I think you had like a prototype. I don't know if you had any paying customers yet when we met you, but you were very early in this journey. And I said to you, you know, gosh, AI is so smart in getting smarter. It's kind of like having, I don't know where you put it at, but I kind of feel like

it's a really smart high school senior or like an average college student. I was like, wow, you know, if I spend more than X amount of time in a brashire browser tab, by definition, that means it's important. So I would like to have a setting when I'm just browsing, if I do something for more than a minute on a page or 30 seconds, like watching a YouTube video for more than a minute, it saves it. If it's a newspaper article, if I scroll up and down, it saves it. If I, you know, stay on it under 15 seconds, maybe don't save it. And then maybe have it say saving in 10, 9, 8, 7, 6, like you do on Ways. And then I, you'd have a very big knowledge base

and then you could do things like you think I remember an article but anyway show me what you built this is one of those VC specific feature requests Actually not It was on our roadmap It was on the roadmap, okay good This was on our roadmap

I mean I think it was definitely from our jam That we got a lot of ideas from that But this was on our roadmap And we did accelerate it Okay here we go So let me share my screen Yeah Basically this is just a normal article

now in this example. So I've been struggling with insomnia lately, so I've been doing a lot of research on sleep, how to improve my sleep. If you open just a normal article, you know,

this is just what it looks like, right? Yeah. But if you open it with our latest, what we're calling the recall augmented browser, you can see these slight links on the page, right? And these links, they're subtle, right? We don't want to disrupt your reading,

your reading experience. There are actually links to things you've seen before. And it could either be things that you've saved or it could just be things that you've seen before. So I actually was shocked when I opened this and I was playing around with the augmented browser and I was like, Matthew Walker, who's that? I literally never seen that person and I've actually seen him four times.

And I hadn't even realized that. So what you just did for people who are listening is in this article by Huberman on Sleep on his website, he mentions Matthew P. Walker. It is in reverse type, so it's white type on black. so you can see it very clearly. When you hovered over it, it brought up a recall card

with other times Matthew Walker appeared in your knowledge basis. Am I correct? That's what just happened? A hundred percent. And not only that, you can actually see exact little snippets that you might have seen them before.

So you can actually go and see this was a book that was recommended. Or in this example, REM sleep is mentioned. Yes. And if I want to skip to a particular element, it actually zooms into where was REM sleep mentioned

in what was a really long article. This is more than I expected you to do. So this is interesting. I am reading an article in the New York Times or, you know, on a substack or something. The knowledge base, recall in just that article, looks in the knowledge base, says what topics or keywords in this article match what Jason's already consumed so that we can help him recall other instances of that, which I would think would reinforce memory and learning. Absolutely. And we just want to bring that experience of the knowledge graph seamlessly, effortlessly, and in the actual content ingestion stage, because that way it's just happening.

It's just working for you and it's part of your workflow without you doing any work. That's amazing. I don't care if you're building enterprise SaaS, a verticalized AI startup, or the next great consumer app. No matter what you are cooking up, you have a brilliant website. It's the first thing that people see or read from your startup and it has to be beautiful. you've got to make a great first impression. People do judge a book by its cover. It's that simple.

That's why we have that expression. But since you're building a company, you're strapped for time. You know that. And wouldn't it be nice to have some world-class help? How about a beautiful website built to your specifications quickly and easily? It's 2024. If you don't already know, HTML, CSS, and JavaScript,

that ship has sailed. Get on Squarespace and check out their Blueprint AI product. You answer five questions, and Squarespace's new AI tool is going to build you a gorgeous website. It's so simple. It's so slick. It's just brilliant. It's the future.

Even better, you don't need to worry about what's under the hood. Analytics, SEO, payments, memberships, e-commerce. You can do it all. It's all included, too. So here's your call to action. Check out Squarespace.com slash twist for a free trial. And when you're ready to launch, go to Squarespace.com slash twist to get 10% off your first website or domain purchase that's Squarespace.com slash twist.

Now, does this automatically? Mac, we put the article into my recall, or I just still have to do that manually? So we're playing around with us. I'd actually love your feedback. Whenever it's something that, so you can see some of yellow links and some are darker links. A yellow link is just something you've viewed. So you've not actually saved it, and this will just take you to the original article. A black link is one that's saved.

