Chamath Palihapitiya: you're always trying to upsell me on AI. Stop with this upselling. Just put the best out there

SPEAKER_01: for everybody. And I'm like, oh my God, I got to pay another $2,000 a year, Grammarly or Coda or Notion or whoever wants this upsell on AI. Just include it.

SPEAKER_03: Yeah, sure. Can you fix the entire SaaS industry for Jason, please?

SPEAKER_06: Please. Just go ahead and talk to everybody. It's enough with the upselling on the AI.

SPEAKER_07: I'm going to start with your two favorite products and then I'll work from there.

SPEAKER_01: Are you going to do that? Will you promise me right now we'll stop with the two different... Well, just to be clear, Coda, Coda. ...the non-AI ghetto and then the AI elites?

SPEAKER_13: This Week in Startups is brought to you by Coda. Coda empowers your startup by bringing words, tables, and teams together. Strategize, plan, and track goals effectively with all your valuable data in one place. Go to coda.io slash twist to get started for free and get six free months of the team plan. Lemon.io, hire pre-vetted remote developers and get 15% off your first four weeks of developer time at lemon.io slash twist. And Gusto. Gusto is easy online payroll benefits and HR built for modern small businesses. Get three free months when you run your first payroll at gusto.com slash twist.

SPEAKER_16: All right, everybody. Welcome back to This Week in Startups. I am your host, Jason Calacanis. With me, my co-host, Alex Wilhelm. How are you, Alex?

SPEAKER_18: Alex Wilhelm. I'm actually really good. I'm starting to get into the festive spirit, Jason. Although I will say in the Northeast, no snow yet. And I'm kind of jealous of you because if I try to snowboard right now, all I'll do is just scratch up my board and be sad.

SPEAKER_16: Ah, well, you got to get out here to Tahoe. I am in Tahoe. Yes, that is a real background. Yes. You're going to see some weird lighting here. Like there's my lighting right in front of me, but it is what it is. Uh, you know, I like to give people the background. You'll be able to see skiers go by here, um, from lovely Lake Tahoe. And I did get out for my first day of skiing yesterday. How was it? It was great. I mean, I, I did too much. I, I came in and I was like, okay, I did my, you know, nice, you know, whatever, six, seven, eight, I did my six or seven, eight runs. And then I saw my other pair of skis. I have two pairs of skis, one for powder, one that's more for groomers, which means when they make the trails perfectly smooth. So one of them is parabolic. They're very thin and then they widen and then the other one's fat. Um, so you float on the snow. And I just took the fatties out for a little bit, because there was a little bit of powder, because we had a little dumping this weekend and, uh, it was great. Gonna try to break my record of 40 days in one year by doing 41 days this year as a 54 year old that I'm old. I'm just trying to get those last, that last decade of skiing. And hopefully all the science and technology we talk about will result in us skiing into our seventies or eighties, but I'm not taking it for granted. I want to get my turns in now. And I do what I call the CEO program, the VC program.

SPEAKER_01: I work in the mornings. We do the pod. I go out and instead of having like a, going out to a business lunch, I do two or three hours in the mountain. Then I come back and I start working again. Success has its privileges. I get to set my own schedule. One of the great things you can do. And what I like to do is I do that little, you know, two or three hours in the middle of the day and get on the slopes. And what I do is I take notes like crazy. When I'm on the lift, I get so many great ideas. All my best ideas happen when I'm on the mountain. And so, yeah, I'm really excited.

SPEAKER_18: As you can see the working class version of, of the ski lift though, is, um, I love to take like a 15 minute shower at some point in the middle of the day, because if I stop staring at my screens, stop reading new notifications, reading new notes, reading new tweets. And I go and I just stare at a wall for 10 minutes, my brain goes, and I figured out titles, ideas, how to lay out an interview. It's an amazing, like cleanse for your brain. So, I mean, three hours must be amazing.

SPEAKER_19: I'm if I can get it. Sometimes it's just an hour and 90 minutes. Sometimes it winds up being three

Chamath Palihapitiya: or four. I saw there was an extraordinary private fundraise, not for SpaceX, not for Stripe. Those are the other two giant private companies, but Databricks just raised $10 billion in the private

SPEAKER_38: markets. I mean, that's like a masa IPO. That's a large amount of money.

SPEAKER_18: It's slightly different than that. And that that's the same headline that I saw. That's exactly what I put in the notes. And then I went and I actually read the Databricks announcement. Here's the thing, Justin, the language gets a little weird. So Databricks series J, they are going to raise 10 billion. J Cal, here we go. J as in J curve. I mean, you know, they've completed $8.6 billion of this to date. So they've gotten 86% of the $10 billion. They will get the other 1.4. I'll explain why in a second. What tripped me out though, was that they wrote it as non-dilutive financing, which to you and I means secondary. But when I was reading other reports and coverage of this, and I couldn't get all of Ali in time to see you have Databricks is I'm just trying to figure out what they're doing here. It's something.

SPEAKER_26: No, they said it was non-dilutive. When you raise a round of capital, you're selling shares in

SPEAKER_01: Andreessen Horowitz, DST, Yuri Milner's amazing firm, Insight Partners. These are investors who want equity, but maybe there's some kind of weird conversion that happens when the IPO happens. This is like a

SPEAKER_48: Mez round of financing, I guess. There used to be this thing called a mezzanine round. The mezzanine round existed when there was just venture capital and IPOs and a Mez round was a year before the IPO or six months before the IPO. People put a slug of cash in and they get first shot. It's almost like preempting the friends and family round or getting an allocation, or maybe you get a

SPEAKER_01: little extra bonus, a coupon, as they call it in the business, an extra couple of percentage points, an extra couple of shares, because you took the bet a year before it went public or six months before it went public. But I guess we don't know here the exact details, or we're figuring them out.

SPEAKER_18: We're going to chase them down, but I want to make a point about the mezzanine rounds, because that's a really fun point. One, you don't hear about those anymore because there's infinite private capital, so you don't need something to bridge you from private to public. But Jason, weren't Mez rounds like 50 or 100 for companies that we're about to list? It's funny how that is now, what, a series C? It's been completely consumed by venture.

Chamath Palihapitiya: Two things have happened. One, venture capitalists have become more ambitious because they know the company so well, because they're on the board of it. Because they know the founder since before the product even launched, they've learned over the last decade or two, well, maybe we should stay involved with these companies. And the key company was Google in this respect. Google made more money

SPEAKER_36: for investors, dollar for dollar, after they went public, then maybe everybody, but maybe the series

Chamath Palihapitiya: A. In other words, why would you get off the board of a company, as VCs tend to do when they go public,

SPEAKER_60: or sell your shares in it? And so Sequoia started staying with these companies longer, they started staying on the boards, and then they started this heritage, or I'm sorry, the Sequoia fund,

SPEAKER_01: where they manage the public equities. And then for the founders, well, wouldn't you like to have Michael Moritz still involved? Wouldn't you like to have Doug Leone still involved, or Ruloff, or Alfred Lynn, or these other great venture capitalists, and keep that continuity going? That's kind of cool. It would be very cool for me if I had had enough equity in Uber to be a board member right now. There would be nothing more glorious than DK calling me up one day and saying, hey, would you be on the board of Uber? And I'd be like, of course I'll be on the board of Uber, because I remember having the conversations with Garrett and TK when they started the company. I remember the discussions around tipping. I remember the discussion. It's just kind of cool, right? To have the historical legacy people around. And so, yeah, that was one thing that happened is

SPEAKER_60: people kind of realized, hey, there's a big opportunity there. Why not take advantage of that?

Chamath Palihapitiya: And then the opportunity set became bigger and entrepreneurs became more ambitious. So as people become more ambitious and the products become bigger and the customer base became bigger over the last 20 years, we went from tens of millions of people having broadband to two or 3 billion people having broadband in their pocket. This is a totally different market size now. There's nobody who's like, I'm not putting my credit card in the internet. There's nobody who doesn't have a mobile phone. Like if you don't have a mobile phone in 2024, it's because you're a Luddite who specifically doesn't want to participate in technology. Like even like Frontier, previously known as third world countries. Those are the places people are going with mobile phones because it's such a huge opportunity.

SPEAKER_01: And they sell a $15 fricking Android phone in India now, $15. It's crazy. So anyway, putting it all aside, that's why we're seeing funding the way it's going, staying private longer, all that stuff.

SPEAKER_71: That's the backdrop, but you brought up, you brought up Google's IPO, which means that I get

SPEAKER_18: to bring up my favorite piece of data of all time, which is let's all look at Google's S1 filing. Oh, here we go. And, um, I just love to pull this up every once in a while to remind myself what a company that becomes Google looks like when it's young and the company's revenue growth went from $220,000 in 1999 to $19 million in 2000 to $86 million in 2001 to $348 million in 2002, and then Jason, $962 million in 2003. That is just one of the most crazy to double or triple revenue.

