Have a “billion‑dollar” idea? CEO of For Sale By Inventor here; I got “scammed” by AI so you don’t have to.

[embed]

I run **For Sale By Inventor (FSBI)**. If you Google us, you’ll see “is FSBI a scam?” and “For Sale By Inventor reviews” right next to our own site. That’s the cost of being in an industry with some truly awful players and real FTC and class‑action history around invention promotion. Instead of arguing with strangers on the internet, I’d rather show you how we actually think about ideas — and how a recent interaction with an AI proved my entire point in the dumbest way possible.

The unsexy truth: we say no to about 98% of the ideas we see. Only around 2% ever move forward.

The “Hall of Fame” bad ideas I’ve turned down

All of these are from people I personally know:

The “Physics? What’s that?” Guy A successful kitchen remodeler I shared warehouse space with. His billion‑dollar idea: “non‑combustible building material so houses will not burn.” That was it. No formula, no science, no testing. Just, “Hey, you figure out the chemistry, and we’ll split the proceeds.”

The Pro Athlete and the Foot Thing A former NFL player I used to play soccer with. Dead serious about a prosthetic severed foot that doubled as an adult novelty item. I thought I was being punked. Then I found out there was already a nearly identical product on the market. He was the second person to think of that exact niche.

The Dunkin’ Dreamer A friend of a friend demanded a private meeting for his “guaranteed winner”: weed‑infused “Turbo” iced coffee, specifically from Dunkin’ Donuts. I had to explain that Dunkin’ probably wasn’t looking to wander into the federal drug‑trafficking business that week.

Those are funny, but they’re not unusual. What kills almost everything we see isn’t passion — it’s this combo:

• No feasible path to actually making the thing.
• Nothing protectable because it already exists or is too generic to defend.
• No real market beyond “everyone will want this” and some napkin math.

On top of that, a huge chunk of what walks in the door are BBQ sauces, home recipes, half‑baked app ideas, jeans brands, and “perpetual motion” machines. By the time you filter those out, you’re already very close to that 2% that might have a real shot — and even those usually need a lot of work before they look like real inventions.

At FSBI, if an idea doesn’t clear those three filters, we say so. That’s where a lot of “FSBI scam” narratives start: someone hears “this is not protectable” or “there is no market for this at the price you need,” and it feels like we’re killing their dream instead of keeping them from lighting their savings on fire.

And that’s exactly why what happened next with AI scared me.

I asked an AI about a time machine. It loved it.

Because I’m building a new platform called TheIdeaPost that will sit before FSBI in the funnel, I decided to test something absurd: a spatiotemporal displacement “time machine.”

I asked Google’s Gemini to help me treat it like a real invention. Here’s what it did (this is all from one session):

• Drafted a full Provisional Patent Application (PPA) spec: background, summary, detailed description, numbered parts, and a filing checklist.
• Generated a drawing with reference numerals and told me it was an “excellent drawing for a PPA” that followed USPTO best practices and would help secure my “Patent Pending” status.
• Walked me through saving everything as PDFs and uploading to the USPTO Patent Center, including mentioning the $65 micro‑entity fee.
• When I asked, “Is this good?” it doubled down that it was excellent and just suggested a small labeling tweak to avoid “claim ambiguity.”
• When I asked, “Do you think this will sell?” it gave me odds of success:
• Patent approval odds: <1% (because physics).
• Licensing/commercial odds: 75% as entertainment IP, movie prop, game asset, etc.
• Funding odds: 15%, with a suggestion to pivot the story to “quantum computing” or “advanced energy storage.”

It then offered to help write a white paper, a pitch deck, and refine my “secret sauce” physics theory with things like wormholes and Tipler cylinders.

If I said that to a real FSBI client about a time machine drawing, I’d be squarely in the territory the FTC now calls using AI to “supercharge deceptive or unfair conduct that harms consumers,” and I would deserve whatever came next.

Why the AI answer is exactly what I can’t do

On paper, Gemini’s answer sounds smart. It cites the utility requirement, admits time travel violates current physics, and notes that a PPA isn’t examined so you’ll still get “patent pending” for 12 months. All technically true.

But in practice, here’s what it did:

• Treated “patent pending” as a product and a win, even when the underlying invention is impossible.
• Framed a time machine PPA as a “solid draft” and encouraged filing.
• Gave pseudo‑precise success odds and suggested 75% commercial potential as “creative IP.”
• Offered to help build investor‑facing materials around something that cannot exist under current physics.

Swap out “AI” for “invention company representative” and you’ve got exactly the pattern that’s gotten invention‑promotion outfits sued, fined, and banned: exaggerating success, implying real commercialization, and selling filings for ideas that will never clear utility or enablement.

The AI doesn’t have a license, a BBB profile, or any duty not to hallucinate. I do. Courts, regulators, and bar associations are already warning that AI hallucinations and overconfident legal‑adjacent advice are a real risk in anything touching IP or law.

How FSBI and TheIdeaPost handle this instead

Here’s how we actually run things:

• We filter out physics‑violating, impossible, or obviously non‑enabled “inventions” up front. Time machines and perpetual motion don’t get “patent pending”; they get a polite “no.”
• We don’t treat a PPA as a meaningful asset unless there’s a realistic path to utility, enablement, and a future non‑provisional.
• We’d rather tell you “don’t spend this money” than sell you a glossy report on an idea that can’t be built or protected.

For TheIdeaPost, the whole point is to give the 98% a cheaper, lower‑stakes way to get reality checks and feedback without getting sucked into an expensive development package. That’s why:

• We’re not putting AI in charge of go/no‑go decisions.
• We’re building a human‑driven review + GOAT score that looks at feasibility, protectability, and market, not just “cool idea” points.
• AI will be used, if at all, as a drafting/calculation tool — not the voice that tells you your retirement savings should fund a time machine.

If you’re an inventor, start here before you pay anyone)

If you’re worried about getting burned — by FSBI or anybody else — do this first:

Write down, in plain English:

1. How does this actually get made? What materials, what processes, what technology exists today that could do this?
2. What, exactly, would be protectable? A mechanism, a configuration, a method — something more than “a better version of X.”
3. Who buys it, at what price, and how many of them exist? Not “everyone.” Name the customer, the channel, and a price that fits reality.

If you can’t answer those cleanly, you don’t need a development firm yet. You need a reality check. An AI telling you your time machine is a “10/10 technical concept” is not that reality check.

Our record is not perfect; nobody in this space is. But it’s a lot cleaner than the horror stories you’ve seen specifically because we’re this selective and willing to say no.

We’re actively building something for the 98% — a way to get straight answers and structured feedback without going all in on a full FSBI engagement. When it’s ready, I’ll share it.

In the meantime, I’m at my desk. If you’ve got a “billion‑dollar” idea and want a human opinion on whether it’s in the 2% or the 98%, or you’ve read the “For Sale By Inventor scam” and “FSBI reviews” threads and have questions, drop them in the comments and I’ll tell you what I actually think.

TL;DR: I run FSBI. We reject about 98% of ideas because they’re not feasible, protectable, or marketable. An AI just cheerfully helped me build a provisional patent for a time machine and talked about its “odds of success.” That’s exactly the problem. The wild stories are fun; the real value is stopping bad bets before they eat your savings.