So we're playing around with ideas. Right now, anything that you've just viewed, it's all local. None of that's going into the cloud. And we could give you the option to then view that in, recall. Maybe it's index, maybe one enrich it. So we're playing around. We're jamming on on sort of how we want to bring this experience to the user. Obviously, it could get, it could get nuts because there's so much you're ingesting. And so that's the balance that we're trying to slack is trying to strike. So I wonder, here we're doing like a mini jam session,

which I do a founders all the time, um, you ever use Waze or in a Tesla, if you're driving home and you find what home is or opposite work is in your navigation system, it assumes and it's been watching you that when you leave in the morning, you're going, to work and that when you leave in the late afternoon or, you know, at five o'clock, you might be heading home. And when you start driving, it just says, you know, navigating home and then it gives you a countdown, you know? And it says 10, 9, 8, 7, 6, boom, and then it starts navigating.

If you don't hit stop and say, don't navigate. So what I would say is if I scroll through an article and I get all the way, you know, let's say I've spent more than 30 seconds on the article, a minute on the article, whatever, I set. It could put a little status bar that says adding to recall in 10 seconds, nine, eight, and you say don't. But if you do, you just watch it go down, boom, and now it's in recall. That's the way I would do it. So then I don't have to take an action. If I'm interacting with it for a long time, it gets saved unless I stop it from being saved, right? It's like an opt out. It's like a very opting out of it, yeah. But I could set it. If I interact with the webpage for more than a minute,

I want to save it. Under a minute, I don't. Then you might have a lunatic who says, you know what, Any webpage I wind up on, I want to download the whole web page and I want the summary. And I think that's like a really power user feature is because web pages disappear. There is nothing illegal in fair use about you making a copy of a web page locally. So as long as you're not storing it in your cloud for like everybody's use, but if I were to download a New York Times article, I can save that, right? I can print it. That's allowed.

That's fair. So I like the idea of having the option to store original article and put into the knowledge base. So then if the article goes down, I have actually the full text of it. You have the backup. Yeah, I love that, Jason. I think this feature that you're showing here where it's connecting my previous consumption to this is great. I would like to also understand my Gmail or superhuman account. What I would like to also have is index my Gmail or hit my Gmail index and tell me if those topics are also in my Gmail. That's super crazy. And now that's a big amount of data as well. Absolutely. And that actually, Jason also opens up more of the opportunity for B2B.

And that's really part of the roadmap and a direction that we're heading towards. Yeah. I think B2B and I told you this earlier, multiplayer mode is going to be phenomenal for this. And when you go into multiplayer mode, what I've seen when founders do that, and you work at Uber for many years, so just so people know the. Yes. Did you see the Travis interview from All In yet? I did.

What did you think? Yeah. I thought it was great. I really thought it was great. Yeah. You know, one of the really interesting things is when you do a group order on Uber Eats or, you know, you add stops or you share, you know, like your affiliate code, those viral moments really help things grow. Multiplayer mode, this is single player mode right now, but if I put it in multiplayer mode, I could invite the investment team to this. And we could say, anytime you see something interesting, share it with the investment team.

Now, I'm on the investment team, and I see that Kelly, Jackie, Bianca, Lucas have shared something, and I see them putting comments on it. Now I'm like, wait a second, my team found an article about product market fit. I didn't know about it. And we added to the product market fit archive automatically. Then when we have a discussion about product market fit, instead of everybody digging through Slack or their superhuman, Gmail, whatever they use, Outlook, it would just be right there. and we go into recall and say, hey, what is everybody contributed to the product market fit,

you know, repository there? And that would be really powerful. Oh, absolutely. We think about that as like a company, a collective company knowledge graph. Yeah. And this is a company, exactly. It's almost like building your company second brain. Because people always talk about second brain, but company second brain, I think that's, I think that's when it gets really, really exciting. I have become obsessed with not losing knowledge in companies because, you know, institutional knowledge is important. But, you know, sometimes people stay with a company for two years. Sometimes they stay for 10. And then all this knowledge gets built up. And then my partner, Mike Savino and I are having this debate over, oh, my God, this person's leaving. We're going to lose all that institutional knowledge. He's old school, you know, and really likes to keep people for a long time. If people don't want to be at the company, I don't like to fight and keep them the company just because they have institutional knowledge. If they want to go on to their next adventure, I don't want to stop them because I think that's unhealthy. And so I've tried to capture institutional knowledge. Notion, Coda, and Slack are the way we do that, but this is another layer of that.

