Why do all my clients want to kill you? What do they have against you? What did you do to all of them?
I can't comment on those other companies, but I guess when you do something that's going pretty well, you've got a quarter billion people using Canva, I think maybe they want a piece of that.
Is Canva tech's most underrated company? Nine years of profitability, hundreds of millions of normie users outside tech that are so hard to get, and now a big pivot from design platform to AI platform. Today on Access, we have co-founder and CEO Melanie Perkins to discuss everything they just announced today at their big event here in LA. But first, Alex and I discuss our waning interest in Open Claw, my new addiction to the AI assisted Poke, which was made by last week's guest, and the big reputational crisis facing the AI industry. Welcome to Access.
My name is Ellis Hamburger, and I'm here with my lovely co-host Alex Heath. We've got Melanie from Canva on the pod today with some actual news, which isn't always the story of our show. We're trying to get the real talk and the real takes, not just the embargoes. But we have a bit of both today. But first, Alex, will you be attending ClawCon in Los Angeles tonight? We need to make LA tech happen.
Hell no. I think it's going to be a madhouse. I'm too old for this shit, man.
Really? A madhouse? Part of me really wants to go, especially when it was first announced. So these ClawCon pop-ups have been happening all over the world really for the last couple of months.
I thought you'd live for the claw. Ain't no laws when we're using claws.
We've talked a lot about the ClawFather, Peter Steinberger, and Open Claw. I have been very interested in it, but my view of it has actually changed in the last, I would say, couple of weeks. And I would say our episode with the CEO of Poke, Marvin, last week actually played a role in me souring on going to this event as well. I just feel like claw is a thing that we're going to look back on and go, 'Wow, wasn't that a really cool nerdy time, like when people used to jailbreak their iPhones?' But then Apple, you know, started adding stuff to the OS and opening up over time. And I think the analogy here is services like Poke, which again, we featured on the show last week, basically simplifying what Open Claw is for people like me who don't have time to chain together a bunch of stuff with a Mac mini.
Okay. So maybe we can cut in all the times you've talked about Open Claw and me being unimpressed for this exact reason from previous episodes of the show, because it just felt kind of so inevitable and also weirdly surprising that it didn't already exist in some way. And we're already starting to see people just natural language it and have the agents hook up MCPs on their own and have it all just work in the server with no Mac mini. But yeah, I've been pretty poke-pilled as well since last week. I mean, it's honestly been pretty mind-blowing having a tool that's just built for this. I think it is shockingly ahead of the game and was a few months ago. My first exposure to some of the agentic natural language programming was actually with Notion custom agents. But seeing with Poke that can handle Notion MCP pretty darn well, if not in some ways as well as Notion AI at this moment, it's like now I can do things like, 'Hey, Poke, when you see client invoices hit on my Mercury account, shoutout to Mercury, who I'm also a new customer of. I just use whatever's on the show, I guess, now.'
Is this the third guest company that you've shouted out in our intro so far? We've got Ivan from Notion, Amod from Mercury.
We do have good companies. And so I say, 'When you see the bill hit my account on Mercury, mark it as paid in my Notion database.' And just like little experiments like that. And I think one of the things Marvin said that was a good prompt was just think of it like an EA. Like what would you ask an EA to do? And so that's one of those things that I would have asked an EA to do. The other thing I would have asked an EA to do is update all my goddamn services with my new Mercury credit card instead of my Chase one. And Poke said it could do that and then messed it up and couldn't do it.
Full Mercury like with the credit card and everything?
Yeah, I mean, dude, my overhead with Meaning is like insanely low. It's like a thousand dollars a month.
But what about those Chase points? I mean, come on.
