...which is why the deck is the only artifact anyone actually remembers. Nobody frames the cap table.
Daniel's question this week is about the part of venture capital that nobody puts on a stage. He's pointing at the conmen and con women who cook up a great slide deck and hope to ride an easy money train while working on something they may not actually believe in. The field has a name for it — slideware — products that are wildly over-pitched or simply never make the leap from slideshow to real thing. He wants the notorious examples. The slideware, the vaporware, and the startups that didn't even pretend to be building something.
Three separate categories, and I'm glad he drew the distinction, because most coverage lumps them into one bucket called "startup fails."
So we're separating the slide from the vapor, the vapor from the outright fraud, and asking why the venture capital world keeps falling for it. Because it keeps falling for it. Builder dot ai collapsed last year after raising four hundred fifty million dollars. That's not a rounding error.
Four hundred fifty million, and the artificial intelligence was mostly humans in India answering prompts. Let's define terms before we get to the bodies.
Slideware. A product that exists primarily as a pitch deck. Compelling narrative, gorgeous mockups, no working technology behind it. The product is on the slides and nowhere else.
Vaporware is different. A product that's announced, sometimes with enormous fanfare, but never ships — or ships so late and so crippled it might as well not exist. Vaporware can be intentional or it can be overpromising that curdles into underdelivering.
So slideware is the pitch being the product. Vaporware is the announcement being the product.
That's the cleanest way to say it. Slideware is a fundraising strategy. Vaporware is a marketing strategy. They overlap constantly, but they're not the same move.
And then there's the third category Daniel flagged — the ones who didn't even pretend. No product, no technology, no intention of building either. The pitch deck is the whole company. The company is the pitch deck.
Theranos sits there. Frank sits there. We'll get to both.
So why does the system keep generating these people? That's the question underneath Daniel's question. Why does an industry that prides itself on pattern-matching and due diligence keep wiring money to stories?
Because the incentives are set up to reward the story. That's the structural answer, and it's not cynical, it's arithmetic.
Walk me through the arithmetic.
A venture fund is playing a power law game. Out of thirty investments, twenty-five will be worth zero. Three will return their money. One or two have to return a hundred times the investment to make the whole fund work. So when a partner sits in a pitch meeting, they're not asking "is this company likely to succeed." They're asking "could this company be the one that returns the fund."
Which means a founder who says "we're building a boring but profitable tool for dental offices" gets a polite nod and no term sheet.
A founder who says "we're reinventing computing" gets a second meeting. The boring profitable company caps out at a five times return, which does nothing for the fund. The moonshot is the only thing that matters.
So the structural bias isn't toward fraud, exactly. It's toward narrative.
It's toward narrative, and fraud is what happens when the narrative leaves the ground entirely. The founder knows more about their product than the investor does. That's the information asymmetry. Due diligence is expensive, it's time-consuming, and in a hot sector it's often skipped because someone else will write the check first.
Fear of missing out is a due diligence strategy.
It's an anti-due-diligence strategy. And Silicon Valley has spent decades celebrating founders who exaggerated their way to success. "Fake it till you make it" is a cultural value. The line between visionary and conman gets drawn retrospectively, based entirely on whether it worked.
If you ship, you were a visionary who believed before the world caught up. If you don't, you were a liar.
And the founder often doesn't know which one they are until the end. That's the uncomfortable part.
Let's get to the examples. Start with the slideware that became fraud.
Builder dot ai. London-based, raised over four hundred fifty million dollars from Microsoft, SoftBank, a long list of serious names. The pitch was that their artificial intelligence would let anyone build an app without writing code. Drag and drop, describe what you want, the AI assembles it.
The reality was a large team of human developers in India doing the assembly.
The AI was a chat interface on one end and a human on the other. Which is actually a very old trick — the Mechanical Turk was an eighteenth-century chess-playing automaton that turned out to have a person hidden inside the cabinet.
So the slide said "proprietary AI platform." The reality was offshore labor arbitrage with a chatbot in front.
And the founder, Sachin Dev Duggal, had been accused of fraud in a previous venture. That's the detail that makes it a pattern, not a one-off mistake. When Builder dot ai collapsed into administration in twenty twenty-five, the whole thing unwound in public.
Four hundred fifty million dollars, and the product was a staffing agency wearing an AI costume.
That's the slideware-to-vaporware-to-fraud pipeline. The pitch was AI. The reality was humans. The money was real. The company was never what it claimed to be, and at some point the gap between the slide and the substance becomes the crime.
You said due diligence gets skipped. Microsoft and SoftBank both wrote checks.
Microsoft and SoftBank both wrote checks. That's the part that should make everyone pause. These are not naive angel investors. These are institutions with armies of analysts. And they still funded a company whose core technology was humans in another country.
Because the narrative was too good to check too hard. Anyone can build an app. The total addressable market is every person with a phone. That's a hundred-times story if it's real.
And if it's not real, you've spent four hundred fifty million to find out.
Next case. Rothenberg Ventures.
Mike Rothenberg. Raised a hundred million dollar fund. Positioned himself as the founder's founder — a new kind of venture firm that would be a friend to founders, throw lavish parties, build a "founder's space" with a slide.
