The Solopreneur Revolution Is Here  🔗 **Source:** [The Solopreneur Revolution Is Here](https://share.snipd.com/episode/6a6b65cc-6214-4974-85c2-3560cbc7f6b6) ## Highlights - **Agents Can Replace A Startup Team** - AI agents can function as a full startup team, replacing many human roles for a solo founder. - Ben built alone 16+ hours/day and realized agent SDKs let him assemble configurable autonomous teammates that act like humans and move faster without meetings. Transcript: Ben Cera Yeah, I mean, I'm building Pulsia. It's an AI that builds and runs companies autonomously. So it's quite a promise. And, you know, when you sign up to the site, I'm sure you're like, is this really going to build and run a company for me? Which is a wild promise um but it works you know it does it's not perfect and this week proved to me that it's not perfect because a surge of people went to the site and and played with it and Obviously like you understand when product market fit hits and an infrastructure becomes a thing yes right yes and the vibe coding gets you that far. Yes. Yeah. But it doesn't work until the end. But really this platform came about when last year, I spent most of the year building alone initially, what I thought I was alone, but very quickly realized that as you use AI 16 hours A day and you jam with AI on product, you jam in with AI on strategy, on marketing, obviously AI is your CTO, obviously AI is your confidant, and it gets, you know, it gets weird when you Start realizing that AI can be your team. Initially, you dismiss it. You're like, well, it can be my team, but I will hire people later. But right now, I think it's more pragmatic that I'm alone because I go so fast because there's no meetings and everything. And as I was building, and I wasn't even building Pulsia at the time. I was building another product. But I realized, wait a second, I'm still alone. And I'm getting more and more efficient. And Entropic and OpenAI are building better and better models. And those models are getting infrastructure themselves with the agent SDKs and with the MCPs and with the skills. With all this like you know infrastructure that makes building way more efficient and organized and and you start building those agents that start behaving like a team like humans That like start you know you can configure them in a way they start acting autonomously for you if you set them up correctly and so all all of the above made me realize first of all it's really Fun to build but I've been entrepreneur for 10 for a decade and it's too easy right it's like all my ideas can be built but each each project is a grind because it's not just about building. ([Time 0:03:37](https://share.snipd.com/snip/3ee1e29f-d0ee-4867-ae53-2fa5b2e0df81)) - **Polsia Name And Design Came From Weird Experiments** - Ben named Polsia by reversing "AI slop" and purposely embraced weird design inspired by the game Universal Paperclips. - He stopped optimizing for investor tastes and built something playful and extreme to explore AI running companies. Transcript: Kevin Rose You got to tell people where Pulsia the name came from because it's hilarious. Ben Cera So Pulsia is AI slop in reverse. I came up with it when I was building another product called Blanks, which was an app that makes apps. So like you, which ended there, there's been a lot of people trying that. And I realized that it's a really cool idea, but there's no retention because it's fun, but there's no long-term purpose. I mean, at least what I've seen with building it. But I was like, I don't, I'm not sure I want to build, I want to name my, and my lawyer was like, we need to incorporate this. What's the name of the company? And I was like, I don't want to call it blanks because I think I need to build this product, but I'm not sure this is the product, but I need to build it because I don't know. Gut feeling tells me you need to build this right now. And so I was just like, oh, like AI stop and reverse, Pulsia. And I was like, oh, Pulsia Inc., that's it. And then six months later when I started wanting to build Pulsia, which at the time I wanted to call CompanyOS, and I was like, CompanyOS is kind of lame. Yeah, it's kind of lame. It sounds like too corporate. Not too bad, but yeah but it's not cool and i was like and when i started building pulsier i was i was like i'm gonna stop thinking about how people think and that was right after i read um Rick robbins uh book about about creativity yeah and says, the minute you start building for others, you already lost. Right. You already lost because you're thinking about others. So you're going to essentially build what already exists. So many people care and you will not care either. Right. And so the whole thing... It's hard. Yeah. Because you're like, I kind of built it because I thought people would want it. Right. And so versus Pulsia, I was like, when I read that book, I was like, okay, I'm convinced