I mean, my personal spending is going crazy on Chase. My business spending, very low. Okay. And when Poke doesn't do so well, I say, 'Hey, Poke, draft a bug report email to the team and send it.' Which is also very satisfying because I've been a bug reporter for my entire life, Snapchat, browser company, etc. And I spend a lot of time reporting bugs. And now I can just fire it off, have Poke do it for me. So I try not to do this, but sometimes when we have people on the show, I haven't had time to really dig into the products that they build. Poke was one of those things where it just looked cool, and I figured it was going to be a thing people were talking about, and I think correctly so. I saw Poke in the news quite a bit last week, at least in our little bubble. But I hadn't quite tried it, so it's good to hear from you that you think it's legit. I've also been getting the daily summaries from it. And it's decent. It forgot that I'm back on the West Coast after a trip to the East Coast last week, so it's not quite as smart as I would like. Still a little rough around the edges, but you're right. I mean, the form factor and the approach that it has with the way it communicates and where it decides to prompt you a little more or go the extra mile seems to be quite good. I honestly didn't know the tone could be as much of a lever as it is in Poke. People talk about Claude having a nicer tone.
Right, but people talk about Claude having a nicer tone, but Poke isn't just a nice tone. The closest thing to it is probably the Friend AI pendant. It's a little sarcastic, a little pushy. Definitely brief. Definitely mention you've made now. And it makes fun of me sometimes, which I actually genuinely appreciate, but it also just knows when to be quiet and work. So I was talking about some potential dinner spots I was looking at for some trip. And I said which one looked best to me, and it said, 'Hmm, that surprises me. Not very Rick Rubin of tech of you.' I'm like, 'Jesus.' You're pulling quotes from my website and roasting me for that.
Wow. Yeah, you weren't looking in Malibu for a restaurant.
No. That would make you Rick Rubin of tech. And it does feel like it knows me better than the others. It feels like they have some good prompts in place to remember your taste and things like that. So it's given me very good recommendations, and it really does feel like a friend as opposed to your AI assistant who's going to write you a five-paragraph essay every time. Okay, so we're long Poke, short Open Claw? I mean, different audiences, right? It's like iOS and Android. Android people want to tweak their home screen with interesting widgets that don't work.
Well, that's great. Another thing that I wanted to talk about with you, Ellis, is that the Social Network sequel is apparently going to be good, I regret to inform you. So there was a teaser shown at CinemaCon this week. I saw friend of the show, Matt Belloni, said that it might be the thing that actually breaks the internet, referring to Jeremy Strong's Mark Zuckerberg in this movie. Gave it five stars, which is high praise.
Stunning casting. Couldn't be better. Yeah, man, Kendall Roy's going full Zuck, and I'm here for it. I think Zuck should embrace this. You know, he famously back in the day for the Social Network, the first one, rented a theater for Facebook employees to go watch it and really leaned into the ridiculousness of it. And look, it's not going to be a favorable movie. It's about Frances Haugen, who I've met and interviewed, and the Facebook files, which I covered extensively, that were all the internal documents that came out about all the rough things that Facebook was doing and the way it was...
As I reported. As you know, shoutout really to Jeff Horwitz, the Wall Street Journal reporter who famously wore a headband on national television as he was breaking this story looking like he was in a dojo. I know Jeff was always a great guy.
Oh, just in the arena. Yeah, he's just in the arena, and now he's being played by The Bear's Jeremy Allen White, and Frances Haugen is being played by Mikey Madison, which will be I think a very great cast. And yeah, I think Zuck, even though this is going to paint him as a savage monopolistic capitalist who doesn't care about children, I think you should lean into this. I think you should rent a theater again. I think you should play into the memes because it's going to be obviously a hard one. And I think if you just ignore it, you kind of give it more life. I mean, what choice do you have? What choice do you really have as I think an executive these days? But to laugh at the joke as well. Yeah. I'm just... Aaron Sorkin's obviously goated. I love a lot of his early stuff. I haven't liked some of the movies he's done in recent years. There was one set in a courtroom that I can't even remember. I think I fell asleep a couple of years ago. He hasn't had a ton of great stuff. So, I'm hoping this is actually a return to form. And I didn't actually look to see if Trent Reznor was doing the soundtrack again. That would be dope. But, yeah, we'll have to go see this. We'll have a little Access premiere. I cannot wait. For any number of reasons, Social Network is one of my top 10 movies ever and it really is hard to overstate how influential it has been on the tech industry. Probably even more so than her and many others because it incorporated music that has set the tone of coding for 15 years since then. So many of the archetypal narratives that executives take part in or have to come that. It really has just been so influential and really just is a stunning movie. And so, I'm hopeful that Aaron Sorkin is not washed and can make this happen again.
