B (3:01)

I just didn't really know like what was, what was the business like, what did it look like and how big was it going to be. It was never going to be a venture company. I never wanted it to be that. It wasn't built to be that way. I don't like managing people. I want everyone to get on with it themselves. So I almost made myself like a checklist of like, okay, if I'm going to work on anything again, these are the things that it needs to satisfy. What? And one of them was like, needs to be a one to many thing. So like a newsletter, a course type thing. Like all of those kind of feel like they're one to many. It's not a service based business. For example.

B (3:38)

Another thing on my list was like, don't get on a content treadmill. Like, okay, if you do this once one week, you've got to do it again the next week. And obviously we could talk about how I didn't do that because I started a newsletter.

A (3:50)

A daily newsletter.

B (3:51)

Yeah, exactly. And there's just things like, okay, I was having twins at the time.

B (3:59)

My work revolves around my life, not the other way around. So when I don't want to have to take calls ever, I don't want to do the like a bunch of things I didn't want to do. And then I had a list of like 20 ideas that I hadn't done anything with, which to me is a signal to myself that I'm not interested enough to do anything with it. Because you would normally in previous times had an idea, I would start working on that idea that day or like at least start a landing page or something because that's kind of how I did things.

B (4:30)

So yeah, I just haven't been ready to start a company. I don't know if I ever will be.

B (4:36)

I kind of shifted to. I Mean, Ben Spides technically is a company, but yeah, that's kind of how I felt.

A (4:44)

You know, no one talks about how difficult it is to start a company. We talk about how great it is. You get freedom and all that stuff, but it really is difficult. You're so uncertain all the time. You're constantly doing every single thing all the time. I loved in Ray Kroc's book how he talked about how they used to have to sweep the floor. Right. You go from being someone who can do like top work to then also being the janitor, literally to also being the person who fixes other people's screw ups. And it's so painful. I wonder though, I thought with no code things were becoming easier. Do you think at least now with like AI making it even easier to create software that all of this stuff goes away, all the pain?

B (5:24)

No, I think it probably gets worse. Why? Because in that example you just gave, which I completely agree with, it's a lot of context switching. There's lots of different tasks going on at different times and different priority levels and all the rest of it. Today if you look at developers now, everyone's talking about like having parallel agents running on your product and like you can spin up a swarm of 100 agents to do like 20 of them will do ads and 20 of them will do growth or whatever it is. Like people are thinking that way and until you've like had any employees whatsoever, you start thinking, how are you going to manage all of that? Like even just for yourself mentally, how can you visualize all the tasks? These people are working.

A (6:08)

I see. Meaning like it's one thing to have all these agents out there in the world and yeah, they're doing great work for you, but are they really doing great work? No. You have to go back in, you're saying, and check on things they're doing and then. Yeah, got it. So you don't have to deal with feelings and family issues. You do still have to deal with outcome and adjusting of directions and all that.

B (6:26)

Yeah. And I think it's just still, it's all those different contexts switching between all these different things all of the time. And I don't see how that.

B (6:36)

Is going to go away. Like it feels like it's actually you can work on even more stuff now. So actually people are going to get worse off feeling like they should be really productive, but actually feeling like I've done absolutely nothing this week except for I tried to do these million things but none of them got through or got over the line. In terms of like, that's the thing that is actually now on forever. And I love the output of that.

A (7:00)

Okay, and you're saying the AI agents are going to be another level of the same old thing of more work piling on top that you have to manage. Okay, that's fair. That is sad and it's scary and it's fair. And is this something that you're feeling now because you're kind of tired of the previous company? Or did you also see that with the MakerPad companies that were being built using no code?

B (7:22)

Yeah, I think it's kind of the same with MakerPad. The whole reason I started it was I couldn't code. I still can't code. Traditionally.

B (7:32)

Everyone told me, if you wanted to be CEO, run a tech company, you have to learn to code.

A (7:37)

Right?

B (7:38)

And I just, I tried. I did try. I tried multiple times, and I couldn't.

A (7:43)

Get it to work.

B (7:43)

Like, it just didn't gel with me. And I couldn't figure it out. And I was like, that's really. That can't be the only way that I can go and be a business person. Like a run my own company. And I wanted it to be in the tech world. I wanted it to be online. I felt like all the advantages come with you. Start a website, you can sell one to many. Like, obviously that is the preferred option. Then starting a shop in my local village or whatever.

B (8:08)

And so that was kind of the premise was, I can't code. I'm going to figure out how I can make tools work together to feel like real software. So I managed to put like a webflow site with Zapier on the back, like as the API connector. And then Google Sheets or Airtable is like the database.

A (8:28)

And then I was.

B (8:29)

Then I built. What I first did was build probably 50 things with that kind of stack of tools, hoping that one of those is going to be my business. That was like, this is why I'm a hotshot CEO now. It's because of one of these. But no one bought a single thing of any of those products.