Chamath Palihapitiya: Growth companies tend to grow. What a growth stock is defined as 15% year over year. Sure. 25%, something in that range. And so, you know, it's really hard to maintain that. And this speaks to the opportunity. This speaks to how much time people are spending with the supercomputer in

SPEAKER_80: their pocket, you know, with infinite storage, infinite bandwidth, uh, and infinite, uh, network effects.

SPEAKER_82: Do you spend too much time tabbing between your team chats, documents, spreadsheets, databases? Well,

SPEAKER_84: it's time to consolidate all the knowledge that is spread out across your entire organization on one platform. And that platform that we use every day here at This Week in Startups and Launch is called Coda, C-O-D-A. If you don't know about Coda, it is a new category of software and it's kind of like a collaborative workspace, document, spreadsheet, apps, all put into one. It's really easy to learn how to use it and it's incredibly powerful. We use it for all kinds of applications. One of the applications we've been using it for is something internally we call the whisper network. What the whisper network is, is we like to whisper when we invest in a company and they're getting some traction, we'll whisper it to other investors just so they know about the company. Well, we track that all in a database. And then we opened it up to our 450 portfolio companies who can request an introduction to another firm. And it's all tracked through Coda, including what they want to include in that email. And then who on our 11 person investment team is the designated contact for that other seed fund or giant venture fund. And this has changed everything for our organization. We

SPEAKER_89: make these mini apps that allow us to be so successful at hitting our goals. And I'm a big fan of having systems as opposed to goals. All the goals we want to set will be done really well if you have a good system. And Coda builds these systems of record for us really quickly. So here's your call to action. Coda empowers your startup to strategize, plan, and track goals effectively and build the systems that you need to hit those goals. So take advantage of a limited time offer just for startups. Go to coda.io slash twist and get six months of the team plan for free. That's coda.io slash twist to get started for free and get six months of the team plan coda.io slash twist.

Chamath Palihapitiya: We take for granted where we don't recognize exactly what's happened over 25 years. Since 2000,

SPEAKER_36: we have seen everybody in the world get a supercomputer. Everybody in the world have unlimited free

SPEAKER_01: bandwidth essentially. And everybody in the world have unlimited free compute, unlimited free storage.

SPEAKER_97: For consumers, that's essentially the case. And that just takes a lot to grok. Now it's been reset

SPEAKER_01: with computational power on GPUs and you do hit max limits as we've heard from various guests on the program or over and over again, the cost of using chat GPT or whatever. But even that is going to follow the same trend line. Right now, consumers don't often worry about storage for photos. If you did worry about storage for photos, you can literally create a second Gmail account and get another free, whatever their limit is. And you'll have 20 accounts with 20 different sets of photos. Why don't people do that? Because it's so it's cheaper to just pay five bucks a month.

SPEAKER_102: I'm paying two bucks a month for like an extra 100 gigs. It's amazing.

SPEAKER_48: It's nothing. It's nothing. I pay for my family plan. I think we're on the four terabyte plan, like just to give you an idea, but I don't mind. I got four daughters. I want all these pictures. I want them taking videos, go crazy, $300 a year. I don't care. Apple earns it. And I would have bought a $1,000 rate array put in my house. I'd rather just have it in the cloud. Easier, breezier.

SPEAKER_18: I remember those days of buying more hard. I mean, do you remember when people were talking about how big their SD cards were like, oh my God, that was four gigabytes. You can have like an hour of video.

Chamath Palihapitiya: We're so spoiled in the cloud. And you know, I just want to give people even more hope now. Okay. Imagine a world in which energy, water, and food follow the same trajectory.

SPEAKER_36: Mm-hmm . That's what we're on the cusp of. We will have free energy for everybody around the world, and it will be a similar de minimis upgrade to have the next tier. But there'll be enough free energy everywhere with nuclear, if we ever hit, you know, other sources, but between solar, batteries, nuclear, it's going to be bonkers, wind. And then that means unlimited free water, because water is a function of energy. And then it means unlimited free produce, because produce is largely going to be a matter of energy to power robots, to pick things in the field.

SPEAKER_01: And that's going to be pretty crazy when you can buy a $20,000 robot that tends to your fields at night, because that's night vision. And it's doing overnight, it's doing Alex and Jason's farm and planting stuff and bringing fresh fruit to the table and fresh vegetables and getting the chicken eggs from the coop and making you dinner, making your breakfast. That's literally happening in our

SPEAKER_117: lifetime, folks. You will have a, not a hybrid or electric car, you will have a robot.

SPEAKER_119: I can't wait. And that's actually why I'm still optimistic. But on the point of things getting so

SPEAKER_18: cheap, they become free. I actually wrote something back in 2014. Wow. And my argument was that as we watched storage becomes so cheap, that eventually everyone was going to have basically unlimited free storage. I took it one step further. I said, cloud computing is going to become eventually cheap, not forecasting the AI boom and the cost of GPUs, which has ruined my point here. Yeah, your point is correct.

Chamath Palihapitiya: You are correct in your thing that compute is essentially free. It's just not AI computer is free, which is a different category. So Databricks has how much revenue? Is it two or

SPEAKER_126: three billion is the last I remember something in that between those two numbers?

SPEAKER_18: Yes. So to your point about there's so much data, who manages that data? Well, Databricks does. So it's growing quite quickly. Over 60% growth in Q3, which is the October 31 quarter, if you care, will cross 60% year over year growth, more than 60% year over year. And it's going to cross $3 billion in ARR and reach free cash flow positivity inside its current quarter, according to its own internal

Chamath Palihapitiya: projections. So that means it's 20 times price to sales ratio, three times 20 equals 60 billion. So that's pretty insane for a company of that scale to get 20 times top line revenue. What would a company like, I guess their biggest contemporary would be Snowflake. Yep. Um, what is Snowflake worth on an enterprise value? I wonder today. And I wonder what Snowflake's

SPEAKER_126: revenue is right now. General Slootman is no longer at the helm. He's moved on, but he really drove that company to huge success.

SPEAKER_18: Okay. So, Snowflake, uh, to answer Jason's question, according to Bessemer data, they are currently, uh, their enterprise value divided by annualized revenue was 14.3 X their enterprise value divided by forward revenue is 11.7 and they have 28% growth, 67% gross margins. And their last 12 months free cashflow margin was 22%.

SPEAKER_139: What's their, what's their just overall revenue. So it seems like Databricks is trading at a

Chamath Palihapitiya: 50% higher premium, I guess, because it's got twice the growth, right? So this is where eventually your market cap, the value of your company, the, the growth rate factors in.

SPEAKER_60: So when I see somebody who's got a company doing a million dollars a year and they want, I don't know, 20 times at a $20 million valuation, or maybe they want even more than that a hundred times, they want a hundred million dollar valuation for a million. I say, okay, well, what was the number last year? We were revenue like 800,000. I'm like, so 800,000, a million. So you grew 25%. If you grow 25%, again, you'll be at 1.25. We're not closing that valuation gap. So what people looking just so early stage entrepreneurs know, well, they'll get eventually your growth rate is going to matter.

SPEAKER_142: This is why sales solves everything. As market would say, sales will drive the valuation.

SPEAKER_01: If your growth isn't there, your valuation does not get a premium. It gets a discount. If you're only going from 1 million a year to 2 million, if you're going from 1 million to 2 million, okay. You're high growth. Great. But if you're going from 1 million to 1.5, 1.25, these are small numbers. It's not high growth anymore to the venture community, because they're trying to figure out how do you get to a hundred million in revenue,

SPEAKER_148: a billion in revenue. They don't see the path there.

