The Exchanges

Q not easy. And so I used Rome and I thought, okay, how do you build challenge into a product? And straight away came up my episode with Rahul from Superhuman, who said about Inserting challenge into a product. And so my question to you is, how do you think about that from the product perspective and how steep do you want to make the learning curve without being too steep?

A So our role model for this is Excel. Excel's got seven hundred fifty million users and the phrase we use is low floor, high ceiling. So a product should have a really, really, really easy entry point. And most of the people who are using Excel, right, they're using it for like shopping lists or they want a tool for making tables. And visually laying out the things they're thinking about. But then on the flip side, you have businesses that have been run for 30 years on a single Excel spreadsheet. And so in some ways, Excel is actually a better programming environment. It's a purely functional, reactive, interactive environment. And in some ways, it's better than a lot of the things we have in JavaScript or other programming paradigms. So we want something like that, where it's really easy for people to get started, and as they gain proficiency in the tool, and as they start to Take on more challenging problems. They're able to learn the tool and extend it and use it really as sort of a programming language for personal productivity and for personal development. A steep learning curve is a good thing, because if the mountain is steep, then it takes actually a relatively short amount of time to get to the top of it. It takes a lot more effort, but you can get pretty far pretty fast, and so that's kind of what we want, is that there can be slow paths up the mountain, and there can …

AI assessment note: “the phrase we use is low floor, high ceiling.”

Q Yeah, absolutely it has, and it's always been entertaining to see. In terms of kind of the cult-like element, you know, go-to-market's always crucial. And when we chatted before, you said we have the Tesla go-to-market. I'd love to dive into this. What did you mean by this, and how does that mean you think about kind of go-to-market moving forward?

A Yeah, so Tesla's business plan was build a sports car, use that money to make a more affordable car, use that money to make an even more affordable car, and for the many years that I was working on Rome, it was a pretty unfundable idea. I had some great friends who were investors or who knew a ton of investors, but everyone was like, well, isn't this just competing with Evernote? Isn't Evernote sort of a dead unicorn? And so the way that we were able to run the company, we were very, very fortunate, and also we made some decisions. I looked into what would be the ideal early user base. Who are some people who are thinking a lot about thinking, who are open to different kinds of habits and workflows that aren't supported by any existing tool, and who are going to be willing to learn those because the problems that they're facing are so complex that they needed a better system, they need a better tool. And so we actually got off the ground by getting customer funding from some folks in the effective altruist and particularly the AI safety community, people who are thinking about AI risk, and I would say like the broader Bay Area rationalist community, and so people who are really thinking about How they think, and how do they figure out what is true, and how do they process their emotions as well to deal with the biases that are inheritance from our ancestors. And so our go-to-ma…

AI assessment note: “Tesla's business plan was build a sports car, use that money to make a more affordable car”

Q That certainly rings true, though, when you look at Rome. As I said, I spoke to Mike and Anit beforehand, and they both mentioned that you've been definitely thinking about, but also working on Rome for a long time internally. What was that founding moment for you with Rome?

A In some ways, Rome's a continuation of the thing that first got me into tech, which was trying to solve problems of collective intelligence. And how groups of people can really tap into the wisdom of the whole group, the different expertise that individuals might have, and then figure out how to make sense of a complex world and take action as a whole. So my first company, which I started when I was 19, was trying to do this for local governments. So we were trying to crowdsource policy proposals and to show that the registered voters in a town thought that one idea or another idea was good. And for many reasons that didn't really work. We did end up getting acquired by AOL, which gave me some more space to keep thinking about the problem. But yeah, in a sense, this is a continuation of the problem that I've been working on for over a decade now.

AI assessment note: “Rome's a continuation of the thing that first got me into tech”

Q weeks at law school, so do not worry at all on not getting a diploma. I think you're a Far stealthier standards than me. I do want to start, though, because when we chatted before, you said Rome is not about taking pet and nose. I guess, what is it about then, and what end stage you really want to achieve from the kind of the product paradigm with Rome?

A Yeah, so the first step of it is about thinking about our thoughts, right? The way we think about it is writing is a tool for thinking, and if you can't get compound interest on your thoughts, if you can't build up a thought structure over time and continually go back and revisit it and revise it, you're limited to just what you can hold in your head. So one of the challenges I ran into when I was an undergraduate writing my thesis, which was Also about the software product that I was building was just trying to put together this, like, grand unifying theory that was pulling from a ton of different domains, and I just couldn't hold the problem in my head. So the first goal for Rome is we want to make it so that you can collaborate effectively with your past and future self, so that as you read things, as you have conversations, and you write those thoughts down, because the act of clearly articulating what you're thinking, it's hard to write things down in your own words. That's something that Rome is not going to be able to solve for you, right? When you read a book, most times people read a book, they forget Write down some of the core insights in your own words. Just that action alone is going to crystallize the thought, test your understanding of it, and help you remember it later. But if you can come across those thoughts and build on them later, if you can say this idea i…

AI assessment note: “writing is a tool for thinking, and if you can't get compound interest”

Q degree, I have to admit. But I do want to ask, you mentioned the product there, and we really delved in there. In terms of the team building the product, when we spoke before, you said something really interesting, and you said, I hire for grit and autodidacts. So how do you think about your hiring philosophy? And I guess, how do you really stress test those two in reality?