And so, just to give a tip to the audience, the way I do this institutional knowledge is I do playbooks. And so in your company, you might have a playbook for how to deal with customer support, right, or how to manage investors. And so what I do is I have my Athena assistant, go to Athenawow.com and we're investors in that company as well. I have the Athena assistant record a zero. Zoom with two people and they just have a dialogue about how they do a certain practice. Then we take that practice and we transcribe it automatically gets transcribed and summarized by Zoom and by Notion and by Cota. They all have that built in now. Then we take those summaries, transcripts and make a playbook. Hey, here's what this video is about. You can hear this person talk

about how to manage LPs, how to write an update for investors. And then we have it there. Next person comes to the company. It's in Notion or Coda. They can just pull it up. watch the original video from somebody who hasn't been at the company for five years. But five years ago, they made the playbook. And I think that what you're building is so interesting in that regard. And the ability for me to attach my notes on it and then share it with the team and have comments, that makes it very hard for people to unsubscribe from the service. So for me to rip out Slack would be very hard because I want to be able to search Slack

for a discussion we had about your company a year ago when we accepted you to the accelerator. I need that conversation. And so when you make this multiplayer, if people start putting comments in it and then you have 100 people in your company who've used it, only 50 of them are still at the company. And it's five years later. And you've had 50% attrition,

which is kind of natural over five years. You'd still have all their knowledge there of what articles they read and why they thought this was important. So I just think this is so, so compelling in the future of knowledge sharing. I'm going to get off, thanks for coming to my TED talk. But I'm so enamored with your company.

But multiplayer will be everything. Oh yeah, I mean, it's one of our biggest requested features. We actually surveyed our users. I don't even think I told you this stat. But 58% of them are using recall for work, for learning, for productivity, for research. And so, yeah, super excited to expand in that way. Right now, startups have to do more with less. We all know that. And founders have to be smart with how they deploy capital. Investors are very tuned in to being capital efficient. So if you need great tech talent, but you don't have to be smart. have the time to interview dozens and dozens and dozens of candidates you need to check out Lemon.io. They have thousands of on-demand developers to choose from, and these devs are vetted, and their experience, and most of all, they're results-oriented. They're going to get you the result you're looking for. They're not going to leave you hanging. And guess what? They charge competitive rates. Great developers can be incredibly hard to find. We all know that, and when you do find them, it can be hard to integrate them into your team, but Lemon.io will

handle all of that for you. Startups choose Lemon.I.O because they only offer handpicked developers with three or more years of experience and strong portfolios. In fact, only 1% of candidates who apply get in. And if something ever goes wrong, Lemon.i.o will get you a replacement ASAP. A couple of launch founders have worked with Lemon.com. I.O. And they've had great experiences. So here's your call to action. Go to lemon.com slash twist to find your perfect developer or tech team in 48 hours or less, and Twist listeners get 15% off their first four weeks. Stop burning money, hire developer smarter, visit lemon.io slash twist. What does this all cost, and what does it cost you to, you know, share this?

Because I told you, like, listen, and when you started, I think these APIs were pretty expensive, like summarizing an article or a podcast, a couple of pennies, right? So it doesn't seem like a lot until you have a thousand people on this doing 100 items a day. And now you're starting to rack up some costs. How have the cost changed? What do you charge for this? Is it possible for you to lose money on this to, you know, to do all this backing up and all this processing, explain for people who are starting AI companies for thinking of it here

on this week and startups, what the cost structure is and for doing all this stuff? Oh, yeah, for sure. You know what's so crazy is you asked us a similar question all the way back in March. And the answer was around 10 cents a summary. And I just checked in on it and it's not less than a. So it's crazy how much more affordable these models are getting. It also depends which one. It depends how you're fine-tuning.