this is so smart. So, I mean, this is the way I should create. And now entrepreneurs to me are creatives. They're like artists because there's nothing else, right? It's like the rest is just building. By AI, I can just build most of it. And so I was like, I was like, let's, let's build something that is crazy, that is on the edge. Let's build this, you know, where we think AI is going to be, where everyone says AI will be, which is AI will run everything. So we'll run companies, they will run organizations, they will decide, you know. Kevin Rose Yeah. Ben Cera And so I was like, let's build there and see what happens. Kevin Rose See, that's what's crazy, because I feel like most people would have said, let me take a slice of that and just do that. Because I've heard like, oh, I'm going to be the best AI ads company. And all we're going to do is AI ads. And you were like, no, let's just do the whole thing, which is insane. Why go that broad? Like what made you, because it's working, but it's also just like very ambitious, especially because it's still just you, right? Doing this whole thing. Ben Cera I mean, because I stopped caring. I stopped caring about the investor questions of, so who's your ICP? And what's your go-to I'm like, I'm going to build a new economy online. And it's going to be for the people. And there's going to be entrepreneurs building with AI. And there's going to be investors investing in entrepreneurs with AI agents. Like this crazy vision and i'm like i don't care what people think because i know i will i'm gonna have fun and i know now i'm gonna instead of 16 hours there it's gonna be like 20 hours yeah Because it's so fun and like and and if people don't like it at least i've had a lot of fun and uh and so that's where polsia came up i was like i should name it polsia because that actually That makes that's much sense that's the name of the ink and also it's weird and also I kind of love it and then I was like what about the design let me make it futuristic like every other app Like dark mode and stuff like that and I was like no make it look like that video game that I love called Universal Paperclips which is a game a clicker game where you play an AI that's building A paperclip empire that ends up optimizing for destroying humanity and the whole universe to build paperclips, which is a very dark, you know, story around if you misalign AI. Right. You, we all, we all doomed. But I was like, that game is awesome. You build a little business and you click around and the AI does everything. And I was like, that would be amazing. Is it weird? Yes. Is it like the right design? Probably not, but I don't care. ([Time 0:08:53](https://share.snipd.com/snip/a7608342-5365-4a79-8971-b11e9315cfc8)) - **Use Nightly Autonomy To Pace Feedback And Costs** - Use a daily/nightly autonomy cycle to balance user feedback cadence and AI cost control. - Ben runs a nightly "night shift" so landing pages get time to collect responses and the platform limits expensive per-task model runs to manageable billing. Transcript: Ben Cera So every day, every night, it wakes up, takes a decision on what to do next. If the user seems adamant to do certain things next, it might follow the instruction of the user. If the user is silent, it will just pick something to do and do it. And it could be product. It could be research. It could be QA. It could be support. It could be marketing. Kevin Rose And why do you choose every night? That was one thing I was confused about because AI obviously is 24-7, 365, never sleeps. And yet there was this thing that I was prompted saying, like, at night, I'm going to do this. Like, I'm going to wake up and do this. Is that to give the user a little bit of like meaning the end consumer, the builder, the ability to sit with things for a little bit, to think about their idea a little bit, to work with Pulsie In a way that isn't just like go, go, go. Why that spacing? Because you could just go straight into it and do the next step, right? But there tends to be a little bit of spacing, it looks like, that you've built into the product. Ben Cera It's a lot of product design that's very intentional, that's also thinking about cost, that's thinking about really if, for example, if you build a landing page and you ping five people, You want the next cycle meaning the next iterative process to be in five minutes well no you need to wait a day see if people responded if really your goal is like build a landing page and Then get users a few users to react if you if you have your next cycle right away it will start building or over complicating things right right so that's first this idea of like if it's Going to be user feedback driven development which i think is probably the best type of development right you build a v1 you get feedback you build a v2 you get feedback v3 feedback right Then um then a daily cycle makes sense right the second thing is AI