Indie Hacking, Different Game
Indie hacking is back. Different game now.
The old playbook still works — small surface-area products, charge from day one, kill fast, ship rough — but the math underneath has fundamentally changed. Build cost collapsed. Judgment cost stayed the same. The economics inverted. What used to take three months of nights and weekends now takes three weeks if you have the right operator at the controls and the right agents doing the work.
The infrastructure was waiting
Years ago I wrote Incorporation for Indie Hackers — a piece on why a serious solo operator should incorporate properly: Wyoming LLC, Stripe, banking, accounting, the boring infrastructure most indie hackers postpone until it's painful. I set up Petreski LLC then and have been quietly running it ever since. Not for any single product — just so the moment something needs to charge real money to real customers, the rails are already there.
That moment is now. The company site got a recent redesign — clean, minimal, no marketing flourish, just the operational facts (D-U-N-S, Wyoming registry, contact). It looks like nothing because it is doing nothing visible yet. That is intentional. The next move is products that ship through it.
What changed
Twenty years of senior software engineering used to mean a lot of muscle memory and a lot of typing. Now it means a lot of judgment and a lot of orchestration. I run Claude Code with multi-agent setups daily. The throughput multiplier on build / refactor / docs / scaffold work is genuinely 5-10x for a senior operator who knows what good looks like.
That changes the indie hacking calculus completely. The bottleneck used to be "can I build it in a reasonable time." Now the bottleneck is "should I build it, and who will pay for it." Build cost approaches zero. Validation, distribution, judgment — those stay expensive. The whole game shifts toward the part of the work AI can't do: deciding what's worth doing, talking to humans, charging money, killing things fast when they don't earn.
Pieter Levels and Tony Dinh and the early IndieHackers crowd built the original playbook on grit and brute-force shipping. The next wave of indie hackers will build the same playbook with AI agents as their build team and senior judgment as their differentiator. Most won't have the judgment. The ones who do are about to have an unfair window.
Same operator, different surfaces
For anyone reading both this and the biohacking series wondering if I am switching tracks — I am not. Petreski LLC has been the consulting and engineering vehicle for years and continues to be. The senior engineering work, the AI-native consulting, the existing client engagements — all of it runs the same as before. The cockpit is a parallel pursuit, not a replacement. Different cylinders of the same engine.
The throughline across what I publish is consistent: take a domain that is treated as fixed, apply systematic pressure with the right tools, watch the constraints fall. Bodies optimize. Businesses ship. Boring problems generate revenue. Same operator, different surfaces.
The cockpit is live
This week I bootstrapped something new. I am not going to describe it in detail yet — show the work, not the slide deck — but the shape of it is worth naming.
It is not a single product. It is an operating system for running multiple ventures in parallel. A cockpit. Idea pipeline, validation playbook, MVP scaffolding, ship logs, kill criteria, marketing channels, the LLC infrastructure all wired together. Each venture lives in its own slot. Cross-venture knowledge accumulates. The fifth idea launches easier than the first because everything that worked got lifted into the shared layer. Everything that did not got documented as a lesson and a kill retro.
The principles I am building it around are old but the AI-era execution is new:
- Money is the goal. Cool is not a filter. Boring + chargeable beats novel + unpaid every time.
- Validation = shipping + charging. Landing pages with email capture are not validation. The only signal that counts is a real user paying for the smallest real version.
- One month per idea. Past day 30 with no clear keep / kill signal, default to kill. Carrying weak ideas is the cost of not starting strong ones.
- The machine compounds. Individual ideas do not. Most will die. The infrastructure, the playbooks, the AI workflow, the LLC operations — that is what survives.
- Honest retros or no retros. Killed an idea and cannot articulate the transferable lesson? The kill is not done.
The ladder
Three rungs. I am not putting the numbers in the article yet — show the work, not the targets — but the shape is: proof of life from a single venture sustained over multiple months, then income replacement from a small portfolio, then a long-term stack of small wins compounding into a number that does not require any other source of income to exist.
That last rung is years out. The first rung is months out. The middle is the one I am actually optimizing for, because that is the one that changes day-to-day life — exit the trade-time-for-money loop entirely, work on what I want, on my clock.
Different game
The infrastructure was waiting. The build cost collapsed. The judgment is sharpened by twenty years of doing the work. The AI agents are wired in. The cockpit is live. First moves coming.
For senior engineers watching the AI-era indie hacking conversation from a distance: the window where the field is wide open and most operators have not yet adapted is now. It will not stay open. Worth taking the question seriously.
More soon.