We also were one of those lucky- Which one do you like best? Take a unpack that because people really want to know the details. Which ones have you tried? Which ones work great? Which ones work the same? Which ones are overpriced? Which ones are great value?

So we started with Bison, which was Google's one of Google's first. We then switched and now we have a combination of fine-tuned ones. But for example, for our question, so if you could, come to a, let me share my screen on this one, if you come to a card and let's say I was, I was on this episode, I can save it in here, and I can run a little quiz, this quiz is then chat GPT4. This is GPD4. So, GPD4 for the quizzes. We've got a combination of models for the summaries. We didn't even explain this. This is incredible. Yeah, I was actually really excited to show you. So basically, once you've saved content in, you can just run a quiz. And once you run a quiz, it gives

us a signal. The signal is, I'm interested in this content. And so once you run a quiz, we actually add it to what we call the recall review. This is Jason what we actually got product of the month for in June. And the recall review takes all your cards that you've shown us the signal that you want to review. And we create a tailored schedule of learning for you. And it's based on your own learning curve. So if you're consistently getting something correct, we show it to you less. If you're getting it wrong, we show it to you more frequently. And so that's the space learning element of the recall review. So you can literally take a podcast and then create questions based on the podcast. Wow. And chat GPT4 is really good at that. Yeah. This is this whole personalized

tutor that is going to change the world because it's adaptive, right? If you already know all this stuff about peptides, you're not going to ask peptide questions, but if I don't know things about skin and sleep, it's going to, you know, ask me those questions. until I really understand sleep. Exactly. I'm sorry, I know I completely then missed your question on cost, but basically it's gotten cheaper.

We use a bunch of models for summarization and entity extraction is our own, and then GPD4 for the actual quiz. And our cost structure right now, it's $10 a month, or if you buy the yearly plan, it's...

I feel like I knew you were going to... Too-chee. I mean, think about the value you're providing. We're visiting our cost. Yeah. So what I showed you. you now on the augmented browser. It's not launched yet. We're probably going to launch in the next few weeks. And I loved what you said about the luxury first model and that we're

actually maybe in the luxury software world. Yes. We're ideating on pricing for the augmented browser. I really think it's $100 a month product and you should store everything I do and you should explain to the users. This is for CEOs, VCs, senior executives who can't afford to forget anything. It's $100 a month. We capture every single thing you do privately in your browser, encrypt it, you know, and it's stored locally, it's stored in the cloud, and it's, you know, high fidelity and high speed. If you look at superhuman, that appeals to the top 1% of email user. You know, if there's 5 billion people using email, that's a big number, you know, 10% would be 500 million. So 50 million people,

you know, superhuman's kind of essential for. Now, if you have multiplayer mode, like I just put in my company, Super Bowl. human, but in the luxury software category, has the ability to create essentially a thread around an email. So if you were to email me, say, hey, we're doing our round. We want to know if you want your pro rata. I could then, in a square box, say, at Jackie, at Andre on our team, have we started the diligence review here? Please let me know what the valuation is and the revenue and our investment history. But they would see your email. So it takes the email as a private email to me, now exposes it to those three or four people and we started discussion on it.

So we don't hit reply, like one of my new researchers did. I heard about this. I heard about this in one of your episodes. This founder has ugly baby syndrome. Yeah. That was right. It was literally this poor, this poor really talented kid who works for me.

Anybody under 30s and kid now. He's like, I just wrote back to the team. Am I supposed to fire for this with him? on there. Oh my God. As a joke, as in like, he didn't look at the CC list. Right, right, right. And just responded to a founder.

He has an ugly product. Therefore, he has ugly baby syndrome. Yeah. And he laughed. He said, I guess you're supposed to fire me. I said, or it's a learning lesson. You'll never do any stupid like that again. That's where he kept him.

Yeah. He's probably obsessing over who's in CC, whatever he clicks in. Yeah. But luxury suffer is a very interesting category. Explain to the audience why that's an interesting to you as a founder, to the concept of luxury software. Why is that an interesting approach to building a startup? I mean, it really does go all the way back to Uber, right?