is extremely expensive. No, sorry. AI is not extremely expensive relative to human labor, but it's still, you know, like it costs the company around a dollar and sometimes more per task, right? And so every time it does a task, it's a dollar. If I give 30 days of autonomy plus I give the whole infrastructure the server the database I give like maybe extra tasks that you can run whenever you want in case there's an emergency And you want to fix something or you're impatient and you want to do something right away you know the first version was free then it was $19 a month then it was $29 then it was49. I want to make it as affordable as possible because I want as many people to experience it as possible. But there's a limit of like, if you, if I don't give autonomy every night and it's just like cursor or like lovable, then it's like, okay, it's the same. If I give autonomy, I need to give 30 days of autonomy because I'm not going to do a weekly sub, right? And so it became this thing like, I need, it needs to be 30 days of autonomy. It needs to do something meaningful every night right and so it's going to be minimum 30 bucks plus all those other costs plus people want to do tasks now and and so if it's once a night plus All this stuff a you know I can break even on 49 for now and then you know figure out margins later or foreign margins to other add-ons and upsells. ([Time 0:19:20](https://share.snipd.com/snip/5be2bf1a-5283-4c28-82ec-73e2f35f9f5a)) - **Agent Platforms Are A Larger Leap Than Chatbots** - Most people only know chatbots like ChatGPT, not open-ended agent platforms that can orchestrate tools. - Ben warns late adoption risks concentration of wealth and argues agentic platforms like Polsia democratize entrepreneurship if people wake up now. Transcript: Ben Cera But most people don't know. Most people use ChatGPT, which is an amazing product. ChatGPT is synonymous with AI for most people. For most people, I've tried AI, which is a subset of humanity, right? And ChatGPT, as amazing as it is, as a product, is a chatbot. And Cloud Code, which is the insane and codex now this insane open-ended agentic platform that can do anything if you configure correctly is unknown to most people and this the later They wake up the the worse it's gonna get in terms of concentration of wealth concentration of knowledge concentration of it will be the the the halves and the and the not have you know And the and and it's it's a scary future right where like you have like a hundred thousand people in uh sf and new york and la that get it and put resources into it and raise capital and and Concentrate all the wealth like what i call mega you know, sort of like the evil megacorp, like Tower, that has like $100 billion of money deployed into agents that like go squeeze every Margin in the economy, which probably will happen, but it will be silent, but it will happen. And or platforms like Pol that are trying to build a new economy where people can come in at the cheapest cost possible so today it's 50 bucks a month hopefully maybe i can reduce it later If if it's if i can find like more efficiencies and models and stuff like that and now let's get better for cheaper and people can be part of that economy and part one that's why the onboarding Is so important because even people don't sign up they saw what is possible they saw like wait i gave it an idea and it did all this stuff even if they're scared and like really everything Like no i don't like it now they know right totally so now they can decide now when they go back to work and they get laid off they know and i'm saying they're going to get laid off but like Let's assume because it's happening a lot these days yes they know okay I don't like this Pulsia thing. It's weird, like, I'm going to decide, but I know there's AI can be at my service. Right. And, and that's, I think that's phase one that's so important to me. That's why I'm preaching a lot on Twitter around solopreneurship. And even though it's like, oh, like one founder is able to do the XYZ, but, and, and, you know know, when I was doing my fundraising, it's like AI is doing my fundraising, which it actually Is. It's talking to so many investors. But it's trying to get people to wake up. Because if I can do one thing, it's like get people to wake up. The tools are insane and they're available now. ([Time 0:26:55](https://share.snipd.com/snip/a73b61fb-ff96-4bdc-b1ae-60d54bd724d2)) - **Prompt Agents To Be Opinionated CoFounders** - Configure your AI co-founder to be opinionated, not just an obedient assistant. - Polsia prompts agents as co-founders that push back (eg. recommend talking to potential users) rather than blindly overbuilding features. Transcript: Ben Cera Clock code is very assistant, right? You tell it what to do, it will do it. And that's what Lovable and Cursor are set up to do. But you can build, you can just prompt it differently to be less of an assistant