SPEAKER_18: No answer your question though, just to add this in a snowflake had 900 million in product revenue in the last quarter, about 940 million. If you include some consulting stuff, that doesn't really matter as much, but that puts it roughly at, call it 4 billion a year, Jason. So it's about one third larger than Databricks, but growing less than a half as quickly in percentage terms, which I think implies that Databricks is going to outstrip it what next year

SPEAKER_147: in terms of total revenue scale. That's an insane number because snowflake has been one of the

SPEAKER_119: fastest growing, best performing, most beloved public companies since it listed. I mean, the Slootman era is known because it was successful. So wait, what was the total, what's the total revenue

SPEAKER_18: right now? What's their last quarter or? 942.1, which is 900 million in product revenue. And then there's some other stuff mixed in there that they tend to discount because it's

SPEAKER_60: lower margin. Okay. So they have about a billion in revenue a quarter. So 4 billion a year. So they're a third bigger and their total market cap is versus Databricks. Databricks is 60 billion at

SPEAKER_153: this 62 and they are worth 55.7. Okay. So this is really interesting just to do back of the envelope

SPEAKER_60: math for folks. Sure. The companies are being valued at essentially the same. One of them has a third more revenue, right? 33% more revenue, 4 million versus 3 billion. So why would they be valued, snowflake a little bit less, 55 billion versus 60 billion? Why would they be valued less if they have more revenue? Well, then you look at the growth rate, one's got a 30% growth rate, one's got a 60% growth rate. That's the reason because people are looking at it. If you're a financier and you work in capital markets, you're going to look at it and say, well, when do they catch up? When do they pass? Databricks, if they're growing at 60%, 60% would be adding 1.8 billion, or if let's say they were 50%, they'll be at 4.5 billion. If they're growing at 30, they'll be at 4.2 billion. So they're about to get passed by Databricks will pass snowflake at this current speed in the next

SPEAKER_01: year, they'll lap them. And so if you were looking at a marathon runner and somebody's got a mile lead and it's mile 20, you know, maybe it's too late to catch up, but if it's mile 10, Databricks could

SPEAKER_18: be going whipping past a snowflake. And also, I just want to say, and this is a vibes point, which doesn't show up as much in the multiples and the revenue numbers and kind of the hard facts of this. But when you think about AI today and AI in the next five years, and I offered you one share of

SPEAKER_147: Snowflake and one share of Databricks, which do you think is better set up right now to capitalize on the AI boom and therefore have more tailwinds? Yeah. And so this is where looking at product

Chamath Palihapitiya: roadmap and talking to customers is the job of a capital allocator. And so when we do our diligence, even at the early stages, seed stage, kind of crazy, we do like to talk to customers because we

SPEAKER_60: want to have the conversation with the founder. Do you have the right customers? And I constantly have do-gooder entrepreneurs, God bless their hearts. They come to me and they pitch me and they say,

SPEAKER_48: Jake, how respect you, been watching this weekend startups since I was in high school. You always give it to me candid. Tell me what you think of my idea. And I'm like, I think your idea is great. I think

SPEAKER_01: your selection of your beachhead customers is idiotic and terrible. And we're going to have every VC

SPEAKER_170: tell you how awesome you are and ghost you. All right, founders, are you tired of doing all your own software development? Do you need help, but you can't afford all this time it takes to find great talent? Are you dreading the endless interviews and email chains just to find somebody great? It takes six months. It takes a year. Well, what you need is lemon.io. Lemon.io has thousands. That's right, thousands of on-demand developers who can help you. And they've done the work already to vet these developers, making sure that they're results oriented and that they're super experienced. Of course, they've got to have competitive rates. So they're going to take care of that as well. And great developers are so hard to find and integrate into your team, unless you're using lemon.io, because they handle all that for you. They only offer handpicked developers with three years of experience at a minimum, and they have to be in the top 1% of applicants, right? If something goes wrong, don't sweat it. Lemon.io will find you a replacement developer ASAP. So many of our launch founders have worked with lemon.io and they've had great experiences. So here's your call to action. Go to lemon.io slash twist and find you're a perfect developer or the perfect tech team in 48 hours or less. That's right. And twist listeners get 15% off the first

SPEAKER_173: four weeks. Stop burning money. Hire developers smarter and faster at lemon.io slash twist.

SPEAKER_01: If they think you're going after nonprofits, that's like going after newspapers in 2010,

SPEAKER_48: when they're contracting nonprofits are like run, you know, not super professionally, not super efficiently. And it's just like, why would you pick that market? Cause you want virtue signaling points. Cause you want to feel good. It's not the market to go after. If you were going to build a finance company to help people raise money for their nonprofit, why wouldn't you build a financial tool to let anybody collect payments like Stripe? Yes. So you've taken the Stripe

SPEAKER_01: opportunity, the Apple pay opportunity, whatever, and you've narrowed it down to nonprofits.

Chamath Palihapitiya: To make it worse. Yes. Yeah. No. So why narrow it down? And if you are going to pick a beachhead customer, which you should, and you should have an ideal customer. Why wouldn't you pick the one with the most money who is the least discerning, you know, uh, in terms of like paying

SPEAKER_179: for it? You know, in other words, you start, and I give them the example, Travis and Garrett and the

SPEAKER_36: team over there, they, Ryan Graves doesn't get a lot of credit. They deserve, they, that he deserves,

David Friedberg: he deserves a lot more. Uh, they get, in fact, he was going to be CEO for a big unknown story. Um,

SPEAKER_182: it almost made me pull my investment. Was that back? Was that, I remember I went out drinking with one of the Uber founders who was not TK in Chicago when they were just having their first

SPEAKER_18: couple of black cars in the city before it was actually live. I think that was Ryan, but that was, I think the event that we eventually labored later called Chicago tech blackout. So it's details are a little spotty. We used to have fun in technology. Oh yeah. Uh, maybe too much at times.

SPEAKER_179: Yeah. And so the, the interesting thing was Ryan was going to be CEO and Travis was like, you know what, I'm not going to take the CEO slide. And I was like, what I'm investing in you. You have to take

SPEAKER_117: the CEO side of getting rid of the ride act. And I think Ryan may have taken a person that he would

Chamath Palihapitiya: be a great CEO, but I think, you know, Travis is obviously, I don't, I don't think anybody thinks that there are very, there are very few people who count them on one hand would be better CEO than Travis. Um, so that's not a dig to Ryan, but Ryan, I think maybe took it personal and pulled my

SPEAKER_117: investment and I got really upset. TK was going to have him be the CEO and then TK to his credit. And it's like, okay, fine, I'll do it. Not because of me. It was his own decision,

Chamath Palihapitiya: obviously, uh, putting all that aside. Um, you know, you have to pick your, who you pick as your first customer is clearly important. They picked Lincoln town car drivers, uh, and people who took Lincoln town cars at the airport. Why? Cause they were already spending back in the day, it was three or $400 to have somebody wait for you at the airport at JFK. You paid for an hour, typically a hundred bucks an hour, um, for them to arrive early, an hour, a buck 50 for your trip,

SPEAKER_01: maybe. So it was like two 50 or something. Now they would come, they would hold the sign. They would get your bags. It was all quite charming, but you would put that versus 40 bucks, for 50 bucks for a yellow cab. And people made a decision, $50 or two 50. If you're trying to keep it with the yellow cabs and those customers, they complain, they ask for a refund, they're price sensitive. And if you told the CEO of a company, like they'll meet you at the curb, you can see them on the map. It's 150 bucks, it's a hundred dollars less, or it's a hundred bucks instead of 200. They're like, yeah, okay. It's cheaper, but it's more convenient and it's easier. It's in my pocket. I don't have to call somebody on the phone. There was like, you know, better

SPEAKER_48: reasons. So always the best possible customer you can. Don't be a do-gooder, be a rabid capitalist. If you want to do nonprofit stuff, then go do nonprofit stuff. My best advice to people, go make a ton of money and then create a nonprofit. If what you're passionate about, if you want to save the mountains and trees, go make a bunch of money in enterprise software.

SPEAKER_160: And you'll buy a lot more trees than begging people for a hundred bucks to plant a tree.

SPEAKER_119: Or back to the payment example, build the broad payments company and then create an arm inside the company that offers cheap payments for nonprofits. There you go. Oh my God. There you go.

SPEAKER_179: Yes. You could do your do-gooders do the do-gooder stuff after you're sustainable and strong and growing.