And Uber launching Uber Black and the way that that entered the market. And then when there was enough demand, then making it accessible for everyone. And I feel like it's a brilliant approach. It's definitely something that we're jamming on. And then when I think about recall and luxury software, where it comes into play is the cost element as well, right? Because today to launch something that's scraping everything, that's doing all of this processing that's providing you with this value add, the reality is the cost

add up. And so we want to strike the balance of, it makes sense from a business perspective, and it also makes sense from like a use case perspective. And I think when we first launched this luxury first could be a great way to go, especially when it also comes to then finding our product market fit specifically for that bundle or that package that we're looking to launch. Fantastic. Yeah, you know, I'm building an app for Inside right now. where it's like summarizing the news. I'm using humans and AI. And I've been really thinking about luxury software,

what would a person like me pay $100 for? And I came to the conclusion, I would pay like $100 if I could talk to the author of the piece. So like imagine you're reading a story at a publication and I had a follow-up question. I would actually pay $100 a month if I could actually know I could, yeah, do that. So in here, I think one of the interesting things about luxury software is if you sent me a summary of what I did yesterday

or a summary every hundred web pages and that's part of the luxury software of 8 sleep. You mentioned before you are, you know, having founder insomnia, pretty standard. Because, you know, when you go into founder mode, insomnia could come after founder mode. I did see your founder mode on the last fall and everything. I'm not endorsing founder mode except in rare instances.

But, you know, if you can't sleep, and then this wonderful thing that ate sleep at it. It's silly but profound. They ask you to put your phone number in. I was like, why did they name my phone number for my smart bed? Anyway, they send you an update, you know, about 11 or 12 o'clock with just a summary. Hey, reminding you, you got a 78, you got enough deep sleep, but you didn't get enough REM sleep. Just be a little cautious.

You might be a little groggy today. Or, hey, you got a 50. You didn't sleep well last night. you're going to need to really focus on sleep tonight. And just that little feature feels to me like so elevated. It's like if you've stayed at a great hotel, I'm staying at one of my favorite hotels in Los Angeles, shutters by the beach, which I've been saying at since 1992, I think,

93 was the first time I stayed here. If you were to call the front desk and ask like, hey, can you bring me up X? After it's brought up, they're going to call your room. Mr. Kalaganis, they just wanted to check. Did you get your ice? Completely unnecessary. They know the ice arrived in all likelihood.

Or they're pretty sure it arrived. because they have good people who work here. A little touch. Cost them very little. It's just a confirmation. And that's where I think luxury begins where necessity ends is the famous expression. Schmach told me that expression that I think the head of LVMH said. So when you think about what's necessary for recall to get the job done, now think about what's unnecessary but very luxurious and feels good.

Storing everything. Like being able to download a YouTube video or, you know, save a clip of a video and record your screen, that's super interesting. So now, think about for me as a luxury customer, if you had a button that said, record my screen when I'm watching videos or record my browser when I'm watching videos, you could literally, if I watched just seven minutes of a specific video, you could screen record my browser. Now, you're not stealing that content, right? It's not stealing of the content. You just record it in my window for me because I set something in my browser to record my screen.

Now that might be the Huberman part where he talks about sleep. Now I've got that seven minutes I listen to, save as a video file. Now, that costs money to store a video file, but it doesn't have to be like the highest res version. That's an example of like unnecessary luxury. And then speed is the other one, you know,

unnecessarily fast. Any other new features? My God, I love the quizzes. I love multiplayer. Yeah, this is going to be great. What do you call the page where it highlights on a new page or previous stuff, what is that?

The augmented browser. Augmented browser. Yeah, we're augmenting your browsing experience. Very slight augmentation to your browsing experience. Augmented brain browser, yeah. I like it. Ooh, ooh, I like that. ABB.

ABB. Yeah, ABB. I was just thinking of the ABB. All right. We all know that building software is hard. And so is getting people used to your code. We're training new users to use your software intelligently, Sometimes it feels insurmountable, right?