Yeah, and we're going to keep the tech meets Hollywood theme going next week on the show. We've got Jonathan Glazer who is the showrunner of The Audacity, a new show on AMC that is the spiritual successor to Silicon Valley, the former HBO show. I talked about it briefly on the last episode, but really excited to have him on. He'll be our first non-tech founder type on the show. So, we're branching out.
Can I get like a pre-read, pre-watch?
I've got the screeners. I'll send them to you.
Is that legal? Did you just cause yourself a problem? If my Poke was working, it would have already sent them to you. But, we'll make sure to address that. So, I had Poke send a calendar invite to someone and I asked the person if they successfully received it and they said, 'Yeah, but I think Poke misspelled the name of my company.' And I was like, 'Ha ha ha. Yeah.' No, that was my fault.
See, that's also what I'm terrified with agentic AI is it's just really messing something up or compromising a deal or...
Oh, it asked me for permission. Don't worry about that.
You're not going to look at everything, you know? You're giving your life over to this thing. Well, the only other thing I wanted to talk about before we get into our convo with Melanie was I think a confluence of two things that to me are...
Yeah, gosh. For those who didn't see, it's important to laugh, I guess, but this is pretty serious. There were two attacks on Sam Altman's home. One of them threw a Molotov cocktail over the gate. And then another was like a drive-by, firing rounds into the side of the property. Thankfully, no one was hurt. And Sam posted a blog about it over the weekend that I'm sure a lot of listeners of this show have already seen. To me, this is obviously a horrible thing and attacks on tech executives have been happening. I remember I covered in 2018, there was a horrific shooting at YouTube's headquarters. And it was about something to do with demonetizing a video or banning someone and the person got mad and this stuff happens, right? But, I think what's happening with AI and I wrote about this in Sources this week is a real reputational crisis that I think the industry has not come to terms with and is not focusing on enough. So, on Monday, Stanford, they do this big AI index every year. It's kind of the Bible of where AI is at, and they release it once a year. And they had a bunch of interesting data this year about how horribly AI is being received by the general public and how the gap between how people feel about AI who work in it and are building it is really diverging from everyone else. And then you see these attacks on Altman's house and it's like, yeah, these are the consequences of what's happening, which is a lot of people hate AI. And they hate what it represents. They hate the mass unemployment narrative that companies have been putting out. And it feels like it's reaching a boiling point. I wasn't sure if you got to look into all of this, but my takeaway is that I think the industry should stop finger-pointing and focusing so much on the rat race of what everyone else is doing around, you know, I specifically I'm talking about OpenAI and Anthropic. And they should think more about this because this seems like the biggest problem of all.
Yeah, I think the first thing I want to say is violence is never the answer to some companies doing some things you don't like. Nope. I will also say, I wonder if in some ways AI is kind of becoming the symbol for the wealth inequality and lack of agency and mobility that so many people are feeling. And while it's always been on the left or on the right to talk about, you know, how we describe the rich or the elite. The AI leaders seem to be something that both sides can agree on are accelerating the demise of a specific group, the middle class or otherwise, that was already really, really, really struggling. And telling them to go start their own agentic small business tomorrow while they're struggling to make ends meet and can't afford groceries is not really going to work. And so, yeah, it seems like it definitely demands a different tack. And I will say I was also scrolling some feeds and half the time I favorite something on X, I now get a notification later that it was actually fake from community notes.
Yep. And I will say I think regulation does happen fastest when everybody is dog-fooding this shit, you know? And so, I mean, certainly it's in the news all day, every day. And could be the future of many companies' livelihoods and populations at stake. So, I would expect there to be a very strong reaction over the coming years. I hope so.
I hope so. I mean, so this guy that attacked with a Molotov cocktail, he's a 20-year-old from Texas. He then went to OpenAI HQ, smashed the glass doors with a chair and told security he was there to burn it down and kill anyone inside. And he had this anti-AI manifesto where he had a specific note to Altman saying, you know, if for some reason I don't succeed in killing you, I guess, then you should take this as a sign to, quote, redeem yourself. And then the memo was talking about AI replacing us, causing mass extinction. And then, you know, Altman puts out his post and he says that he sympathizes with this anti-tech sentiment that is clearly so pervasive. But, then he said, 'I believe technological progress can make the future unbelievably good for your family and mine.' And that may be true. Like, I believe that mostly, I think. I think you probably do, too. We talk about AI all the time. We're nerds about it. We love to try it. It's helping us with our businesses. But, most Americans don't think this, Sam. And this is the problem. So this Stanford report said that 38% of Americans view AI positively and that's the lowest ever recorded. And in China, it's 84%. It's happening with just the rest of the world. And then the trust in the US government to regulate is like 31%. So, people seem to just be incredibly pessimistic about AI.