SPEAKER_59: Yes. Just don't do it in reverse. And they're wondering, they're wondering, Jake, why can't I get funding? It's like, you literally signaled to venture capitalists who are

SPEAKER_01: on edge, they're on tilt about getting returns, that you don't care about returns. Now, you may have done that, or, you know, you may not have done that on purpose, but you just have to understand that's what you did. You just said to them, I'm going to burn your money and not get your return, as opposed to, I'm going to print you money and get you a huge return. So just know what you're

SPEAKER_18: signaling. I want to do a quick point about the ghosting point, because you mentioned how VCs might say you're great and then not talk to anymore. How common is that? Well, there was a great thread over on the startups forum on Reddit that I saw the other day. I've gone through and I've highlighted every use of the word ghosting from this founder who pitched a great number of VCs. So there's ghosting,

SPEAKER_198: and then there's ghosting, and there's ghosting, and then there's ghosting, and then there's

SPEAKER_201: ghosting. Yeah. And then there's even more down here. Oh, you blocked it. Oh, no, you're on,

SPEAKER_202: you're just on the, wait, how did you get all the ghosting? Oh, I just did command F ghost. Oh,

SPEAKER_205: got it. Perfect. But like, look how many people, the ghosting was pretty pandemic for this, this

SPEAKER_207: founder. I mean, it's the reason for this is, is it's a hard conversation to tell somebody,

SPEAKER_36: I don't want to date you because you're ugly. And I'll just tell it to you straight. That's the beauty contest here. If you think that men pick women because of beauty, let's just say you pick that, you think that that's how men pick mates. Okay. We can debate it if you want or not. In that dating scenario, are they going to say to you, I don't want to date you because you're ugly,

SPEAKER_01: or I don't think you're attractive. Or are they going to just say, I'll just respond and I'll move on because that's, which is more comfortable to do. That's what VCs do. We take the approach. We say to founders, we have other investments that we have to prioritize that, and we just leave it at that, but we would love to get an update from you. If they ask us, we'll give them more details, but I just brought it up internally for our firm launch. Again, should we be giving founders more details of why we're not investing or not? And I have switched my philosophy back and forth, back and forth, because once you give founders, Hey, we're not investing because of the margins and because of the growth rate and because of the team, they say, well, you're wrong. And they argue with you. And then you have another email to do. That's why the best ones say we have other investments we need to prioritize, or this isn't a fit for me. I do want to do a quick news

SPEAKER_147: update though on TikTok, just so everyone knows we've been talking a little bit about TikTok and its

SPEAKER_40: legal travails. So today- Yes. The ban is in January 19th, right? Isn't it January 19th? This is fast coming. We're about a month away. We're 30 days away from this. Absolutely. And TikTok's done.

SPEAKER_18: Yeah. Well, so on December 13th, the DC Court of Appeals said no to them asking for an injunction against the law. Then on the 16th, TikTok filed, dear God, Supreme Court, please come help us. We don't have any time left. Today, the Supreme Court said no to their request for an emergency injunction. But what they did was they've squeezed an entire Supreme Court petition in between now and January 19th.

SPEAKER_219: So it's going to be, yes. Wow. The Supreme Court doesn't usually do that,

SPEAKER_221: right? Aren't they usually slower? Slow. Yeah.

SPEAKER_147: So slow. But the commentary that I saw that made the most sense to me is that they're compressing this so that way the current DOJ and the current set of government that brought this bill and passed it are going to be the ones defending it, which makes, I think, reasonable sense. So very quickly,

SPEAKER_224: everybody- There's going to be an administration change. The new administration might have different feelings about this. The old administration should get to dispatch justice

SPEAKER_227: here as they see fit or the court should. And so they just want to wrap this up before it gets nixed

SPEAKER_228: or there's a transition of people, I guess. I guess that seems like a good idea.

SPEAKER_229: I think it's reasonable. And then the Trump administration could do whatever they want. Congress can pass a different law. They can go back to the drawing table.

SPEAKER_18: Oh, right. We live in a democracy. There's laws and there's a process for passing these laws. Yes. It turns out there is a process. We don't use it very often anymore, but we do have one that is in reserve in case we ever decide to legislate as a nation. Quickly,

SPEAKER_147: opening briefs, December 27th, right after Christmas, we'll get the big notes from the government and TikTok, and then essentially oral arguments, January 10th. So this is going to be

SPEAKER_206: bang, bang, bang. We're going for it. No Christmas for people working at the law firms of TikTok or the DOJ. You've got work to do. No skiing for that group.

SPEAKER_101: Jason, this is going to be the biggest, no, what? No. Okay. Look, if someone wants to pay me

SPEAKER_205: what they're going to pay those lawyers to work over Christmas, I think this is literally the best

SPEAKER_242: Christmas ever for the lawyers involved. Those families are like, guess what? Christmas is in February, wherever you want it to be. Yeah, exactly. Pick a continent. We'll do it there.

SPEAKER_18: We're there. That's what's going on. Everyone, we have our eyes on it. We'll bring you the briefs when they come out. It's going to be a big deal. It's going to be politically charged. Very interesting. And it'll help determine kind of where things go. But Jason.

SPEAKER_202: Divest. I retain my position. Divest, divest, divest. I'm still on that team. Do you be powerful for the Chinese to have this tool? Yes. Or any country to have this tool, I'll be honest. Especially in our nation.

SPEAKER_173: All right. You didn't start your company to run payroll, did you? Of course not. We all know that.

SPEAKER_170: Gusto is here to help. Gusto is going to help you run your payroll and handle all your benefits, onboarding and HR all in one place. The market agrees. 300,000 businesses. Trust Gusto today. And you can too. As your startup skills, Gusto is going to grow with you. You got state and federal taxes handled for your staff around the country. Gusto does all that. And hey, maybe it's finally time for you to offer a 401k plan for your team, right? Gusto's got you on that. And you might need to get your compliance sorted, right? Well, three out of four employers say Gusto helps them be government compliant. And even better, Gusto is simple, easy to use software. So you can focus on what matters, building your startup. So here's your quick call to action. Do you want all Gusto has to offer with no hidden fees? Well, how about a discount? Try Gusto and get three months free. Gusto.com slash twist. That's G-U-S-T-O dot com slash twist. All right. So coming up next, we have a guest that I'm very

SPEAKER_18: excited about because Jason, it is rare that I talk to someone who runs not one product that I use day

SPEAKER_147: in and day out, but two. And so I want to welcome Shashir Mehrotra to the show. He is the former CEO of Coda and now also the incoming CEO of Grammarly because news broke this week that Grammarly and Coda

SPEAKER_252: have decided to become one entity. They are merger acquiring one another. And so we have him here.

SPEAKER_255: Shashir, hey, how are you doing? Hey, I'm doing great. I think technically I'm still CEO of Coda.

SPEAKER_256: It closes in January, so it closes in January. Shashir, good nice to see you. As you know,

SPEAKER_36: I love Coda. Coda has changed our company. We use Coda for so many interesting projects internally that you're aware of, but that I have told you you're not allowed to tell people about. We were going to do a case study on some of the stuff. Big secret, I know.

Chamath Palihapitiya: I just, you know what? It's like my secret weapon. We do really interesting projects here.

SPEAKER_01: Now the Twist 500 is a public one, but we built this Twist 500 database here on Coda. What I love about Coda is as a workspace and a tool, it's kind of like the Wikipedia, plus a database, plus Google

SPEAKER_60: Docs, plus, plus, plus, plus Zapier. But we still use Zapier with it because it complements it quite

SPEAKER_01: nicely for a lot of the things we're doing. But man, I love the program. Twist 500 is the public facing thing we're doing, where we're putting the top 500 private companies, which we're going to include your merged companies. So congratulations, you're in the Twist 500 as a merged entity. But also Grammarly, I love because as you in a 20 person company, like we are slightly more, there's usually two or three really great writers and they become, so it's me, Alex and Jackie are

SPEAKER_48: the best writers in the company, I think. And people come to us and say, Hey, we wrote copy. Hey, I got this memo. Hey, we got an LP update. And we read it. I pay, not only for everybody to

SPEAKER_60: have Coda, but I pay for everybody to have Grammarly because Grammarly has helped everybody in the company go from being a bad, okay, or good writer to a good, great, or excellent writer. So I love both products, but tell us what happened here. Why are these two companies coming together? Because they seem like

Chamath Palihapitiya: maybe they're different products, but we know that SaaS has headwinds. You both raised a ton of money, 200, $300 million each. So why are these two companies being put together? Is it a sign of

SPEAKER_36: weakness, a sign of strength, a little bit of both, a way to consolidate and to be more efficient? What's the background? How did this all come together?