B (8:47)

So I was like, okay, well, rethink this. But everyone's asking me how I built it if I can't code. I was like, okay, those are gone. Do teach. So I will start a company that teaches people how to do this. I can. Like, I think I saw there's a guy, I think his name's Chris, who started a company called Go Rails, which is a effectively a tutorial site for learning a framework, a coding framework, Rails, Ruby on Rails, I was like, oh, well. And I think he was on Indie hackers at the time, making like 50k, 15k a month. And I was like, that sounds great. I'll do that. And I'll just record how I build stuff with no code. So all my weird ideas that I have, I can still build them. Just record my screen and then sell access to it. Like, great. So that's how MakePad started. And then sort of fast forwarding through the journey of makepad. The thing I didn't realize until I don't know whether it was within the makepad life or after, it was like, I'm still teaching programming but in a different format. Like, there's different. The tools just look different. Like, they don't look like a code base or a code file or GitHub or anything like that. It looks like you've got issues. Okay, your Zapier connection didn't work and you got to debug that and actually got to fill in all these different forms, form fields in Zapier to like, do that and test it and then like, oh, let's run it again. Did that work? No, it didn't. Let's try a different thing.

A (10:10)

Do it again.

B (10:10)

Like, that's bug fixing. It's the same. It's the same kind of thing. And I think it helped me a lot to understand a bit about like, sort of system thinking on, what's the front end, what's the back end, what's the API layer, how do these pieces connect? But it wasn't the true. Like, my mum can now build an app if she wants to with no code, like I thought was gonna happen. But then, I mean, we can fast forward to today and thinking, my mum still won't build an app, even though you can use something like lovable, because I think it's still. You've got to want to do the building bit. Like, that's still the issue. It's either someone like my mother wants the app built, doesn't want to spend any time in the building of the app. So no code. You're spending time in that building phase fixing Zapier and Airtable and whatever else.

B (11:04)

So I think that's sort of happening today on a much quicker scale. So you can get something working quicker. But if, if you're having to type 1015 prompts to any tool to get to work, you have to either enjoy that process or have a very high threshold for going through that process to get to the end. That's kind of how it feels similar.

A (11:26)

And different Today you are now not at Zapier, where you sold your company. Are you using Zapier still?

B (11:36)

I am not.

A (11:38)

Because.

B (11:40)

Because I have a very high threshold for pain and I want to code it.

A (11:45)

So you are actually getting into coding then?

B (11:47)

Yes, I am.

A (11:49)

How do you square that circle that you are not a developer but at the same time you code?

B (11:54)

For myself, I would need to feel like I truly understand.

B (12:00)

All of the workflows that like the engineering team I work with do and like why they do them and why they're important and like why you shouldn't do this, but you should do this. That's like one that I'm like, okay, I can learn some of that stuff at some point, but I haven't got it all like learned yet.

B (12:19)

I couldn't sit there if you took the AI tools away to build anything either. Like, I couldn't sit down and write lines of code. I would have no idea where to start really.

B (12:30)

And I think on some level it's definitely helpful to be able to do that, then get the most out of these tools. But then I don't also think the future is coming faster than the present is like sticking around and I think you don't need to, like, you don't. I feel like the original thesis of makeup was right, you don't need to learn to code.

B (12:51)

But like there's asterisk. Asterisk is all the way through since then, which is like, but it might help or like it sure would help or like it might help now. So that's kind of how it feels is like I can, yeah, sure, I can get stuff done not at the level of a good top programmer, but more so than a non technical person. And like some. Somewhere, I'm like a technical, non technical person somewhere.

A (13:14)

Here's what I'm understanding about how you do it based on what you've told me before and what you're saying now. You've really created these AI tools for yourself that allows you to communicate with developers by having it translate what developers are doing and tell you in English and having you tell software in English what you want and have it converted to developer talk. And at the same time, all the work that you did with Zapier, which is basically these, I don't know how to describe Zapier. It's basically this. If, then this, then that type of experience that's very visual, it still taught you how to reason through getting what you want with all these different possibilities. If this, then that, and if that, then this and all that. And so you've learned the logic of coding and now you don't have to learn the language of it because AI is your translator.

B (14:05)

Fair? Yes, very fair.

A (14:08)

Okay. So that's actually the dream that we're all aiming for here with AI Now. Not all. Your mom is not aiming for this dream. You know what it is becoming easier and easier to do? Plumbing, for example. I still don't want to be a plumber. I still don't want to go and fix my own pipe. That's where people are. So a lot of people are not going to want to code. The dream of AI is going to make us into coders. All of us into coders. Not there. But for those of us who want to be, who have the desire and the understanding of the logic behind it, it's more acceptable. Fair?

B (14:37)

Yes, I think so. I think AI generally is making people more technically able to. And also making people more curious.

B (14:46)

I think in order to want to get anything out of life, I guess, is the curiosity is a key piece of that. So for me, I'm really curious about all of these systems. How to work. It doesn't feel like work to me. And building this system to understand the code base and the roadmap and our documentation, as much as it's to help me figure out answers for the community. It's also like me learning the product, like in and out. Like all these questions I probably need to ask once and then I know it. And then also learning how to build a system and like tweaking the system, get better at that system. Like, there's all learning opportunities for me. And I was not interested in learning until I could start building stuff. And that's when I was like, this doesn't feel like learning. So that's why I love doing it.

A (15:37)

Okay, let's look at one of the things you built and then we'll come back. Away from screen sharing.

A (15:42)

Here's how I would describe what you're about to show us. You've got a team of people who are all using LINEAR project management software to keep track of what they need to code.

A (15:53)

Into the factory software, right?

B (15:55)

Yeah.

