# Jake Ostrowski, Indie Maker / Fractional Dev at Ostrowski Works — read of CrankGPT, 2026-06-17

> "6 years trying to ship something that sticks. Two apps in the App Store, one micro-SaaS I killed at month 8. I read idea packages the way some guys read fantasy football reports."

## How I got here

Someone dropped this in the #ideas channel of a small Slack group I'm in for indie hackers. The message was just "anyone looked at Wishdeal Factory yet?" and a link. I had 20 minutes on my lunch break so I clicked it. I was expecting a regular SaaS landing page. What I got was something different enough that I kept reading.

## What I clicked first

The score hit me before the hero copy did. "59/100 Adoptability" right up top. I've never seen a product page lead with its own mediocre grade. Then: "$-21,890 Year-1 take-home (Fermi)." That's negative. The product is telling me it will likely lose money in year one. That's either the most honest thing I've ever seen on a startup page or a very clever hook. I'm not sure which yet.

## Where I paused

"financial upside: 1/10" and "pain intensity: 4/10." I sat on those two for a minute. They're burying the lead on the actual CrankGPT idea, which is supposed to be what I'm buying into. If the upside is a 1 out of 10, why would I spend $99-199 on a starter kit for it? The framing is: we're being honest with you. But the honest answer seems to be "this specific idea isn't very good." I kept waiting for them to explain why I'd still want it and the page never quite got there.

## What I distrusted

"Voice capture: One interview. CrankGPT learns your style, tone, and positioning. Every output after feels native to your channel."

I've heard this exact claim from Jasper, Copy.ai, Anyword, Typeface, and at least four others I've forgotten. "Learns your voice" is the "We use AI to..." of 2025. The page doesn't show me what the voice capture actually looks like, what the interview consists of, or what "native to your channel" means in practice. The live demo section is referenced ("Try it Live result Before With CrankGPT") but in the stripped text I can't tell if there's actually a working demo or just a static before/after mockup. That gap matters a lot.

Also: "Honest disclosure: we don't have live customers on this idea yet." This is good to say, but it also means the "credibility: 9/10" axis score is based on what exactly? The concept's credibility? The founders' credibility? That number is doing a lot of work without explanation.

## What would convince me

I'd want to see the Fermi math shown out in the open, not just the summary number. Walk me through the $-21,890 estimate. What assumptions are baked in? What's the customer count, what's the price point, what's the churn rate assumed? If the math is laid out and I can poke holes in it or see where I might beat it, that's actually useful. A black-box negative number just makes me think the idea is a dud.

Separately: if someone has actually done the voice capture interview and shown me 5 outputs side by side with what they'd have written themselves, and the outputs are close, that would make the core claim real instead of theoretical.

## What I'd ask in an email reply

1. The "credibility: 9/10" score -- what does credibility mean here? Is that about the founder's credibility, or the market's credibility? Because the market is brutally crowded.

2. The $99 adopt tier says I get "working code starter." What stack? What does it actually do out of the box -- is it a Claude API wrapper with a UI shell, or something more substantial?

3. You score this 1/10 on financial upside. Is the point that this is a lifestyle business and not a venture bet, or do you think the idea itself is weak and you're selling it anyway?

## Verdict: on-the-fence

The honesty frame is genuinely interesting and I haven't seen it done this way before. But the scores they publish for the idea are, charitably, not a strong pitch for the idea itself. I'd probably pay $5 for the dossier just to see if the underlying analysis is as sharp as the framing suggests.

---
*Memo by skeptic persona, generated 2026-06-17. Studio breaks own self-grading loop.*