Daniels Pitkevics

If you want to stop configuring and start shipping, I packaged the entire workflow into the Vibe Coding Starter Kit. • Native Desktop App • Next.js 16.1 • Stripe, Supabase, SEO ready Grab it here: vibecodingstarterkit.io Watch demo: youtu.be/4SywZzPKOgA

Hit $5K MRR alone in my apartment. No team to high-five. No office to celebrate in. Took myself out for coffee. Bootstrap wins are lonely. All the risk is yours and all the victory is yours. Both feel weird.

No co-founder means no one to tell you an idea is stupid before you waste a week building it. I've shipped 8 features nobody asked for this year. VC founders have boards to say no. Bootstrap founders have regret. The freedom is also the problem.

AI lets me say yes to every feature. 40 features in 6 months means I maintain 40 features. Speed created maintenance burden I didn't expect. Each yes costs me forever. Fast building makes it easy to accumulate debt you can't pay back.

Vibe coding changed what I'm willing to build. I tackle problems I'd never attempt from scratch - payment reconciliation, email parsing, complex forms. The AI lowers the intimidation factor. But now I have code in production I'd never have confidence to write myself. Empowering.

I vibe code daily. My typing speed for boilerplate is getting slower. AI writes it faster than my muscle memory can. When the API goes down, I'm worse at coding from scratch than I was 2 years ago. Speed has a cost - your fundamentals atrophy if you don't use them.

Took 3 weeks to train my AI agents on how I want emails written. Now they save me 5 hours a week. The upfront investment killed my productivity for almost a month. Automation has a learning curve that nobody talks about. You go slower before you go faster.

@JamieBrock19060 @csaba_kissi The format matters too. I keep a decisions.md file - why I chose patterns, what I tried that didn't work, tradeoffs I accepted. When you need to change something months later, that context is worth more than the code comments. Good luck with the launch.

@DanielsBuilds @csaba_kissi Good advice. Evolving documentation is a huge part of the process - no need for coding experience to teach you that.

If you can still code without AI, you’re a real developer. Are you a real developer?

@TheConfigGuy @psomkar1 The diff review is everything. I've seen entire repos converted to TypeScript with subtle logic bugs hidden in type assertions. Looks clean in the diff but breaks at runtime. If you can't explain why every changed line is safe, don't merge it. Speed without review is gambling.

@psomkar1 That's Claude Sonnet 4 converting an entire repo to TypeScript in one shot. Vibe coding or bold automation, depends entirely on whether you reviewed the diff before merging. My guess is the answer is no.

Dear developers, is this called vibe coding??

@SuspiciounDude @om_patel5 Marketing forces you to articulate value to strangers. Building lets you hide behind features. That's why it's harder. You can vibe code a product in a weekend. You can't vibe code someone into caring about it. Persuasion requires clarity that code doesn't.

@om_patel5 Marketing is 10x harder than vibe coding

when you spend 48 hours vibe coding your startup and the stripe dashboard doesn't instantly show $1,000,000 MRR

@casperbarn Congrats. Now keep them. That first customer teaches you more than the next 10 combined. Ask them why they paid. What almost stopped them. What they expected vs what they got. Their answers shape your next move. Don't hunt customer 2 until you understand customer 1.

I just got my very first paying SaaS customer. The adrenaline is insane, but now what? I honestly can't believe it. I just saw the Stripe notification pop up on my phone, and for the first time, I have actual MRR.

@JamieBrock19060 @csaba_kissi The magic feeling is real until you need to change something 3 months from now. Document your architecture today - what talks to what, why you made key decisions. Future you won't remember the prompts. You need to own what Claude built, not just deploy it.

@DanielsBuilds @csaba_kissi Built entirely myself. 95% Claude, 4% Codex, 1% Gemini. Complex is an understatement. I sit there watching in awe as actual magic is happening on the screen. Working through the pre-deployment process now, polishing the landing page. Launch is imminent.

Had 12 people say 'this is exactly what I need' before I got my first sale. Words are free. Money is truth. The gap between 'I need this' and 'I'll pay for this' taught me more than any customer interview. Watch what people do, not what they say.

@2027launches @levelsio @marclou 600 users with $0 revenue means you're solving a problem people don't value enough to pay for. Or you haven't asked. Try this: pick 10 users, DM them a price, see what happens. If they all say no, you built the wrong thing. If they say yes, you've been giving it away for free.

Got hyped to build in public after being inspired from @levelsio and @marclou . Launched my first website which is at 600 users right now. None paid. Would love feedback. Link in first comment.

@ericson4smith Exactly. Claude Code ships faster than you can think through edge cases. That's why code rescue exists - people skip planning and ship AI's first pass. I've fixed projects where features worked on launch day but broke everywhere else. Planning is the bottleneck now, not coding.

Claude code literally creates the feature faster than it takes to plan it. It just shows how important the planning process is, and if you don't have experience in coding - your vibe code output is gonna be shitty.

@JamieBrock19060 @csaba_kissi That's a complex stack for no-code. Did you use AI to build it or hire someone? Either way, the 10 years of domain expertise is your real asset. You know the problem better than any developer could. The code is just execution. How close to launch are you?

@DanielsBuilds @csaba_kissi It's a vertical SaaS for service/retail-based businesses. Full stack — Next.js frontend, FastAPI backend, Postgres, Stripe Connect, Xero/QuickBooks integrations. Solves a specific operational pain point that I've lived with for over a decade. Happy to share more once it's live.

Customers ask 'will this work for my use case?' I always say yes. Because it will - with work. Starter kits aren't plug-and-play. They're educated scaffolding. If you expect zero customization, you're buying the wrong thing.

@Iwasonceblijuy Same workflow here. The 'while I sleep' part sounds magical but the review in the morning takes hours. I've had Claude build features overnight that looked perfect but broke 3 other things. 800 users is solid. How much time do you spend reviewing vs letting it ship?

My SaaS has 800 users. I’ve written a good chunk of the code myself. But Claude Code still builds features while I sleep. I describe it. It scaffolds it. I review it. It ships it. Embedded systems engineer by trade. Solo founder by choice.

@JamieBrock19060 @csaba_kissi No-code works great for validation. The problem comes when you need something custom or hit platform limits. I do code rescue for projects that outgrew no-code and had to rebuild. If your problem is simple enough, no-code might be enough forever. What's the product?

@csaba_kissi I’ve never written a line of code in my life, and am just about to launch a B2B SaaS. I definitely don’t consider myself a developer or even founder - I’m just a dude with a problem no one else had solved and plenty of free time.

@wryter_the @dorgoclips @mwilcox This is what I see doing code rescue. Vibe coding lets you ship without understanding what you built. Works fine until something breaks and you can't debug it. The danger isn't the AI - it's thinking you can maintain production code you don't understand.

@dorgoclips @mwilcox No, that's not right. Without vibe coding, many “AI startup BS” companies (=one-man shows) wouldn't be able to offer anything at all. Now they can offer something, but they still don't have any know-how. This is highly dangerous for all sides.

@victor_bigfield Agree with most of this. The 'real problems' part is key - too many people build what sounds cool instead of what someone will pay for. I ship fast but I only start if someone's already asked for it or I can find 10 people with the problem in 48 hours.

want to make money as an indie hacker? start by solving real problems build tools people actually need automate the boring stuff ship fast, iterate faster stop overthinking the tech stack start shipping solutions wrote about my approach to building profitable side projects check it out