A (15:55)

You don't want to go into linear because linear only tells you what they're working on and what they've done. And what you want to do is you want to also understand what did they say they're done, but is not fully done, what did they say is ready to be published, but they forgot to tell the project management software is actually part of our code base and it's published. And so you wanted to have this check and balance on the system, am I right?

B (16:17)

Yeah. I wanted a source of truth that was anything that's being worked on should be in linear and then anything that has been done should be in the code base. And then everything that is actually public and shipped and is like feature flag on, so everyone has access to it should be in the documentation. So all those three stages should be like anything we're thinking about all the way to everything that users can use and touch and feel today.

A (16:42)

Okay, let's take a look at it. And this is you using Factory's own software to manage all this. Show me the screen. Let's take a look.

B (16:50)

Yes.

A (16:50)

How would you describe Factory in one sentence, by the way, while you bring it up?

B (16:53)

It's the number one coding agent in the world.

B (16:57)

So it's kind of like a core code, but we get more out of the models than the model companies do themselves. And we're model agnostic, so you can use any of the models in our software. Okay, so this is Factory's Droid CLI tool. So this is in my Terminal, which is something we can touch back on, which is something I was scared of for a long time. But you can just think about it as like a chatgpt chat effectively. So I've asked a question, and this Droid exec is one of our products. So it's like a way to use Droid without.

A (17:35)

So you're using Terminal as like a chat experience. Highlight what you're talking about as you're describing. What are we looking at here, where it says linear at the top. And then there's a checklist of two items that have checks next to them and four items that have this push pin next to them. I'm assuming the push pins mean they're almost. They're not ready yet. They're almost there, yes.

B (17:54)

Correct. Yeah.

A (17:55)

And the check is. And so this is work that your team, the developers, have worked on, right?

B (17:59)

Yes, exactly. Yeah. So this, like the summary here is showing me what did I ask? And then gives me a quick summary of. This is what it is. Answer to my question. And it talks about. Okay. There are some enhancements that are tracked in linear that aren't yet shipped. Okay, helpful to know.

B (18:21)

And then here we'll show. These are the tickets specifically to your query. And like you can link to them and go and click and check them out. So for this. This one, for example, I looked at previously. So I set this up for this demo, but this one, for example. Yeah, Says ready for release. We want to check it in Linear. It's got a bunch of sub issues or subtasks within it that haven't been done. So the full thing isn't fully done. But whereas this ticket says it is ready. So there's like that. That discrepancy does happen.

A (18:57)

And the reason that happens is because somebody said in Linear, your project management software, it's done, but they didn't look to see that the subtasks still needed work.

B (19:06)

It could be that. It could be that actually for the top priority, we just needed to get like these four things out of the way and then these six things are to be done in Q4 or Q1.

A (19:17)

Okay.

B (19:17)

Things like that. So I think there's like developers have their own way of working within the developer team. They probably all got their different ways of working with Linear. I've never used it, so I'm coming in being like, okay, where is everything? Um, okay. And this is a. Really. For me, like I said, it was a sort. It's a source of truth for me. I don't need to. I'm in the uk, my team is in San Francisco. I don't want to ping all these questions on Slack that effectively sort of support, like questions.

A (19:46)

Yeah.

B (19:46)

Feel like I should be able to answer them and answer them well and understand them and understand our users of what they're actually asking, like me learning the product. So this will have a bunch of linear tickets mentioned here. It will then specify where in the code base any of the, like any of the sort of response it's giving me is linked. So like this is where this happens and this is where the session stuff happens, or this is where the streaming output happens. And then we have our docs as well, which then shows this is where we documented anything to do with this query that you've asked. And then it gives me a suggested reply, which I don't ever use. It's just I sort of added it to the template and it's just. And then this is just a way to show me what it did to check all the above information was correct. It looked through the different, like these different search queries across the different files and systems I've got. They reviewed the docs and the key implementation in our code base and then it queried Linear through another little tool which I built. And then it sort of combines everything into one place, but there's no actual code written for this. This is just one big instruction manual.

A (21:07)

That you've given Droid and Droid now Knows what to spit out for you and you've also told it on my computer in this folder is where you're going to find our code base.

B (21:15)

Yeah, exactly.

A (21:16)

And go search that. Okay, give me an example of something you were able to do here that you couldn't have done if you just spent a little time with linear and just learning linear instead of learning this.

B (21:26)

Well, I think it's important for things where stuff is on the roadmap, but.

B (21:33)

There might be some things within the code base that, okay, it's ready for this feature in the future, but this is something that maybe the team wanted to do a while back, but now we decided in Slack that we're not going to do it. So there might be tickets in Linear for that piece of work. Nothing in the code base or vice versa.

B (21:56)

There's very. There's varying levels of what this can do. And also just helps me even with things like billing. Like billing can be complicated with AI tools. There's like tokens, there's free trials, there's like credits, there's all these different terminology and like different ways to work with it. That even if you just asked how does our billing work from an internal point of view, like, how does the billing work? So I can have a really clear.

B (22:24)

Most of it is to have a really clear answer in my head or really clear, like visual in my head of, okay, I understand where that works. And actually now I know where it is in the code base that I can. Now, if I wanted to go and look at it and look at the code and then go and ask questions of the code, I can go and do that quite easily.

B (22:44)

So that's kind of. I don't know that I could get this any other way because I built this linear. So the way I've got linear in this is that there's a small tool that uses my API key and just searches linear. So it has like different search queries based on whatever I've sent the system.