A So I spent a really, really long time founder dating for this company, because I knew this was my life's work. My first company, we had a small exit, That gave me some financial freedom to go live in a van and work on this and not have to worry about a paycheck for many years, which is great, but this is the project that I want to be doing for the rest of my life, and so I wanted to be really careful about who I brought onto the team and who was somebody who I wanted to make that sort of, like, lifelong commitment to, and actually, I was still in India at the time, and I remember hearing about FortyTwo, which was a tuition-free peer-to-peer coding university that was funded by a billionaire in France, and I saw Paul Graham tweet about it saying, you know, this isn't another coding boot camp, this is another MIT, and their filter was Processes. You know, there's like a two-hour long online test that you take that's basically just a logic puzzle. Sort of feels like an IQ test, but it's also a test to see will you keep going all the way through the two-hour time limit? Will you actually, like, keep doing this hard thing and have that dedicated focus for two hours? And once you get there, you get invited for a month-long application process, which in the batch that I was in had basically a 70% attrition rate because you're teaching yourself C. There's no teachers. There's just some…

AI assessment note: “you're teaching yourself C. There's no teachers. There's just some YouTube videos”

Q We mentioned those tools earlier, and I kind of find it clarifying often to think about kind of the binary statuses of bundling or unbundling. When you look at kind of the collaboration tool space today that we mentioned and that people, I think, probably wrongly put you in. Do you think we're in a state of bundling or unbundling when analyzing that market today?

A It's interesting because I've used and I've seen other people use all sorts of tools for organizing their thoughts. Like, they'll use their Kindle highlights, and then maybe they've got some dedicated software for reading PDFs and for building mind maps of the papers that they're reading, and then using something like Figma for laying out a bunch of thoughts, digital whiteboarding, all this stuff. But none of them have a paradigm of, like, reusing ideas in different contexts, taking the same data and showing it in different contexts. So I think sort of moo, like, not yes, not no, right? Like, in some ways, it's the wrong question. I think Notion's done a pretty great thing of saying, hey, actually, you want your notes and your tables together. You want your Airtable to be, and Airtable's also working on notes as well, but if the underlying paradigm of all of it is not one that's generic enough to fit the domains that you're thinking about, that's one of the reasons we're so bullish on graph databases, is that they're a better representation for the relationship between thoughts, the relationship between projects, the relationship between people, and yeah, most of the tools that we have out there are based on the SQL databases, and like tables and rows, that doesn't have the same kind of structure. So yeah, we're starting to see some rebundling But I want to live in a world wher…

AI assessment note: “So yeah, we're starting to see some rebundling”

Q It's an incredible story of kind of persistence. Before we dive into the founding story for Rome, I guess, how do you think about that then? And how do you advise founders on the balance between really sticking to your vision and mission versus acknowledging when something could have been wrestling? It is not working, giving up.

A I mean, I think the thing is you can have the long-term vision, but then you also need to tactically be looking at how can you see yourself really clearly? How can you see where are the places where the strategy you have right now? You actually don't expect that continuing on this strategy is going to succeed. So Yeah, figuring out how to maintain faith in yourself, and maintain faith in your ability to eventually solve the problems that are not solvable or tractable to today, but figure out how to make progress, and it's interesting, because I'm sort of a believer in the, I guess, the power of positive thinking, like, luck favors the prepared mind, and if you're going into situations believing that you will find an answer to them, you're just more prepared to see the opportunities that are there, so there's one thing in terms of psychologically priming yourself to find the opportunities, but you also have to be able to change strategy and not be like, Too committed to a particular path. So it's sort of like strategic focus and, like, long-term vision and grit, but then tactically being able to see that there might be many different paths you could go to reach that end goal. It's a tough balancing act, though.

AI assessment note: “strategic focus and, like, long-term vision and grit, but then tactically being able to”

Q I do love that as a story. It's an amazing to hear the relationship there, but it's also amazing because you do also choose to live with the team. So Mike told me about this. Why did you decide to do this? And what do you think of the benefits of living with the team?

A I mean, it's a family, right? I mean, for us, it's like, this is much more than a business, right? This is the future that we want to see. I want to build this graph of the world's knowledge so that my kids don't ever have to go to school. They can learn the things that they want to learn and find peers and find mentors and establish credentials based on the quality of their work. I think that maybe this is also part of the reason there's this sort of cult behind it is like Rome is a philosophy. It's a belief about how the world should be. So we try and put that kind of thing into practice and have an orientation towards truth seeking in our personal relationships and everything, but just having those like really close bonds and having all that time to know each other as full people. Now we're growing the team. We currently have one of our first hires. He was brave enough to brave the airports during the pandemic and come out here and stay here for six weeks, but we are going to be Primarily remote company, but it's still great to have people come out and say for six weeks or a month and just like really get immersed in the culture and for us to just be able to know somebody as a whole person and know like what are they wrestling with personally and just have those kind of bonds because I believe in we can do a lot with a pretty small core team and it's good to know people for …

AI assessment note: “just having those like really close bonds and having all that time to know each other”