Yeah, it seems like a little bit of a catch-22 for whichever administration is in power today or tomorrow. They don't want to cap the American growth and hegemony over tech. But, it also in some ways is just kind of like a repeat of some of the stuff Bill Clinton gets criticized for, accelerating globalization, which took so many jobs abroad. And sure, it did lower a lot of the costs of goods here, but for many people, they don't think it was worth it. And for many people, their salary or their wages haven't realistically, relatively gone up in, you know, 20 to 30 years. I always say think it's the only chart that matters. You know, you talk to people building AI and a lot of them privately believe that there's going to be more jobs in the future. But, they all also acknowledge that there's going to be this period where a lot of people lose jobs really quickly. And I think that's the tricky part is communicating this vision on the other side of, you know, technology tends to enable new kinds of jobs. I mean, like we would not be doing a podcast 50 years ago. The idea that this could be a business was ridiculous then. So, it will unlock a lot of new things, but that transitional period is going to be really messy and no one in the industry seems to know how to talk about that. And again, I just keep coming back to the industry is so focused on itself and competition and recruiting and virtue signaling to everyone in the bubble that they aren't paying attention to the fact that people are rioting about this stuff.
Has Dario written the only bad news essay?
I guess the citrine as you say. But, that was not a frontier lab. Altman has a lot of stuff about, you know, needing universal basic income because of such mass unemployment that's going to come, right? He has that Worldcoin startup that he's involved with. I mean, they all talk about this stuff. I just think there's a recalibration that needs to happen and I'm not really sure what's going to look like, but I mean, I think the attacks on Altman really show the stakes here. And I just hope that the industry wakes up. I mean, you kind of probably get, if I had to guess, you get lulled into a sense of this is just the way it is when you're an OpenAI employee and you're leaving the office and you see a sign reminding you to hide your badge as you walk out, which is something I noticed when I was walking out of OpenAI a few weeks ago.
I remember even going a couple years ago and this was well before all this mass unemployment rhetoric and I was asking like, 'Why is this an unmarked building? Like why is like there's no way of knowing that this is the headquarters?' And people were telling me at the time that they were already getting crazies showing up talking about AGI and trying to find AGI and this is just the stakes of this technology feel so high now that I just hope that people don't get used to going into their unmarked office every day and ignoring it, you know? Well, on a lighter note, today we have Canva co-founder and CEO Melanie Perkins to talk about some upcoming news they're dropping. Excited for this one. It's been in the works for a while. We've wanted to have Melanie on the show. She's obviously a super impressive founder and Canva's a very important company. We recorded this last week right before their big event in LA which is happening the day this episode drops and they've got a lot of big AI news which we get into. So thanks again to Melanie. Let's roll it.
Hi Melanie, how are you? Great to see you.
I'm well, how are you doing? Good to see you.
Are you in the throes of Create rehearsals or what's going on?
Yes, it's all happening. Create rehearsal, keynote prep, like every product's coming. It's all happening right now, but it's exciting.
Ellis has done some keynote prep in his time. Yeah, I'm curious like what does that look like these days? I feel like some folks including Apple are just like we're just going to do this all in post and just record it and then make everybody sit in an audience and just play the video. But there seem to be a lot of different approaches on like where is the leverage with a keynote these days? I don't know, what's your take?
Oh, that is totally... That'll be a very relaxed approach, wouldn't it? No, we have an audience of 5,000 people, all of our community come. It's a big event. We have a whole day of workshops and training and it's a very live experience. I think the keynote goes live like an hour or something after it's recorded, so it's minimal post.
I got to see Jensen do his keynote at GTC a few weeks ago and he doesn't have a script. He just freewheels it for like almost 3 hours. Are you very scripted? How regimented are we talking? Urge to freewheel your 2-hour keynote?