B (23:03)

If I did the same thing in Linear, I might get some tickets.

B (23:09)

But then I don't know. Again, if one says ready for release, is that in the code base? Is that live? Is that live in production? Or is that on our sort of dev environment?

A (23:17)

So what it's doing is it's searching all these different places at once so you don't have to. Yeah, there's a brand new idea that somebody wants to do. You can come in here and it can search and say, you know what, we actually considered this before and we scrapped it, but we did some work in putting it in project management. We don't have to rethink how to do this. Here is the. Here is the origin of what we thought about before, and we can use that as a beginning for what we're doing next.

B (23:39)

And that's a perfect example for. I think, if. I think there's something that, like, oh, this would be a really cool feature. I would use this system to say, okay, firstly, research around, like, is there any scaffolding or anything else in place that would be crucial for this new feature that I want to add? Is there anything in linear already? Have we talked about it? Have we scrapped. It wasn't an approach before that we want to avoid all that kind of stuff. So it kind of gives me context across product management, the code base and what's been live, and then gives the system the context to, like, do a better job of writing the right code.

A (24:15)

Okay. All right. We can stop screen sharing unless. Is there something else you want to show?

B (24:20)

No, I don't know if you want to show the instruction manual. I don't know if it's.

A (24:26)

Yeah, let's see the instructions that lead to this.

B (24:28)

Okay.

A (24:29)

And what you're going to show us is how you're searching multiple platforms to get answers.

B (24:34)

This is my instruction manual, which is the only file that.

B (24:42)

Factory sees before it goes and does any of those searches. So here I basically got a number of examples. So, like an example question that someone's asked me and then an expected output. This is what it should look like. You should see the linear tickets. You should see what's done.

A (25:00)

Scroll up, let me see what that summary is. Let me see if I can read that a little bit more. A little bit more. Okay. Summary. Provide a one or two sentence answer that states what's shipped, what's in progress, and any caveats. Got it. And that's the beginning that you showed me earlier. So before you see all these details, you want to know, summarize for me what's going on. Number two thing that you ask it to give you is for each provider, include a subheading with the status and evidence. And then underneath that it says core support. And there's the checkboxes that we saw that. What you want to know is check linear, tell me what's done and show me what's in progress and update it for me in this format. And that format includes a link back to linear in case you ever want to click and see the ticket itself. Got it.

B (25:42)

Exactly. Same with the code base and same with the Docs like, just tell me if it's supported, if it's in progress, if it's not supported and linked to the things that are relevant, and then.

B (25:54)

Give me a suggested reply. Then there's further examples, but effectively there's.

B (26:04)

Instructions. Start here where it says, okay, do parallel searches across the different.

B (26:12)

Files and folders that I've told you that you should look at.

B (26:16)

Do an initial exploration, then deep dive into any key files. Then do targeted feature searches with the context. And then there's like a research strategy here, which is like cast a wide net. First focus on documentation, drill down with context, verify completeness, and then bring it all together for the output. So this is like, yeah, here we go. There's some more.

B (26:41)

Instructions here. So treat this example as like gospel. You should follow this exactly. Okay. Execute the tools in order. This might be something I'll change. But like, this is what's worked for me. Cross check findings, cross linear code and docs before drafting the output. Reuse suggested reply format for linear. Use this tool. So I built a little CLI tool that it can just call with this command that basically just uses my API key, looks up linear and does that. Droid obviously built that little tool for me. And then get your product data from my code base on my machine, and then also the documentation from the docs on my machine.

A (27:26)

How did. This is a really detailed document, really detailed recipe for how to take care of Ben when Ben wants to know what's going on at the company. How did you write this?

B (27:36)

I didn't tell me I was trying to do this work without having this already done.

B (27:44)

Just playing around with Droid back and forth. And at one point I got an output that I really liked and I was like, that's perfect. I really, really like that. So you can see all the different tool calls and things it did. So I kind of just copied all of that and then said to Droid, like, give me step by step instructions exactly on what you just performed to get to this output. I said, okay, great. Now put that into an agents MD file as instructions for another agent to understand and follow this exactly. To get the same output as I would with any input and kind of just do it that way. And that's how I build a lot of these systems, is that way.

A (28:26)

That's so interesting because when I asked Gary Tan, the head of Y Combinator, how he created his prompt for ChatGPT to create the perfect Gary Tan YouTube script, because he's big on YouTube, he essentially said the same thing. He kept Going back and forth and saying, look, here's the finished product. I want you to create something like this. Then once you've done that, I want you to give me a prompt that will get to that and then just keep giving it the prompt, asking for edits and asking at the end, recreate the prompt for me that got you to this end. And that's the way that he did it instead of writing the prompt himself.

B (28:59)

Yeah, I think AI is so good at getting you off the blank page problem and that I really struggle with the blank page. I'd always rather have something and with AI, you've always got something. And sometimes it's helpful and sometimes it's not, but it's always. I find it always helpful for me, whether it's. I hate all of this. At least I know that. Okay. I know when you have any of that in the end of what I'm working on.