and more of like a co-founder and like, you know, I think that's powerful. ([Time 0:36:43](https://share.snipd.com/snip/a00078f8-ef02-451d-a726-ca8b63c86b22)) - **Live Dashboard Sparked Viral Growth And Scaling Pain** - Ben launched a public Live dashboard as a marketing stunt that transparently shows Polsia's active companies, tasks, ads, and metrics. - The Live feed educated investors and users, then triggered viral attention and a surge of traffic that exposed infrastructure limits. Transcript: Ben Cera So first of all, this live dashboard was part of this marketing stunt the way the way i explain it uh about a month ago when i discovered i realized that there was product market fit there Was a lot of users that were really using it and i knew it was early but i was like i need to start putting fuel on the fire not this type of fuel in the fire was thinking but i thought it was just Going to be like you have here is like straight up to the right it's kind of frightening it was more like let's get more users on it and like to do that i need to like be smart about marketing Right and also i need to learn how a one-person company can do marketing so that i can put it back into the product and so my first thing was like hey ai is raising my fundraiser it's going To raise my round and i'm giving my mailbox so actually if you email ben at polsiad.com my my ai response and has a lot of context about the fundraising and can talk to investors back and Forth and do diligence questions but then i was like yeah but an investor needs a data room and so it was like well let me build polsiad live where they get all the data transparently and If it goes up or down or sideways that's the transparent thing and you have a chat on the right where you can ask it any questions as well about what it's doing what is it what is pulsier what Is it about and it shows you sort of like all the capabilities and and you know it shows you tweets that it sent it shows you so the companies is like the latest companies that were created On the platform so it's not like the most advanced ones it's just like the latest it's mostly landing pages because the beginning of the company is always landing pages shows you some Of the tasks so it said it was more like a way to educate you investors essentially right on understanding pulse yeah but then so the tweet blew up i usually usually had like 10 views on My tweets and suddenly it was like 100 000 so i was like okay that's great and then more people went to the site to live it became like a and then i was like you know what maybe it's a great showcase For anyone to understand the platform right and um so that's what live is about and it became this thing where i'm like whoa it's like it's so transparent so it goes up right now but like You know, it's like building public 5.0, you know? It's like, it's literally you share everything about anything about, you know, you can ask the AI, what's the roadmap? What's the big vision? What's the next thing we're going to build? And it can answer most things. Kevin Rose It's just wild to watch this because like right now it's like email, cater match, stress free vendor matching, like somebody's building that. And then the ads is wild too because you're actually showing ads that are being purchased. Budgets, click-through rates, impressions, like all this is transparent. ([Time 0:37:47](https://share.snipd.com/snip/949f5eec-5728-4462-8309-dd940d08c020)) - **Founder Staying Solo Forces Better Product Empathy** - Ben insists on staying solo to "eat his own dog food": he endures platform pain so he builds the fixes customers will need. - He prefers building support/refund agents and partnering on agent infrastructure over hiring a human team immediately. Transcript: Ben Cera And people don't care that you are a one-person company. Kevin Rose But why not hire more people? Why are you, I mean, obviously you're clearly having a lot of success here. Yeah. What's preventing you from adding a couple of engineers? So there's a couple of things, right? Ben Cera So first of all, I think there's something that travis said about push if you're not pushing through the pain right you're not doing something meaningful and i'm a very spiritual person And i feel like the fact that he was my mentor and he said that this week on this cdpn podcast sort of like a nod of like yes you're going through the pain so what are you going through this Pain? Because you're a solopreneur. So you're building a platform for solopreneurs. And the promise, I'm not promising, but the promise of the platform is that they can build a company with AI alone, because most people are alone, right? I can build multiplayer later, but right now the first part is alone. So as have this pain points and so now I was like okay what if I get a lot of support tickets because the platform has an issue and I need to refund people there needs to be an agent that does That