SPEAKER_269: Yeah. First off, thanks for the glowing review of both products. And I won't reveal all your secret

SPEAKER_255: Coda use cases. And someday, someday I'll convince you to let the world know that we don't have to do

SPEAKER_258: that today. Yeah. I'm super excited about this where we, as you mentioned, we announced yesterday, Coda and Grammarly are coming together. I've personally been like you, a deep personal user of Grammarly for years. I was actually an early investor in Grammarly as well. So I followed along the business, um, uh, as well. And, uh, you know, it's been a huge fan and I don't think I saw it as an obvious pairing either. Uh, but then we got asked to get together. We have a common investor, uh, who put us together and said, you know, you guys, sorry, you know, amazing person and, and, you know, great connector and so on. And he said, you know, you guys don't realize it, but I've sat through both of your discussions and you're actually headed to the same place. And it was kind of surprising, but we took a day, we sat down and we literally took each other's vision memos for the next year and just went through them line by line. We compared some of the, some of the, the, the mocks and graphics. We compared some of the, the, the observations and realized coming from two completely different places, but actually headed to the same spot. And the end of that meeting was basically, are we going to each try to go kind of build each other's approaches or are we going to do this together? And it was just obvious we should do it together. So that's the,

SPEAKER_60: what is the joint vision? If you were to describe it, because I use Grammarly as an editor when I'm writing an important piece, I will write in Grammarly, but I have it installed on everything. So if I do a tweet, it's going to, on my desktop, at least Grammarly. Grammarly keyboard, a little bit clunky, needs a little bit of work. I'll give you my notes privately on that because it's, doesn't

SPEAKER_36: do a good job of predicting the next word, putting us on my little tiny insight there. Um, and then Coda I use for building, uh, workflow products. What we used to call workflow in our industry,

SPEAKER_60: as you know, things that Lotus notes used to do where, you know, Hey, we're putting some data in, and then we want to have things happen. Little scripts happen based on, uh, and reporting happens based on what happens with that data. So what is the common vision for the two companies? If you had

SPEAKER_255: to describe it in a sentence. Yeah. So, so at the, at the core, we're building a user-centered AI

SPEAKER_258: platform of applications and agents. And, and those two things together are what makes this magical. So maybe I'll, I'll just sort of divide them up quickly. So agents first, uh, you know, that words tossed around a lot these days. I think of Grammarly as the original AI agent, the, the one that, you know, 40 million daily active users, 200 billion words a day. It's in the middle of everyone's work and follows them everywhere they are. And the, the real magic of what Grammarly has done, which, which is, as you mentioned, you experience well as it works everywhere. Doesn't force you to change your behavior. They've done integration with 500,000 applications. So the, the Grammarly experience can be like this, this agent along, alongside you as you're, as you're writing. Hmm. One, one analogy they use is like, it's like an AI superhighway being built into every other product. And the, the analogy that, that Max and Alex, the founders here at Grammarly talk about is that they, they feel like they're only driving one car on that superhighway. They have this one agent that comes along and helps you with proofreading your, your, uh, your work. Another way to think about that is that superhighway, now that we're in the world of AI and there's all of these different types of agents, I said, what else could we put on that? And so this part of the conversation mostly focused on a new product we're building in Coda called Coda Brain, which is basically we've taken all the integrations of Coda. It's one of Coda's favorite, uh, features. Everybody, everybody loves to use Coda to connect to other tools. Uh, we have about 1700 or so different integrations, 800 or so that are public and they, they connect everything. So it's email and Salesforce and Jira and so on. And you can just picture that all of those now become agents on the same superhighway.

SPEAKER_276: I like the vision. I understand it.

SPEAKER_258: Right. So this first one is really easy, right? So you're going to come along and rather than this, this agent helping you with, Hey, you got your grammar wrong. It's also going to tell you, oh, here's the context on that investment. Here's, here's a little bit of what, you know, that customer said last week in their, in their meeting. Here's, here's a little bit about the

SPEAKER_278: feature you're talking about. So on.

SPEAKER_206: It's, it's like a co the, it's a co-pilot for writing better, but if it was reading everything

SPEAKER_36: I'm doing, which it is, and it's thoughtful about it, it could say, by the way, uh, I know you're responding to this founder in your founder Slack, other conversations with this founder exist in your superhuman, in this other CRM. Oh, by the way, you're not just talking to this startup here. They're also an advertiser in there, in the HubSpot where they're in this marketing database.

SPEAKER_60: That's like a really interesting observation for sure that Grammarly is following around and kids

SPEAKER_01: love Grammarly. So the thing that, um, I was just talking to somebody and they said, uh, Grammarly just got banned in their class. Their teacher doesn't like Grammarly because it's too good. They'll let you use ChatGPT, but they said no Grammarly. And I said, that's really interesting. Why not? I'm, I've been teaching my daughters to use Grammarly, uh, as they learn to write because it makes you superhuman. It's so powerful that, yeah. Interesting.

SPEAKER_258: By the way, it's on the, I mean, I have, I have two teenage daughters of 16 and 18. So they're, they're my, they're my first test for many of these things. And, you know, of the, of the, the many wonderful text messages and everything I got yesterday, the ones from my daughter's friends were definitely the most heartwarming and from their teachers, which is also great. I love it. Yes. They do. And I actually, that the experience you described is it's usually the

SPEAKER_290: opposite, actually. Usually it's Grammarly is approved and ChatGPT is not.

SPEAKER_160: That's what I said. Yes. But there was some teacher, it was an English teacher who was like, I want you to write without Grammarly. I understand that. I would, if I was a teacher,

SPEAKER_48: I would have them write something without Grammarly. Then I would put it into Grammarly and walk through each correction and then debate each one. Then they would learn how to use these tools, not, you know,

SPEAKER_142: without relying on them, but to use them to augment their brain.

SPEAKER_269: We, we should get to the second half of the vision, but I would say, even on just this topic, this idea of an agent that can do everything with you and knows more than Grammar. What, one way to

SPEAKER_258: think about that, just to picture a world is I actually want that agent to be the teacher's agent. I want the teacher to be actually part of that. You should, one of the amazing things about Cotapax is anyone can make them. So there, a lot of them have been made by many people around the world. Many people build them inside their, their, um,

SPEAKER_296: Jim O' Explain what that is, a Cotapax.

SPEAKER_258: Jim O' Yeah. So Cotapax is an extension of Coda. So that, that, uh, most of them are integrations. So they're mostly integrate between Coda and some other product, but there are ways...

SPEAKER_87: Jim O' You can think of them as Zaps for Zapier, or if this, then that.

SPEAKER_258: Jim O' Exactly.

SPEAKER_290: Jim O' Two other great companies that, you know, do the same... Jim O' So the common ones will be two systems. They connect to Gmail, they connect to Salesforce,

SPEAKER_258: so on. But you can build ones that have a particular point of view. So you can absolutely build one that is, you know, here's the chemistry agent that knows everything about, about chemistry. And for that student use case, you can really, uh, you know, imagine that AI superhighway having a whole different class of agents, even for just that audience. Jim O' It's also, I think, interesting to think about, Grammarly obviously grew up known for students, very core audience, but it is actually the vast majority of users are actually professionals. Um, and so, yeah, that's right. I mean, you can't, you can't get to 40 million daily active users without that. And so the, there's a huge number of people out there that rely on Grammarly. It's marketers, salespeople, lots of CEOs. I mean, the number of messages I got yesterday from CEOs who said, this has saved me many, many times as I, uh, as I have, uh, you know, tried to communicate with my board, with my team, with whatever it is, I think it's, uh, it's awesome. Jim O' So one half of what we're doing

SPEAKER_269: is, is that, and that's, you know, Grammarly is a big AI superhighway. We're going to, uh, use it as a platform for the future of AI agents. Jim O' The other half is applications. And, you know, this is something the, the Grammarly team's approach to this was, you know, up till now, Grammarly has mostly been the magic is that it works everywhere else. Um, and it's always kind of lacked

SPEAKER_258: a home. It's lacked a, a, uh, a destination that feels like your true writing space. There is a

SPEAKER_36: Grammarly editor, which is quite good. Jim O' Yeah. But you know what? It's great, but they don't support it now on iPad and iPhone anymore, I don't think. So they seem to have gone away from trying to compete in the document space and they just want to be agentic. They want to be

SPEAKER_255: a co-pilot. Jim O' Yeah. I mean, it's a, it's a, uh, as you might imagine, a common debate inside Grammarly was, do you want to, do you want to work everywhere else? I like to say, Grammarly spent

SPEAKER_258: the last 15 years enabling every other blinking cursor. Coda spent the last 10 years trying to build a new one. And so there was this big question about does Grammarly want to be in that market? And so as Alex and Max walked me through their vision, they said, we've made a decision. We really want to be in that market too. And the reasoning really resonated with me. Their view

SPEAKER_269: was that, you know, we're going through this change and how people think about their tools. So they sort of see it as three phases. There was a phase of the PC phase and all the tools came out

SPEAKER_258: and they just felt like digitized versions of our physical tools. And that was the office days and so on. And then the second phase happened with the web and all of a sudden, uh, tools became about collaborating with people. And so it was Google docs and Slack, and it was all about communicating with

SPEAKER_269: all these other places, uh, with all these other people. And then this phase is AI and all of a sudden the applications work like people themselves and they're, they're sort of working alongside you. And that's the sort of role of the agent. And what they said that I, I thought was really interesting is they get a lot of pressure from, from, uh, users of, do I have a home for my work? But more importantly, what is that first class place to understand an AI native collaboration surface? And they had this plan for how to go build that. And we sat down and looked at what we're, what we built.