B (29:26)

And I think it just. Yeah, it's just a way to. It's an agent itself. It's going to write in a way that it would understand. So, like, I can write these instructions, but maybe. And maybe some of these would be better off written differently. And I haven't tweaked it too much yet, but I realized at the bottom of this, it's something I did look at doing, but I haven't actually used it for anything yet, which is okay. If I want to make any tutorials or how to's on how to use the product, I should also have instructions on how to do that using the same tool that I've built. So this actually looks at. Okay, looked at, recently completed, like, linear issues. Because, like, is there a new feature that I want to cover and do? Like, as soon as it's launched, I'll do a video that day and then show people.

B (30:14)

And then. Yeah, writing documentation for it and all that kind of stuff. So this is something I forgot was here until right now and I should go and try and do some of these.

A (30:23)

How did you start getting comfortable with terminal? When you're a person who is comfortable with flowcharts of Zapier and make.

B (30:30)

I felt weird, like probably most people that like everything was now a chat, like everything was just okay. You just type in a question and get a response. And like sometimes depending on the task you're doing, it'll show you code, it'll show you not code, it will show all sorts of different things. Things. And then I came across a product called Warp, which was a terminal product but with like chat deeply embedded in it. So a terminal, if people don't know, is on your machine. So it comes with when you buy your new laptop. And it's a way to enter commands into your system. It can read your files, it can move files around, it can delete files. So be careful of that and do all sorts of things at the sort of machine level.

B (31:19)

I knew developers use the terminal and have experience there, and they know the commands. Like, you need to know very specifically what commands you're typing in. Otherwise the system just says, nope, doesn't exist, nope. And you're like, okay, there's literally no margin for error. Whereas with, like, ChatGPT, it felt like you could make a spelling mistake, it would still know what you mean. So with Warp, I felt like, I can type in, hey, can you create a folder in this? No. Can you create a file in this folder? It was like, okay, yeah, you meant whatever the commands are like mv, whatever. Like, it's already weird commands. So it would like translate that. And I was like, okay, so I can actually just talk to it like a normal person. It will talk to the computer like a computer, and I still get the best out of those things. And then CLAUDE code came along, which is a CLI tool again. And within that, you can just talk in natural language within your terminal and it will do all of these different weird and wonderful things. So that was definitely my sort of feeling like, I can just talk to this. I don't need to worry about what it's going to do. Over time, I've worried a bit more. Like, I can see it, like, deleting files now and I need to know what that the hell it's deleting. But you find yourself watching a lot of what's going on and like, picking up, okay, I know there's a command for that now. And sometimes I might use a command where I wouldn't before. I would have no idea what it was before. So they definitely got me comfortable. And actually, I now use the CLI tool over using ChatGPT or Claude on the web for almost anything.

A (32:55)

Give me an example of something that you're doing in Terminal that you couldn't have done without it.

B (33:00)

Well, it's kind of like, okay, well, I want to start a new project, so create a folder in my. In my system. And then I need four files in there. I need one which is the instruction file. I need one which is like, oh, just go and copy this thing from this PDF. Or like, get that PDF information and put it in there, actually rename that file and Actually that other one we don't need now so we can delete it. Like you can do all of that with clicking, dragging, sort of copy pasting, but like in a sentence or two. I could have that done and handled and like all done correctly and set up.

A (33:36)

What's a project that you would, that you needed to do in the early days that you were able to do because of this? What's one that made you go, yeah, this is amazing.

B (33:43)

I think one I tried to build was like an email tool where.

B (33:48)

AI search on email is awful. Whichever tool and like wrapper you use, I've tried a lot of them. It's, it's just, it doesn't work.

B (33:59)

And I was like, okay, well I'd love to build an email tool that is like AI powered somehow. Like I can just search all my emails. How do I do that? And I think I built it three or four times that never really worked. I would try replit or try something else that felt like an all in one builder or just tried like cursor and try and build the project. I'm like, I've, I just don't know enough of what I'm doing here.

B (34:25)

And then it sort of came back to like if I built a tool in order to call my email so like a CLI tool and then just have an instruction file that says when you're looking for emails for Ben, use this tool and then search in these kind of ways and this will be a better experience than everything else he's experienced.

B (34:46)

And I think it's because what I tend to do is build first and think about it later. And I kind of learned that way. Like I build ahead of myself where I shouldn't really be here, but like I'm here now. So what you, what are you going to pick up in the process?

B (35:00)

And I go back to like, this could be simpler. This feels like I'm in a. And maybe if I was a developer or better equipped I'd know how to navigate some of the harder pieces there. But I, I almost don't need to in a way that I've built some of these things which is just have a really clear instruction manual with an agentic tool. There's not much else you need in between those to be honest with you. Like you obviously need to call with an API your Gmail stuff. So you need to set up an app on Google Cloud, which is horrible in of itself, but other than that, that's kind of.

B (35:36)

They'Re the kind of systems I've ended up Doing and building myself.

A (35:40)

Okay, let's talk about now marketing. When you were doing MakerPad, you had to build a community in order to get paid members. How did you get enough people in your community that you could sustain the community, have them help each other? What did you do?

B (35:57)

My first response is like, I didn't really do that much because I didn't feel like I was. I was, I wasn't like planning, okay, I need to do this, because then this will do. This will have this follow, like this knock on effect. What happened with MakerPad was when I was building those tools that no one ever bought, people were asking me how I built them. So then I actually sent out a type form. They had like 800 people on my list, sent a Typeform and said, if, if you pay 15 or $50, maybe 15, I think it was, if you pay $15, I will make this tutorial site. And like 15 people paid. So I was like, okay, I've made money now.