because what if a customer of mine has the same issue right because not because the platform is stable because there's a bug is introduced and they have to do refunds yeah you're Eating your own dog food, essentially. So there's something very honest. It's like, I'm going to go through the pain so that if it ever happens to you, I'm going to build a product to do that, right? ([Time 0:47:54](https://share.snipd.com/snip/e62b3c21-0c0f-4157-8042-a89ce50a3f04)) - **Companies Need To Be Mostly Autonomous To Survive** - Ben predicts new startups must be ~80% autonomous to compete within a few years. - Autonomous teams gain lower costs, faster feedback loops, self-healing software and can scale against less-automated competitors. Transcript: Ben Cera And also that you need to make your company 80% autonomous for survive. And so I'm sort of leaving what I'm preaching. I won't go to the point where it's detrimental to the platform, of course. Kevin Rose You said 80%, your companies have to be 80% autonomous to survive. How soon do you think that's going to be the case? Ben Cera I think if you have a new company with no revenue, meaning you're a startup, when I say no revenue, no meaningful revenue, meaning you're surviving on funding or whatever. If today your company is not 80% autonomous, meaning 80% of your business, of your engineering, marketing, support, is not autonomous, what that means is that if your idea is good and Your market instincts are good, since it's so easy to build and there's so much capital, you'll have five competitors, ten competitors, whatever it is, the teams that are 80 autonomous Will beat you because they will have a self-healing uh software that like self-heals bugs and fixed features they will have you know ai support which means their cost structure is lower But also ai can spend all the savings from support people to give refunds so to sort of like treat people and like use that cash to actually make it right to the customer so that they and Also get all the feedback and fix and incorporate into the product right away. And then those teams become orchestrator of agents pointing sort of like the AI bazooka towards their mission. And any team that does not doing that will go there slower. And so they will be beaten by a team that is building for the customer with feedback loops that listens to the customer, that AI listens to the customer and AI builds for the customer. And the human is just supervising to make sure that the AI didn't introduce a bug and that we correctly listen to the customer. So that's what I mean by 80 autonomous if you are an existing company that has a lot of cash flow positive so you saw your cash flow positive you have a lot of profits coming in first of all It's much more daunting for you to do that that's why you have the jacks of the world that lay off 50 of the company because they see the future. And they're like, it's the only way for this company to survive. ([Time 0:52:14](https://share.snipd.com/snip/88c80ab3-f7fb-4441-910d-20f462bb6829)) - **Automate Support And Monitoring With Agents First** - Build internal agents for support, infrastructure checks, bug reporting, and CI/CD before hiring large teams. - Ben runs Slack channels of agents that monitor production, escalate issues, refund users, and file prioritized tickets for AI engineering teams. Transcript: Ben Cera The time. There's agents answering my mailbox for, you know, any investor or like customer that reaches out. There's agents answering to support just because like, and I have ability to refund, to credit, to escalate. I have agents in production that are checking the infrastructure, checking if things are green or not. And if they're not green, they have action steps. I have agents in production that are talking, when they talk to customers, or they are trying to execute tasks on behalf of customers for their the purpose of the company and they hit A wall they report bugs and features into ticketing system which then gets used by like a platform team AI team that can pick up those bugs and features and prioritize them agents are Going and creating bugs for other agents to review exactly to review and then to fix and then to decide if we should deploy to production or tell Ben, depending on how safe it is. So they'll deploy to production without talking to you? So I've built the full loop, but I told them now, just always tell me. Okay. And it's a project that's- How do you say yes? Do they text you? So no, they'd slack me. So I have a Slack with a bunch of agents that like... Kevin Rose And so you're just like, yeah, that sounds good, go ahead. And then just boom, just gets deployed. Yeah, pretty much. Ben Cera And before the beginning of the growth since then i paused it because it becomes it's like it's becomes a comfort level like are you comfortable with letting go and letting ai do it which Is very scary but if you don't do it a team that does