SPEAKER_308: Yeah. And the plan was your roadmap notion and what other people are doing. So,

SPEAKER_36: hey, now the two of you come together, peanut butter, chocolate. It's such a great combination.

SPEAKER_312: Peanut butter chocolate, I'll use that one. I've been using Lego pieces and puzzle pieces, but I like peanut butter. No, it's peanut butter chocolate,

SPEAKER_179: you know, like you guys, and this is the thing that makes me crazy about the whole lot of you in SaaS,

Chamath Palihapitiya: is that you're always trying to upsell me on AI. Stop with this upselling, just put the best out there for everybody. Because what I'll tell you what happens with AI is 1% of your, or maybe 5% of the

SPEAKER_179: people at your company use it. But then you guys demand that I pay a hundred percent of my users pay for it. Now you put in this cognitive dissonance. I got the top performers saying, I need this tool. And then you guys are saying like, well, everybody's got to pay for it. And I'm like,

SPEAKER_01: oh my God, I got to pay another $2,000 a year, Grammarly or Coda or Notion or whoever wants this

SPEAKER_03: upsell on AI. Just include it. Yeah. Shazure, can you fix the entire SaaS industry for Jason? Please,

SPEAKER_06: just go ahead and talk to everybody. It's enough with the upselling on the AI.

SPEAKER_07: I'm going to start with your two favorite products and then I'll work from there.

SPEAKER_01: Are you going to do that? Well, you promise me right now, we'll stop with the two different

SPEAKER_320: non AI ghetto and then the AI elites. That is the Coda approach. Coda AI is included for

SPEAKER_258: everybody. Thank you. I mean, Grammarly is the obvious approach because the whole product is AI. So how we bring that together, we'll get to, but yes, I agree with the sentiment. I understand very much that the haves and have nots of the creates a weird choice for the buyer.

SPEAKER_328: Alex, do you have any questions? Oh, I got a bunch.

SPEAKER_48: I mean, we have two ways to go with this. There's product, which I've taken up all the product discussion, but you love markets and you love M&A. So I think you probably have some M&A questions or some business architecture questions.

SPEAKER_147: Oh, hell yes, I do. So one thing that blew my mind when I was reading about the deal was that it was

SPEAKER_18: done in stock according to Axios and was done, I believe at the last kind of valuation marks for the companies. And Grammarly was valued, Jason backed me up like 13 billion. And I think Coda was valued at like 1.3 last time. So Dan Premack writes that it's about 90% Grammarly stock and like 10% Coda stock. I'm curious why you guys went with those prices at, to set the rate to consummate the deal. Cause to me, those are older numbers.

SPEAKER_334: Are those numbers correct or ballpark correct?

SPEAKER_335: Let's start there.

SPEAKER_258: I can't, I can't comment on those numbers. So the numbers are correct on the last valuations on our last round was done at 1.3, 1.4 post. Um, and the Grammarly last round was done at, at 13 billion. So I can't, I can't really verify anything about the actual numbers. You can draw your own conclusions on that. Uh, maybe I can comment on the process of, of coming to a decision like that because I think it's, it's always hard. Um, I think the, the, there are different ways to, to, um, to approach these situations. So sometimes you do it based on revaluing companies and you sit down and you try to come up with, you know, the metrics and you come with cash flows and so on. And sometimes you take the approach of we're about to build something new and it feels a little bit more like, I'm sure you've coached founders on this before, a little bit like when the founders are coming out and saying like, Hey, we're going to kind of split things up here. What's a reasonable way to split things up based on the future. And this was definitely a case of us saying, we're going to build this together. We have these two vision memos that are, that are quite similar. And we came up with what we felt was a reasonable split between them that I think is appropriately fair for everybody. Um, but I think that's, it's much more looking to the future than it is looking to the past and focus on profitability. I assume,

SPEAKER_126: uh, I mean, you're in investment mode right now. I would also assume that,

SPEAKER_311: Brad Granley has been profitable for a long time. Yeah.

SPEAKER_228: Oh, wow. That's fantastic. So, you know, there's a big discussion. I don't know if you saw the Satya

SPEAKER_36: Nadella comment the other day that likes on the BG2 pod, they had this like interesting thing where he said, listen, there's databases on the backend and there's just going to be AI agents that actually kind of aligns with your vision. So, uh, but these companies are going to take less and less in terms of infrastructure and people to run them as an, you know, entrepreneur for a while now. Like, what do you think about how many people it takes to grow these companies? Another CEO, I don't know who was the Klarna CEO or somebody, you know, said a flipping comment, like, why would I hire anybody? I can just build agents. I can just build, use AI. And, you know, he said that

SPEAKER_179: and people were like, he's lying. It's not true. I think he said the quiet part out loud

SPEAKER_01: is that the next hire is harder than hiring somebody is sometimes, sometimes, oftentimes. I don't know. Is it all the time? Most times? I don't know if it's most times or often. I would

SPEAKER_48: say often, perhaps somewhere between often and most times. The most, the best use of your resources

SPEAKER_04: is to automate something and use technology than to hire and throw a human at it where it hasn't always been that way. How do you feel about that?

SPEAKER_269: So I really like Sebastian. I think he's done an amazing job with Klarna and spent a bunch of time

SPEAKER_258: with him on this topic. I think they, and he does, he makes actually a couple of different observations. One, one is that he thinks AI can build, rebuild all the software you need to run the business. And the other one is about, about hiring. As you mentioned, those are sort of two separate observations. And I think I, I count myself as an AI optimist. My, my view is it's going to be a net additive to the world. I mean, there's definitely a view of the world that is, as technology advances, every wave of it, as, as you both know, there's been this fear of, will it replace jobs? Will it make it impossible? And obviously it does in some, in some places. So technology over the last hundred years has shifted jobs, but I think in almost every case, it creates them instead of removing them. And I don't think it's, I think it up levels people. I mean,

SPEAKER_269: the idea of, I'm going to have agents to go and, you know, help me do my research and help me prep for my day or, or, or help me write code or so on. It's great. What does that allow me to do? It allows me to up level and allows every person to up level what they're doing and changes the jobs. And I, so I, I think, yes, it's going to change the nature of jobs. I think it's going to create jobs. Maybe on the second part of it, the, will it replace applications? I think that's slightly misguided. And the reason I, maybe just give you analogy. Imagine I came and told you, AI can write great assembly code. You can like automatically generate as much assembly code

SPEAKER_258: as you want. And you'd say, who gives a shit about that? Yeah. I got them writes assembly code. I mean, there are people who do,

SPEAKER_269: but there's not a lot of them. Yeah, that's not, not, not. And so, yes, of course, can, can AI, it's a magic trick. Can it generate this thing? But I actually think AI

SPEAKER_258: working in a surface with building blocks that I understand is very important. The fact that AI can go and create something that I understand that I can still work with. And I think this is the idea of AI should feel, we call this user-centric AI, but the, the idea that AI should feel like something that's working with you in a surface you understand, and doesn't produce a bunch of, uh, you know, I think one of Sebastian's statements was we can just generate Salesforce or Workday or so on. I'm sure you could, I'm not, I'm not, I, I, but then what do you do with it? Yes. What, what, where does it go next? Go ahead, Dan.

Chamath Palihapitiya: I, I got a double click on this one again. So glad you came on and I appreciate you coming on. I know you're very busy. It's the holiday season and you're in the middle of this, but you guys, it's such an important discussion for us to have. And I have to say, you know, I have found, um, young people, uh, who are what I'll call AI first people. They use ChatGPT first before a Google search or just a search because Google now has Gemini. So they'll do a Gemini deep research. Have you played with deep research by the way? 1.5 yet? Oh my God. Literally the second you get off this pay for 20 bucks on your Gmail account to get Gemini 1.5 with deep research.

SPEAKER_60: It's scary. Scary. Good. Um, putting that aside, I have people on my team now. We have 20 people.

SPEAKER_142: We're paying you what? 20 bucks a month for a Coda? 30 bucks. What is our, what is the rack rate?

SPEAKER_361: You're probably at 30, I guess.

SPEAKER_01: Okay. So I'm paying you 360 bucks per person. I'm spending 7,000 a year with you, maybe 10, whatever. Um, they just did a, a function. I'm not gonna say what it is, but this is something that's time consuming in a venture firm that processes a lot of applications. And they said, you know, we save this amount of time. And I said, this amount of time per what? And they said, well, per like transaction or whatever, or per processing thing. And I said, well, how many of those do we do a year? And they

SPEAKER_366: said 20,000. 20,000 times five minutes, you saved a hundred thousand minutes. So yeah, we saved a

SPEAKER_48: hundred thousand minutes. And I was like, okay, people work 2000 hours a year. Like I can literally

SPEAKER_01: put numbers on what's being saved now. And I can tell you having high performers, bar raisers in your organization who are AI first, who use code first, who have Grammarly turned on everywhere. And they took the time to turn it on. These people are 50 times more valuable for my organization than the people who are not doing it. I'm not saying that to scare everybody inside my organization, please clip this and send it to the people who aren't using it. Like you're not long for this earth if

SPEAKER_48: you're not doing this in my organization or others, because there's a group of people behind you who are making applications that save a hundred thousand minutes. And you're thinking in those terms.