B (36:32)

And I kind of, I take that approach with everything is like, don't. I don't know if this contradicts what I've. What I was just talking about, which is like, I build first. I kind of just jump into stuff and I think I'll figure it out along the way. And like, nothing's really that catastrophic or bad of an issue if, if you, if you need to revert something.

B (36:53)

So I just started recording what I did, how I was building certain things and putting it on Twitter. And just a ton of people wanted to like, learn more. And it was an easy way to just like plug a new tutorial that I made. And then there was a slack group. And the thing I, when anyone asks, asks me for advice on Community, the number one, I think, number one thing I say is you're gonna lead by example. Like, if you're not in there talking, no one else is gonna be in there talking. If you're not in there asking seemingly stupid questions, other people aren't gonna feel like they can ask those questions either. And this kind of came back to when I worked at Product Hunt, where like, we'd. We'd want every single post to have a good comment thread underneath it. But, like, people, it might be a really technical product and like, oh, I don't know what to ask about that. And I would just go in and say, cool logo. Where. How do you make it? I'll just like try and break the ice. Like, start the conversation going, I saw that. This doesn't need to really. You're on everything. Yeah. I know it's kind of my job to be, but I, I didn't really think of it like I'm gaming anything but looking back it looks like, oh, actually that's a really good way to like kind of game that thing. And I just think that's it's important. Like I think the leading by example make another feel comfortable that other people feel comfortable enough that they can ask their question no matter how technical or not. Like it's bringing accessibility of the conversation to anybody. Which it's just something I've done since the product and days and then has carried through to like today. Like I still do it now. Like I know people will come after me for not being technical, but I'm like, yeah, but like I can do a lot of stuff that non technical people definitely can't do and I can understand a ton about the developer stack and all the rest of it.

A (38:39)

If I'm understanding you right. What you said was first you built up a mailing list of people who are following along as you were, as you were doing no code development. Then you said to them, I would if you pay me $15, I'll create a tutorial on how to build using no code. Fifteen people paid, by the way, that doesn't seem like even a lot, but it was enough for you to say, okay, at least somebody loves me, somebody wants this. You went and you created something, you sold it to them. Then how did you get people into a Slack community?

B (39:06)

I probably just put the button to open Slack on webflow as like when you paid, when you joined the membership or when you paid, it was like hit oh and welcome email. Here's your invite to Slack.

A (39:17)

Oh, that's all got it.

B (39:18)

Yeah.

A (39:18)

So all the paid people were then in Slack and then what you did was you kept that Slack active by asking questions, sometimes basic questions, answering questions, engaging with people. And then so then from there, how did you go, how did you build up the Slack community?

B (39:31)

Slack.

A (39:32)

It was only for paid, right?

B (39:33)

Yes. Yeah, it was. Well it just becomes.

B (39:38)

It just becomes its own thing. If you do, if you get community rights, they do do the selling for you and like they spread the word for you. The problem is people start thinking, well we'll set up communities so that we can do that, so that that end piece is what we're setting up the community for. Not truly for like hey, I want to talk to people who are building in this way or like wanting to learn this stuff. Like I was interested, still didn't know all the things that I Didn't know then it was like a way for me to learn. Like, oh, what do you want to see? Like, what should I try and build? Like, I'm. This is all new. And like, I'm trying to push the limits of what no code can do. So, like, give me ideas. I'm going to run out at some point. And I think it's just, yeah, you've got to build an authentic community first. I think too many people, especially for the over the last few years, have been trying to build community for it to be your sales, like viral moment to other people.

A (40:32)

Okay, how did you get more people to. To join your email list? To join and buy and join the community?

B (40:38)

It was just consistently putting tutorials on Twitter. That's all I did.

A (40:42)

Twitter. It was simple tutorials on Twitter. People would join to what, buy to get the full tutorial. I don't remember that.

B (40:49)

Sometimes. Yeah, sometimes they would be just some free tutorials, just like get people in. And then sometimes it would be, I would do like a quick trailer almost of like, this is what I built and like full tutorials here below. And then. Okay, that'll be paid.

A (41:04)

Yeah, okay. And then I know from the. Every interview that you did, you also used to go and under like Sundar Pichai's post, you quickly commented. So we all would see you in there. And was that around this time?

B (41:17)

No, that. That was for Ben's bite. So that was the newsletter I created when at Zapier. And again, at the time I didn't think I'm doing some like, big gamified thing. Obviously looking at it, it looks like that now, definitely. And like, I probably knew somewhere that like, oh, this would be a good way to hack, getting visibility. So I would basically go, anytime Sam Altman or Sundar would post a big new AI thing, I'd be the first one to post under and say, awesome, I'll put this in my newsletter tomorrow. And then that just like seeds, a bunch of people signing up for my newsletter. But my newsletter was the first, like the first and only AI Daily newsletter @ the time other people started newsletters and also started doing the same thing. And then I was like, oh, I don't want to do this anymore. This is now looking like a Twitter growth hack. I didn't mean. But I thought about it recently when I was. And I was talking to someone, I think last week about it, where I think that comes from the product and days where if someone had built an interesting product and shared it on Twitter, my job was to say hey, you should post this on Product Hunt. And I didn't really connect it at the time of like, actually that was a growth hack thing for Product Hunt. Like, I wasn't just reaching out to the community, it was like tagging Product Hunt, linking it. I was like doing all of that. For anyone else who was following.