I have to say we're a little more prepped than that. But we also have a lot more, let's say, creative flair, you might see coming out on next Tuesday. There's a lot of elements and creative industries coming together which we're excited to bring to life. So there's a little more than just freestyling it for 2 hours.
How many cooks are in the kitchen approximately at this moment? Because I feel like keynotes are just so high pressure because they're live and I feel like everybody wants to have their two cents which can get a little crazy at times.
It starts off pretty small, but then as we come up to the day, we've now got hundreds of people working because there's going to be like hundreds of people on stage. I don't know exactly the number on stage. So there are a lot of people prepping all sorts of things to bring to life on the day. So there's I'd say hundreds of people involved actually at this point.
Melanie, how big is Canva these days? Like how many users do you guys have? I don't know if you disclose revenue. I know you're private, but give us a sense of the scale. Because Create is such a big thing.
I mean thousands of people. We've got more than a quarter billion people using Canva each month and we reached 4 billion in annualized revenue in December.
How do you think about the growth? Like has it always been kind of the last few years pretty steady or have you had kind of an inflection moment recently?
It's been pretty steady for the last decade. We've been launched for 12 years now and it's just been growing rapidly actually since day one which has been pretty epic and continuous and steady, but I think because so many people have been using Canva like one of our favorite charts is a cohort chart. You can see people that joined in 2013, there's more people using Canva and paying for it today than they were back in 2013. There's more people from that cohort upgrade to our paid products and use it more over time and then the same for every year since. So I think that's why it has been very continuous over the last decade.
I'm curious, how would you describe the growth? Is it moving into new teams within companies or is it just kind of more companies using the baseline stuff that's been there and been popular or how are you thinking about that?
It's all of the above, so it's new people that have never used Canva before. It's people using Canva for more of their workflow. So they might come in to create a presentation and then they stay for all of our docs and whiteboarding and all of the other things that you can create in Canva. Having that intuitive visual suite at the heart of it means you can learn one product and then you can take that to every other product like websites and so forth. And we've got 100 million teachers and students using the platform around the world. We've rolled out into 21 countries and then thousands upon thousands of districts, small businesses, individuals. It's really the full gamut. It's quite fun when you look at a chart breakdown of the different departments of a company using Canva, it actually mirrors their company. So if there's more sales folk at their company, there's more sales folk using Canva. If there's more engineers at the company, there's more engineers using Canva. So it really is a very broad split.
Ellis has a question that he needs to ask you about his clients. Ellis, just want to make sure you ask this.
Thanks for teeing me up, Alex. Why do all my clients want to kill you? I work with tech startups and I feel like we've got the deck ones. We've got the video ones. We've got the doc ones. We've got the design ones obviously. What do they have against you? What did you do to all of them?
There's such a vengeful spirit. Is it really vengeful or just like there's people want what Melanie's got?
I mean I can't comment on those other companies, but I guess when you do something that's going pretty well, you've got a quarter billion people using Canva. I think maybe they want a piece of that.
You just made two acquisitions, right?
Yeah, so maybe you could buy some of these that Ellis is talking about.
Well, I think like part of the question is just kind of about the classic bundling unbundling idea. I mean every company starts small, gets big, adds more features, becomes a suite, wants to be for everybody and then the startups come and peel out the individual use cases, make them simple again. Where do you feel like you stand in that trajectory right now?
I think if you look at the first decade of Canva and the genesis of the company was that there was a very fragmented industry. To just create a piece of design, you have to go to all these different products, go to different template sites, photography and illustrations and layouts and a different piece of software for everything you wanted to design. Our whole thing was taking all of that fragmentation and bringing it into one simple platform and making it accessible to all. I think what's really interesting is actually if you look at the parallels of creation today, actually the fragmentation is occurring now with all the different AI tools, all of the different things that have your memory and your context. And so the modern process of creation, we're really excited to again re-bring it all into one simple platform and making it accessible to the world which is exactly what we've been doing with Canva AI 2.0. So we're very excited about everything being in one place once again.
Melanie, you're from Perth, right?
And it's one of the most remote cities on Earth if not the, right?