And I have to tell you, most tools built to help users get up to speed with new software are annoying as heck. Or like these low-rent chatbots that users despise or even worse, they ignore them. There's one company using generative AI to help users onboard without frustrating them. And that is the guide on your side from Command AI. It has a chatbot that gives personalized responses to use your questions. instead of like a basic Q&A and the thing seems dumb and it doesn't understand you and what you're trying to do. Nope. Command AI is going to show users around your product like a live guide with the cursor included.

Command AI can detect when a user needs a little nudge, right? And we'll help them down their user journey, giving them like a little product hint there or a special offer maybe to close a sale. It's used by all the unicorns that you know and love, gusto, HashyCorp, Angelus, Sixthens, and so many others, and that's why they're killing it with enhanced user experience. I want you to understand that you can take your user experience up another level. So, integrate an AI power guide into your software so that your customers can navigate your product intuitively and quickly. Visit command.a.i slash twist to get a live demo and start transforming how you and your customers interact with software. Go to command.comat.a.i slash twist.

really powerful. And then I would really think about LinkedIn as a very specific, interesting data type. Because like podcasts, YouTube videos, tweets, and LinkedIn are interesting data types that, you know, important people use. If I knew I had been to that person's profile before and you're storing the URL of every LinkedIn profile I've ever visited. And when I visited again, it says last time you were on this page was August two years ago. I love that. Yeah. But you don't have to take notes because we're, we're, we're, recording. Oh, yeah, but we're,

for people who don't know, I let you, I was like, let me write that. For people don't know, this is what we do in the accelerator. All these great companies come, and then we will sit around

with 10 founders and just jam, which came from Ticken. Travis came up with this concept of jam session, and we just jam out about different ideas. So here we are, jamming out. Any other new features we should share before we were up? I wanted to just show you the last thing. So,

you know, Jason, I've been talking about this knowledge graph, right? What is a knowledge graph? It's just a type of database. It's behind the scenes, connecting everything. But what we also launched

was just the visual representation of it. When you look at your entire knowledge graph, you know, it looks a little bit nuts. I've been using recall for ages. So this is just a full view of my entire graph. It's nuts, you know? So beautiful. I can hone in.

I can just hard what the audience is seeing for those listening. It's basically a node graph of all of the content that you're consuming connected into this web of content. It's really magical. A constellation of stars in a way. A constellation of, absolutely.

Or it could be like neurons in your brain. Neurons in your brain. And this, for example, you know, it looks nutty when it's everything, but I honed in on, I actually imported a bunch of topics on economics. I knew you were studying that, Jason, and I was doing an example for a student. I thought this would be a good one. And you can say zoom into something in the center, like what's connecting everything? Let's say in this example, it's demand curve.

And I'm like, what is that again? I can hit expand and it would then enrich what demand curve is. So it pulls up now a bit more information on what demand curve is. And the last is if there's things in here that these links aren't, you know, aren't things that you actually wanted, you can edit them, you can delete them, you can create your own. I could say that this was recommended by a particular person. I can create my own card.

So everything is fully editable. And it's basically a combination of ingested content, but also user-led. So it's this perfect hybrid. this is what I would really like because I as I had mentioned I went through MIT's course and you have the course up here that I took and you know I remember the section on demand curves and I had I understand what a demand curve is hey as you change the price it could impact you know consumption so if you make something really cheap people might drink more of it and then if you raise the price they might drink less of it so you take a cup of coffee you know if it's free in your office you might have five or six cups of coffee a day. If it's, you know, $7 at Starbucks and you have to, you know, take 15 minutes to get there, that's also a cost. You know, you might do it once a day, maybe twice. So, you know, but if it was $15, you might do it once every three days. And if it was $2, you might do it three times a day, right? So you kind of figure that out. But this is where

I'd like to take the quiz. This is where I'd love to start taking quizzes. Should we do it? Oh, here we go. Okay. What is the definition of demand elasticity? The percentage change and quantity demanded for a percentage change in price. Okay. The change in quantity demanded for a change in income, no. Price at which quantity demanded equals quantity supply. I think the percentage change in quantity demanded for a percentage change in price is demand at elasticity. I think that's, yeah.