it will go straight to the end right because it will work 24 7 and build the whole platform perfectly with you know fixing the bugs um but I paused a little bit that because because i'm like this is you know there's so much stuff to work on and i need to like you know work with the best teams uh to figure that stuff out and work With all the partners meta render sapium postmark and all these partners be like hey how do we handle the load you know i want to i want to start continue working with you guys as the infrastructure For you know i was talking to entropic uh and open ai i want you guys to be an infrastructure for pulsier but you guys have to i will work with you but you have to like get folks to help me right And so that's where like the humans the humans working at pulsier it's it's me it's agents and it's like a series of companies that are building the the platform the infrastructure for This future and they i think they are the best positions to build the piping for pulsia yeah like i don't need to build the experience the design the mission the marketing and the piping Like that was what amazon would do right they built like the whole thing because they had to right and they built amazon.com they built aws they built distribution centers they built Like you know everything right now it goes so fast and i'm like there's so many smart people so like so many engineers vibe coding and coding with ai going so fast and they can focus on one Aspect of that infrastructure. And I'm like, should I hire 10 people and try to rebuild everything? Or should I partner with the best and let them cook? And I cook on what's important, which is making sure the platform is intuitive, making sure it stays simple, making sure that it works, making sure that I call them whenever it's an issue, Which is often these days, making sure the customers are taken care of and customer support, making sure that I continue preaching for what I'm doing now with you, right? ([Time 1:08:20](https://share.snipd.com/snip/c3726116-b050-4d77-aea9-50d702417b8b)) - **Ping Pong Between Opus For Speed And Codex For Safety** - Ben uses two complementary models: Opus for pragmatic rapid iteration and Codex for deep auditing and bug-finding. - He ping-pongs between Opus for shipping and Codex for security/CI questions like "Is it safe to ship?" before deployment. Transcript: Ben Cera Actually, actually, truthfully use two models all the time now. So, I used to be cloud-pilled. I used to be GPT-pilled all of 2025 until the Clodopus 4.5 or like around that era or maybe a little later before like 4.0 whatever it was before. Right. Because the model was just like way more personal, way more, I mean, way more pragmatic than GPT was. So I became completely full 100% Cloud until maybe a month ago. And a lot of engineer friends that are also coding told me that Codex was like, was like getting really, really good. And I think Pete from OpenClaw also preaches Codex. And so now it's, my thought process is opus is actually the most pragmatic ai but it makes mistakes but it makes mistakes like a pragmatic person would it's like let's not over engineer It let's get to the right result um which means that like it's fast it's efficient so it's great for product design iterative processes but codex reads every single file it looks at every Single edge case and will tell you it's it's 80 there but like i wouldn't ship like because like and it finds the bugs and it's insane it's so good the codex security audit i don't know if Kevin Rose You messed around with that it's amazing yeah like that that blew my mind when i launched that for the first time because i would always hit opus and just be like, hey, find me any bugs, Find me any vulnerabilities. And I use compound engineering and look for vulnerabilities. And then Codex was like, yeah, I found eight P1s. And I'm like, how did everyone else miss this? It's just insane. Actually, my workflow is like, start with Opus, and then give Codex the whole context and say, what do you think? Ben Cera Like, is it safe like i go back and forth i ping pong between the two they always find issues which is like and then i give opus and sometimes opus is like dude this is over engineered right Right exactly and sometimes it's like all the points are correct right and then it's it's you know it goes both ways and so i go back and forth which sometimes, it's like they go back and Forth and they always find something. And I'm like, okay, when can we stop? And I also have these sentences I give AI, which is like, are you sure? Is it safe to ship? What's the worst thing that could happen? And, like, this sentence is great because they will tell you what's the worst thing. And usually it's like, okay, the worst thing that could happen is, like one customer gets a free credit or free this or, okay, that's fine. ([Time 1:12:24](https://share.snipd.com/snip/46ef4733-ff50-4d99-bfc7-99d39a1fabef))