A (42:29)

These product launches, here's the part that I used to see. I would go on YouTube looking for a tutorial on any kind of no code software and I would see a video from MakerPad. But it wasn't you creating it who did.

B (42:42)

That. So I hired a few people from the community to do this. So I, I got to a point where I was like, okay, I need to be like proper business person now. I need to like hire people and like get out of the day to day of it of like creating the content. And to make this big, I feel like I need a few people who can make 10 times more tutorials than I can. And I'm struggling to like think about an idea, sit down, record it, edit it, all that kind of stuff that comes with it. And I put out a call to the community and said, hey, is anyone loving note code stuff and like building on their own and like, would you be interested in doing any tutorials? And then there was a guy called Tom and then a girl called Amy who reached out and I said to Tom, have you done any tutorials before? And he's like, no, I don't. I said, I think he said no. And I said, okay, we'll just do a loom of like you walking me through a build that you did. And he did that. And I say, great, you've done a tutorial. Now we're going to post this. Like that's, that's, that's the job. Like this that you've done it in that way. I think just making it feel like I'm talking to one person and then describing what I'm doing and how I'm doing it. It's kind of all the job of that. Even though I couldn't convince myself to continue doing it, it was like that's how I sort of started that. But it was, it was always people in the community that we hired. Like, I never, I think I very rarely hide outside of the community, which helped because they knew what the mission was, what we're trying to do, who we're trying to teach, who the community looked.

A (44:09)

Like. They know the aesthetic, they know the vibe. Did you give them any guides, guidelines to doing.

B (44:15)

It? I think at some point we Must have had some star guides of like, okay, this is kind of how the. How it should. How it should work. But I think Amy and Tom both had their flavor of tutorial. I didn't really want to change it too much. I think at one point we did have.

B (44:33)

Quite a structured. Okay, this is the intro now, and this is the outro. This has got to be how we do it, because this is what businesses do. This is what it looks like when it's professional. And I've actually gone back in my thinking now where I think any video that is just like a loom style with a floating head, that's all you need. Like, that's kind of all you need for.

A (44:50)

Anything. I don't know. I really liked that there was consistency in it, especially if you're not part of it. I like the consistency, and it made me feel. Feel like it was a bigger company than it turns out it.

B (45:01)

Was. Yeah, I think that. Yeah. And that's kind of one of the reasons we probably did it was like, let's look professional. Let's look like we are a bigger.

A (45:09)

Company.

B (45:12)

Yeah. But I mean, even thinking about just like, for YouTube, I mean, YouTube's a different game that I never wanted to play, and I never really was good at thinking about how to, like, get a load of views for that. It was more. I want to share something organically. The best place to do that is Twitter. For me, everything in my life has happened through Twitter professionally anyway. And I think the best videos on Twitter are the ones that are just, hey, I've just clicked record. I'm just doing this thing. I've just posted it and that's kind of it. And like, I like that. That feels like that's a community member talking to a community member. It's not like this company presenting to you this video that we've spent hours.

A (45:52)

Doing.

A (45:54)

Okay, I got it. Let me just close out with a couple of obnoxious questions. I'll start with this one. You said you didn't need to work after the sale. What did you sell the company.

B (46:01)

For? I am not allowed to say, which is frustrating because I wanted, like, that, as the data point, to show you didn't have to learn to code to get rich and, like, do this stuff. So.

A (46:13)

That'S. You literally would not have to work right.

B (46:15)

Now. And you do now. I spent a lot of money, so I. I do need to work now. No, I think in seriousness, I. I could have made it work, but I think it was the.

B (46:29)

Figuring out what made myself work like What I want to be doing and like, could.

A (46:33)

I. Millions we're talking about.

B (46:35)

Then.

B (46:37)

I can't.

A (46:38)

Say. And then when you say you spend a lot of money on what.

B (46:42)

House, cars, holidays, all that kind of stuff, I mean, obviously I invested a bunch of it. I haven't really spent all my money. It's more that, like, there's some money tied up in shares.

B (46:56)

And I think after thinking for two years I wasn't going to work again in a normal, like a typical capacity that I then realized that I only care about working. Like, I actually need to work. There's a creative part of me that like, is essential. Otherwise I'm just sitting here and like, wasting away. Like, wasting my.

A (47:15)

Life. You spend two years doing.

B (47:16)

Nothing. I spent two years vesting shares, working as Apia and really struggling with what am I doing? What am I going to do next? Like, what has.

A (47:27)

Happened? Like, I did miss your voice then. Honestly, I feel like you were so present and then it felt like you were less.

B (47:33)

Present. Yeah, I just died.

A (47:35)

On. Yeah, yeah, it almost. It was like in top secret mode, which was. Which I get now because I don't know what you're like, what should your role have been? Should you have been the community manager for Zapier? No, they've got people for that. Should you have been the guy who's cheerleading this Maker Pad? I guess, maybe. But MakerPad became a different thing.

A (47:57)

Okay. I think I'm seeing then where this goes and then the way that you've discovered Factory, where you're working now, is you have a Ben's Bytes fund that invested in it. How did you find companies to invest.