SPEAKER_60: What this means to me is you've got a great future, by the way. I mean, I do think like people

SPEAKER_36: building apps is the future and the agent following you along, but I'm going to disagree. I think static team size is going to be the future. If you are an elite CEO, elite leader, Alex, I think you're going to look first at automating with the existing team rather than throwing a body at it,

Chamath Palihapitiya: hiring more people. It's just too much work. And having the same team size, like you're the Navy SEALs. You're like the same team size, but everybody got a little bit better. Everybody's a medic. Everybody knows how to use the defibrillator. Everybody knows how to use CODA. Everybody knows how to use a sniper rifle. Everybody knows how to fly the Apache, whatever. Like if you can actually do that at every Navy SEALs and Apache sniper and medic, which are three distinct fields in, you know, the military, then all of a sudden you got like a really killer team. So anyway, I'm taking the other side.

SPEAKER_18: Well, no, you also do away with middle management. I mean, what's the bloke problem at every single major tech company? They always say, we have too many layers. Where did all these layers come from? I'm like, you literally made them out the door.

SPEAKER_269: Just to be, just to be super clear, I'm in the productivity business. I've been trying to help

SPEAKER_258: people be more productive for years. So it's not, I'm not averse to the, the, the idea that using the right tools can make you individually and your teams and your companies a hundred times, thousand times. I mean, obviously that's, that's my whole business. And maybe to give you one, you were mentioning earlier about things people do with Coda and you can kind of use it as a workflow tool and you can do automations and so on. I think last year we automated something like 4 billion tasks. I mean, the, the, the, just on, on, on Coda. So, I mean, you know, go multiply that by how much time you would have spent and every one of those emails sent or every one of the notifications or the updates to Salesforce or whatever, whatever it might be. So I'm, I'm, I'm definitely aligned with you on that. My statement isn't so much the, you know, any of that's going to stay static. It's my job to go make that a thousand times, a hundred, a hundred thousand times better. My point is more that I think the, the human imagination is really high. And the, the idea that my company size is going to stay static because I can do the same thing with the same amount of people. Yes. If you're just trying to do the same thing, I think that's true.

SPEAKER_269: But then you're going to sit down and you're going to dream up to like what I could do something else.

SPEAKER_379: Okay. Fair enough. Yeah. That's true. But man, see, this is the natural tension. I think

SPEAKER_179: you're going to look at that and say, yeah, you know, we want to start this new year. You're

SPEAKER_36: not thinking, okay, I'm taking, I'm going to pluck these two people out of this unit and they're going to start it. And then these people are going to pick up their slack and automate whatever

SPEAKER_255: they did. But maybe I'll, I'll use you as the example, like it's nuts to go run an investment fund and multiple podcasts and so on. You could have stopped. You could have said, I've got this

SPEAKER_386: kind of fund. I've got this magical Coda thing. I'm trying to stop. You didn't. Now, because of

SPEAKER_269: Zoom, you, we don't like, this took no time to go get this set up. Technology made it possible. You know, because of Coda, you now can invest with like so much less of your attention than you used to be able to invest. That's all true. What'd you do? You filled your time with new productive things to go do. That was my, that's my AI optimism view. Productivity, I think produces

SPEAKER_391: humans. Alex, who's right here?

SPEAKER_101: Alex, who's right here? Neither of you. I'm right. But critically, I want to ask about the agent point, because I know we have

SPEAKER_18: to go in a second, but in your discussion of the plan for the combined company, you said you're going to weave the best of Coda and Grammarly together. Of course, you're going to combine company knowledge, Gen AI, chat features, full productivity suite, and hundreds of agents. Now, the thing that I struggle with is what does hundreds of agents mean and how do I touch them? Because I feel like people have their own view of what agentic AI means. And so in your future vision, where do the agents live and

SPEAKER_24: how do I know to call them? How do I know to interact with them? And what does that really look

SPEAKER_269: like in your view? Yeah. I mean, it's a great question. I don't know if touch them is quite the right phrase I would use. Interact with them very politely with consent. Yes, absolutely. Yeah. And I mean, I think that, and I think this is like defining this word is really important.

SPEAKER_258: So we spend a lot of time on what's an agent. I think of an agent as a mix of a few different

SPEAKER_269: things. An agent, if you sort of analogize it to human, an agent is something that has context, so it has a certain set of knowledge. It knows certain things. The Grammarly agent knows the entire entirety of everything about grammar. The Gmail pack on Coda knows everything in your email, right? So there's context first. Second, skills. So there's things that this agent can do, right? The Grammarly agent can go and revise your work. It can go and answer questions about it. You know, the Coda Gmail pack can go and send emails on your behalf. I think there's jobs. There's things you've asked it to do all the time. You said, hey, auto send emails out to my investors. You said, please correct my grammar in these places. And then finally, I call it the instructions. I'm going to talk to this thing. I'm going to say, hey, right now, I'd like you to do this. Right now, I'd like you to answer this question. So you take

SPEAKER_258: all those pieces together. And in my mind, that's what agent ends up meaning. It's something that's context, skills, jobs, and instructions that that's it. And so I'm working on how to, um,

SPEAKER_340: context. We understand, right? And the skills. Yeah. The knowledge. Okay. Yes.

SPEAKER_269: Those are the actions. Skills are like the buttons you can press. These are the activators.

SPEAKER_400: Got it. Update a database, whatever it is. Send something to Slack. Yeah.

SPEAKER_269: Jobs are your work list. Jobs are, hey, I hired this person. I said, this is what I'd like to get done. And then instructions I think of as one level up is kind of the personality. It's the, it's the current thing I'm trying to get done and the, what I, how I want you to behave with, with me. And at that point, I expect to interact with an agent in many ways, like I interact with humans. I want to be able to assign tasks to agents. I want, you know, what, what, one magical thing about the, the, the Grammarly experience is it kind of feels like somebody came along, this really good proofreader, like Jason, the, the, the, the way you were talking about it earlier of like, hey, I'm sick of reviewing all my team's writing. Right. What does Grammarly do? It does what I would have done. I come through and

SPEAKER_403: uses, you know, red and blue marker and says, good, bad. And then teaches you, by the way,

SPEAKER_227: this is why you're using then and then wrong. That's right. Yeah. Time and yeah. Yeah.

SPEAKER_258: So that, I think, and I know you had to squint a little bit to see how these things all come together. And I know the term is going to feel slightly unfamiliar, but if you start, if you think about it, like applications act like people, I think it's the easiest way to think about it.

SPEAKER_408: I like that. Apps acting like people. Yeah. It makes total sense. But the substance is,

SPEAKER_48: and Grammarly studying all this. So it reminds me of what Apple intelligence is doing. Alex, we talked about this on a previous episode. Yes. In Apple intelligence, you can go in and you can say, let Apple intelligence study what you're doing in this app, essentially. And it's on an app by app basis. So people are turning it off for things like signal or telegram or iMessage, because they don't want Apple intelligence listening in on those conversations, uh, for obvious reasons. Um, you know, but they do want it watching. I use Spotify or co-buzz another high res music service because co-buzz hasn't written Siri integrations yet. But if Apple intelligence is studying how you use co-buzz when I'm on the ski slope and I say, play this song in co-buzz or play me, you know, whatever, uh, it should know how to do that. It doesn't because it hasn't been written. So you have

SPEAKER_36: Grammarly studying how people do things. You have code explicitly having people write apps. So this is like a pretty awesome combination. And Athena, Alex is also down this path. Uh, everybody go to Athena, wow.com to get a month of, uh, I'm an investor. Athena has a virtual assistants and they're studying and watching them work and then using that to back into agents and

SPEAKER_228: technology. So a lot of people are onto this business process studying and applications. I think it is the future. I think you guys are perfectly positioned, but, but the thing that

SPEAKER_147: we're getting to here is that it's not going to look like one single thing. So there's ambient agents like Grammarly, which follows you around everywhere you go. There's things that Shushir

SPEAKER_18: mentioned that you would have to prompt directly. And then there's other things that might look more like a, like a to-do list or a task list that you assign to it. So I think the things I struggle

SPEAKER_411: with is agentic AI seems to be everything. That's not chat GPT. No, here's an idea.