Yeah, I think so. There we got that right. Yeah. So we don't have to go through all of them if it's too much. You're enjoying it? We'll do one more. We'll do one more. What happened to the demand for a normal,

good when the price decreases. Quantity demanded decreases. Quantity demanded stays the same. Quantity demand increases. Well, it's a normal good. It should make, the quantity demanded should increase. I would think.

Increase? But it might be stay the same. It might be, because it's a normal good. I think normal good. It's just normal. I don't. Yeah, I think maybe let's go, it stays the same. Because normal good would be, yeah, let's say.

Oh my God. No, it was. quantity demand increases. Yeah, it increases. I think that's the, yeah, you did. But that's the paradox of it. I just got thrown by normal good because I don't remember the definition of normal good, but a normal good might be like a banana.

It is a normal good today. If the price, but bananas used to be a luxury item, we were very expensive when they came from South America. They would be very perishable. It was hard to get them here, et cetera. Now, they aren't as perishable as you might think because, you know, get the green ones. They do last a little while.

have a case on them, but even still, they used to be a luxury item. Now they're a commodity. We make banana bread out of them. Come on. So that's like a perfect example. If it was a luxury item, you wouldn't make banana bread out of it. You would covet it so much. And I guess meat would be in this category as well. If you had very little access to state, when you did get a steak, you would covet it,

you would grill it, and you would eat it. You wouldn't make it into dog food. You wouldn't make it into meat bowls or hamburger. You would want to, you know, really covet it. So very interesting. Yeah. This to me is just extraordinary. Okay.

And so the knowledge graph lets you kind of dance around and see the different connections it makes. So you might have the demand curve come up in a breaking bad episode that you watch on Netflix and the demand curve for meth. Not founder mode, stay away from that. Very bad. That's not founder mode.

That's berserker mode. But it's kind of interesting when you think about it, right? Yeah. I mean, in that example, it would just be reminding you of the concept of demand curves. You're like, what's that again? And it's just this like subtle sort of active recall of things that you've saved in the past. Demand curve really does apply to actually drugs.

If you look at the horrific in San Francisco use of fentanyl, fentanyl became so popular because of its addictive properties and its price. And the price is so cheap that I think you can get high for five or ten bucks. and if it was heroin, it might be like five times that amount and if it was some other drug, I guess it might be even more. And so what the police say is,

if you do enforcement, even though it doesn't seem like it's working, when you do police enforcement, the price of the drugs go up. Do you know why? I'm guessing it's a little bit riskier to take it. It's a very practical thing. If a drug dealer gets arrested,

the tradition, is that they get a stipend from the cartel to keep paying their family, even though the person's in jail. So you get paid for being in jail. Oh, I didn't even know that. I didn't even know that. So the DEA and the DOJ, you know, they came into California and they said, hey, listen, if we want to solve this problem, you got to arrest the drug deals.

I'm like, oh, well, two more will pop up. It's like, yes, two more will pop up, but the price of fentanyl will go 3x because you now have two drug dealers in jail, one on the street selling. You have to pay for all three of those. cover the cost of everyone who's not making it. And you have to cover the cost of lawyers and you have to cover the cost of it being turned off for a day or two

while you find another drug dealer and recruiting a drug dealer takes time, money, effort. So that's like very interesting when you start thinking about demand curves. This is amazing. I'm so excited for multiplayer mode and for the augmented brain browser. I think it's just so great what you're doing

and, you know, I'm a paid subscriber as you know and I love using it. So continued success. And if people want to go use it, they can just go to getrecall.ai. Get recall.ai. If somebody has recall.com, call me. We'll negotiate it. Maybe you can get a little bit of equity in this great company, whoever's got recall.

com. Just so great to be in business with you, excited that you let us come on the journey with you and invest. And if you would like to come to our accelerator, founder university, the pre-accelerator, or get into direct investment from our firm, launch. dot post slash apply, launch. dot CO slash apply. No M there.

It's not a dot com, it's a dot co. slash apply. It's a common application for all three ways we invest. Pre-accelerator found a university, launch accelerator, which you went to, which is $125K, and then direct investing,

all at the same exact application URL. And we'll see you next time on this week in startups. Bye, bye.