B (48:12)

In? Lots of people have been readers of the newsletter have reached out. I've helped them with launches, I've given them advice or something along the way, and then they reach out and say, oh, by the way, I'm working on this new product with Matan, who's the CEO of Factory. I think one of my friends actually introduced me. So someone who also was in the no code space, who also has a fund. I think he introduced me actually through email. And there's, yeah, a few other people I. I made friends with. So every. Everyone in my network, I've sort of made through the varying places I've worked. And yeah, that kind of. I got introduced to Matan in October 23, I believe. So it was the first investment of the. Out of the fund. Okay. All.

A (48:58)

Right. Can we say that you're raising another one or should I edit that out? Because we're not allowed to talk about.

B (49:03)

It. I think you can talk about it. Yeah, I think it's.

A (49:06)

Fine. I think you can talk about.

B (49:07)

It. All.

A (49:08)

Right. Congratulations on the fun. Congratulations on getting more comfortable with coding yourself. And I really appreciate that you can be open about how, look, I'm not using Zapier. I've evolved into this other thing. I think there's a place for.

B (49:19)

Both. Yeah, I think what I really want to do now and what I feel why I've jumped into a job is I'm not old and just wasting away. Like, I'm not, I'm not done. Like, I really feel like I'm not done. Like, this is the most interesting time in technology ever in this space. In like this space. Especially given my background and like where I am on the non technical to technical scale. There's no better, like meaty challenge that I could jump into and feel like that'd be a really, really good opportunity for me to be in. And I just want to. I truly believe that the right message all along was you don't need to learn to code. But there's a new way of like what things look like now and we can figure this out together. I'm figuring it out. I'm trying to help figure it out myself so I can teach other people how to figure it out themselves and like sort of run back some makepads pastime stuff that I can, I can share and do some video. I did, I actually did a video a couple of weeks ago on a coded thing that felt like, oh, this feels like however many years ago, six years ago now. So yeah, it was awesome actually to do that. So yeah, I just feel excited. And one, like if I believe non technical people should be using tools like this to do work and there's different way to do that. Who's gonna tell him? Like, I'm willing to stand up for.

A (50:42)

That. Who should I interview? Who's doing this? Who actually is using AI to create this one person operation that is standing up and being a model for others. Who do you know who's doing.

B (50:52)

That? We hear this all the time. Is going to be one person, $10 billion company. Like, okay, well, it's boring by yourself. Like people want to have other people around them. I think it's natural to, well, just. I could just use some help with just this thing. So I think.

B (51:09)

See, we'll see how that truly plays out and when. And if it does, you know.

A (51:13)

What, Even if it's not exactly where we're going to end up. One person company. I think is a direction for what's power, what's possible for individuals. It's absolutely there. Even in the early days of the Internet, I interviewed one of my first. One of the first entrepreneurs who I interviewed was a woman named Rosalind Resnick. And I think she had had this vision that everything was going to work on the Internet. And we're talking about like the year in the 90s. And she said, no one's going to need a salesperson. I'll just put the buy button on the web and no one's going to need this and I'll just. And she just thought it would all be like that. She ended up hiring salespeople because she realized that if you really want the big deals from companies like IBM, which she got, you need a human being to talk to them. It doesn't matter that she was wrong about the detail of whether you need a landing page or a salesperson. She was directionally right in the sense that the Internet was capable of doing more than people thought. And if her vision was the Internet will do everything, including closing sales, but she didn't completely, completely get there. She got a lot closer because she kept that North Star. And I think if we don't end up with one person, billion dollar companies, but we end up in a direction where one person feels much more empowered. And the business is not this big operation of humans, but it's this big operation, maybe agents, it's.

B (52:19)

Okay. I agree. I think actually that feels very similar to like the American dream, which is why I think Americans typically have, yes, like a lot more optimism than people in the uk, for example, who are just a bit more cynical and less like. Obviously there's, there's many people here who think like me, but it's, there's something to go for. Like if you go to Silicon Valley, San Francisco and spend some time there, you can think, you can come away with kind of two responses. I think one is, God, these all these guys have got their, like, heads in the sky and like they're drinking all the Kool Aid and this, like they're away with the fairies. Or you can come back and thinking, well, this is where all the biggest companies in the world are built. So maybe like, yeah, I take a bit of this optimism and forward thinking and try and be like that. And yes, there's obviously going to be grifters and opportunists in the, in the, in the mix of it, but it's no different than anywhere. But I think it's the true. Why think really negatively and poorly on it when you can choose not to.

A (53:19)

Yeah. All right, thanks for doing this interview. I'm going to keep doing interviews and understand this space and I want to see if I could be a part of helping more businesses get closer to that one person, billion dollar business. In fact, I don't necessarily even. I'm not driven by that. I need a different mantra that's more mine. But I do love the business side of business and I think that I want to hear from more people who are building the business side of AI, who are not just saying, look, here's how I develop something, but here's how I develop something and here's the growth hack that got me to sell it to customers and understand their needs and build a business around it and all that's.

B (53:51)

Fun. Yeah, I agree. And that's like kind of what I want to do in my role as devrat is like, how can I build tools to help me do my job as a one person operation rather than there could be 10 people on all parts of this.

B (54:05)

All right, thank.

A (54:06)

You. Thanks. Bye, everyone.