SPEAKER_269: No, I would include chat GPT. Why, why I think of chat GPT is, and I go back to my analogy, apps acting like people, what did we all love about chat GPT? What was the thing that blew our minds about chat GPT is that I could go to this thing. I mean, we've all had the experience, just sitting at dinner, somebody asked a question, what would you do? You turn to your smart friend and say, what do you think? And instead I can go talk to this thing and it acts like a person. It's an amazing person. It's a person that read the entire internet. What's the context for chat GPT?

SPEAKER_349: It's the entire internet. Yeah. It's all human knowledge. I do see the vision here pretty

SPEAKER_36: clearly. If Grammarly, one of the concepts that people always need is give me an idea of what to write next, right? The blank sheet of paper, what Grammarly and Coda should do next is they should say social media. I use my Athena assistant to go find everybody talking about Founder University or Accelerator, yada, yada. And then they share that in a Slack group. It's a manual process, but it's

SPEAKER_60: well worth the 33,000 a month I pay for it. It's about one thing the person does. Instead of having a

SPEAKER_36: social media manager, they just put, here's a list of the people on LinkedIn, Twitter, Instagram, talking about us, our firm and our companies, go engage. And then they go and say, who hasn't

SPEAKER_60: engaged? Who did engage? And they just remind everyone on the team very gently. These people have

SPEAKER_23: engaged. If everybody else could engage, that'd be great. Just means go find the Founder. If you were one of our Founders and you were talking about this merger- I did this exercise yesterday. Yeah. But now you do it. How do you do it? You have somebody prepare it for you and give you a set of links? No, we have a Coda doc for it. Oh, you have a Coda doc, right. So people just dump them into

SPEAKER_60: there. So now social media monitoring as an app with Grammarly, plus you combine what should you write next? So here's everything. These are people that you should engage with. This person from

SPEAKER_36: com.com, Alex from com.com tweeted about a new feature. Would you like to engage? Here are possible responses, click one and edit it and then post it. So a social media agent should be the next one you do. Yeah. And it should just be built into Grammarly. It should just follow you around, but it should all be documented in the single source of truth and the database should be on Coda. We,

SPEAKER_60: I would have people build that right now, but I think it's probably your next best use of time

SPEAKER_427: because everybody needs help with that. Yeah. Those are great ideas. I think this is the way I get

SPEAKER_165: myself to the top of the product roadmap. But sometimes it's gotta be a startup doing this too, like a social media, where's the social media monitoring startup? I think, I think part of the

SPEAKER_269: reason, um, there are, I mean, there's a bunch of them, but the problem I think is most, most applications built that way. They they're forced to be rigid cookie crudder things. So they'll build a version of it that is designed to run a very specific kind of company. Why do you guys use Coda? Because nobody builds software for your unique way of investing for you. So the thing you want, I think, and the fact that Sebastian and Klarna and so on is I think he's right. You don't know, not everybody needs exactly the same Salesforce. And I do think you should, you should have your own version. Like you guys have your magic Coda doc that runs everything.

SPEAKER_36: We did not specifically use any of the CRMs out there. And there are CRMs now, verticalized ones for venture. We looked at them and we're like, but that's not how we work. That's how the average venture, it's not my business. I need something that does these customizations. And I looked at it and I was like, you know what people were going to charge us for

SPEAKER_01: that stuff? A hundred thousand, 200,000 a year. Cause they're like, well, you're a VC firm. We'll just take a 1% tax on your, whatever. We'll just tax you one or 2%. I was like, that. I bleeped that out. I was that. I'll just build it in Coda or Notion. I'll use Zapier if this, then that, you know, and I'll have it done. It may take me two or three weeks to get this done, but I'll just have some of the young guns on my team, build it. And by the time they build, it would have been us debating the price with you and training and creating logins for everybody.

SPEAKER_435: I would rather just go into Coda or Notion and just do it.

SPEAKER_255: And just to be clear, I think, I think that's a good example. I mean, if you think about what we're trying to get done here, agents and applications, what do you really want? You want

SPEAKER_258: a social media agent that really feels like a human that does all the things that you,

SPEAKER_269: you know, would you actually hire a million people to go do that? Probably not, but now, now, now you have this thing that can do it. And you want an application that feels like your unique perspective on how to do it. You don't want the one everybody else does, that everybody else uses. You want your unique way to do it. Cause that's your strength. Like that's, that's what you do. Superpower.

SPEAKER_258: Putting those together. That's the heart of what we're doing here at the combined family and Coda.

SPEAKER_228: All right. Listen, I am a super fan of yours and the products. Uh, why don't you come on in six months? We'll book it right now. Welcome to the twist500.com powered by Coda. Uh, you're not a sponsor of it, but we give you the plug anyway, because it's such a great tool, uh, twist500.com built on Coda. Okay. It is what it is, folks.

SPEAKER_36: It's a great product. So you can always tell, you can always tell like which products I love. Cause I'm just like, this is, if I use it, we love it. You know?

SPEAKER_439: Authentic. Yeah.

SPEAKER_36: Yeah. And so I'm pretty authentic. Like we're using it. Um, and so maybe in six months, let's talk about how the integration work. Cause I would love to hear, is this the first time you've done an M and a kind of deal like this?

SPEAKER_398: Well, I mean, I was, uh, I record obviously, but you know, I was at Google for a long time. I was YouTube. Google was, you know, a different version of the same thing.

SPEAKER_302: I bought many companies. So definitely been through them. This one's unique. Every one of them is

SPEAKER_36: unique, but the, uh, I want you to just, if you could be thoughtful when you're doing this and maybe write down the lessons you learned in this merger, in this time period, would you be willing to maybe in three, six months, come back and just tell, tell, tell everybody what you learned about doing this kind of M and a. Cause I do believe we're going to be in a new M and a moment where companies like yours, uh, and other companies, like I, I can't imagine Zapier, if this,

SPEAKER_60: then that yourselves, et cetera, that we might not see more of these get togethers and these products can remain independent and you can buy them that way, but they could also be deeply integrated and come together. Just a great vision. So congratulations to your, uh, general catalyst,

Chamath Palihapitiya: uh, partner who, uh, who was that, that came up with this idea? Give them a handbook.

SPEAKER_446: Heymonth brokered it. Yeah. Heymonth's an agent. Yeah.

Chamath Palihapitiya: So he's, he's your agent. He's your, he's agentically looking at the portfolio and saying, Hey, sometimes one plus one equals three, you know, and I think that's how VCs look at this great job.

SPEAKER_60: And we'll talk to you in three to six months when you're, when you're out of the weeds.

SPEAKER_450: Sounds great. Thanks for having me.

SPEAKER_60: Ah, appreciate it brother. And that's, you know, that's what you get here, Alex on the short term is, uh, we're able to pull the guests who are breaking news. This is all news that happened

SPEAKER_04: in the last 48 hours and we got them first on this week in startups for founders by founders,

SPEAKER_147: getting it done. Absolutely. Also, this just reminds me of my favorite headline of all time, which was from 2014 entitled pure storage CEO acquisitions always suck worse than you think.

SPEAKER_18: So I'm looking forward to seeing how that was wrong 10 years ago. No, no, it's, you know,

SPEAKER_60: M and a is hard for every YouTube deal or Instagram deal. You know, you got another 20 that just

SPEAKER_01: returned nothing. But if you look at those two deals and use them as a proxy, those two deals

SPEAKER_48: define the modern you, the modern Google and the modern Facebook, Facebook didn't buy Instagram.

SPEAKER_119: Facebook would be in decline right now. I predict. Oh, oh, massively. And it would have been, I think, for the last three, four or five years. I mean, seriously, when's the last time you fired up Facebook

SPEAKER_228: on purpose? Never, never. I mean, I, the only reason my Facebook's being updated right now is because once in a while I'll do something on Instagram and it gets syndicated there. Another

SPEAKER_36: great guest. All right, everybody. What an amazing episode. We'll see you again on the next one. Remember 2025, we're Monday, Wednesday, Friday at 10 AM Pacific noon in Texas time. And then Alex time,

SPEAKER_60: Pacific coast, 1 PM. We're live, live, live on YouTube and X, X.com slash Alex, X.com slash Jason, X.com slash TWI startups. And YouTube, just search for this week's startups, hit subscribe, hit the bell, subscribe to Alex's amazing newsletter. Cautiously optimistic. Cautious optimism. Cautious optimism. Cautious optimism. Cautious optimism. And that's on Substack where you just type in cautious optimism, Alex Wilhelm. Thank you to our partners and we'll see you all next time.