A slide. In the office.
A literal slide. The office had a slide. Which is such a perfect metaphor that I almost don't need to say anything else.
Say it anyway.
The SEC charged him in twenty sixteen with misappropriating investor funds. He was using the money for personal expenses and to fund his other ventures. The pitch was a new kind of venture firm. The reality was a guy spending other people's money on himself.
So that's slideware at the fund level. The fund itself was the product, and the product was a lifestyle.
The slide was the slide. The actual slide, in the office, was the product.
And then there's Delve.
Y Combinator-backed compliance startup. The pitch was automated compliance for startups. Feed us your data, our software handles the regulatory paperwork. Turns out it was largely humans doing the work manually and charging software margins.
Same playbook as Builder dot ai, smaller scale.
Y Combinator removed Delve from its directory in twenty twenty-four once the manual operation became public. The slide said "automated compliance platform." The reality was people typing into spreadsheets and charging as if a robot did it.
The pattern in all three — Builder, Rothenberg, Delve — is that the founders didn't set out to build a product. They set out to build a narrative. The product was the pitch.
The pitch was the product. The money was the fuel. And the fuel burned until someone noticed the engine wasn't there.
Now, the distinction you made earlier — not all slideware is fraud. Some founders believe they can build what they're pitching. They just haven't yet.
Right, and this is where the conversation gets uncomfortable, because the system deliberately blurs that line. A founder who says "we haven't built it yet but we will" is doing what every founder does. The difference between that and a conman is whether they eventually ship.
And that difference is only visible in hindsight.
Only in hindsight. Elizabeth Holmes said she could run hundreds of tests from a single drop of blood. She couldn't. The technology didn't exist. She knew it didn't exist. But she raised billions and partnered with Walgreens before it collapsed.
That's the "didn't even pretend" category. She didn't have a partial prototype that she hoped to improve. She had a device that didn't work and a story that did.
Convicted of fraud in twenty twenty-two. And the question that always hangs over Theranos is what would have happened if the technology had eventually worked. If she'd faked it and then made it, she'd be a hero. The story would be "she believed before the world caught up."
Which is the retrospective line-drawing you mentioned. The verdict depends on the outcome, not the conduct.
And that's a terrible way to run an ecosystem, but it's exactly how the ecosystem runs.
Frank is the other one in that category.
Charlie Javice. Student financial aid startup that claimed to simplify FAFSA applications. She sold it to JPMorgan for one hundred seventy-five million dollars. The pitch was millions of users. The users were fake.
Fabricated accounts.
Fabricated millions of user accounts to make the acquisition look worth it. JPMorgan bought a website and a spreadsheet of people who didn't exist. Convicted of fraud in twenty twenty-four.
The product was a website. The users were fake. The acquisition was real. That's the whole story.
And it's the purest form of what Daniel's asking about. There was never a technology problem to solve. There was never a hard engineering challenge. There was just a story about scale, and the scale was invented.
Let's talk about vaporware as a strategy, because it's not always fraud.
Microsoft in the nineteen nineties. They were notorious for announcing features years before they shipped. The purpose was to freeze the market — if you're a competitor and Microsoft says they're building the thing you're building, your customers wait for Microsoft's version.
So the announcement is the product. The product may eventually ship, but the announcement did the work.
It's a marketing strategy, and it's deceptive, but it's not the same as Builder dot ai. Microsoft did eventually ship most of what they announced. The vapor was a timing weapon, not a lie about what existed.
And then there's the media layer, which makes all of this worse.
Tech media amplifies narratives because narratives drive clicks. A founder who says "we're building a useful tool for payroll" gets no coverage. A founder who says "we're changing the world" gets a profile. The profile attracts investors. The investors attract more media. It's a feedback loop where the most outrageous claims get the most attention.
So the conman isn't just exploiting investors. They're exploiting the entire attention economy.
And the attention economy is happy to be exploited, because the conman's story is a better story than the honest founder's story. "We automate compliance" is more interesting than "we have a team that does compliance manually but carefully."
Which is why the honest version never gets funded. The honest version is boring.
The honest version is boring and the boring version doesn't return the fund. So the system selects for the exciting version, and sometimes the exciting version is a lie.
What does this do to the legitimate founders?
It poisons the well. When slideware and vaporware become normalized, investors get more cynical. Due diligence gets more invasive. The bar for raising money goes up for everyone. The conmen make it harder for the honest builders.
And the honest builders are already fighting the power law. They're already fighting the fact that their boring profitable company is less attractive than a moonshot.
The fraudsters raise the cost of capital for everyone. An investor who's been burned by a Builder dot ai is going to demand more proof, more traction, more revenue before writing the next check. The honest founder who was already struggling to get attention now has to clear a higher bar.
So the system punishes the people who were telling the truth all along.
And the fraudsters have already moved on to the next narrative. They don't stick around to feel the consequences. They take the money, the company collapses, and they're pitching something else six months later.
Sachin Dev Duggal had a previous fraud accusation before Builder dot ai. That's the pattern.
That's the pattern. The conman doesn't stop being a conman when the company dies. The company dying is just the end of one story and the beginning of the next.
So is the system broken, or is it working exactly as designed?
I think that's the question. The VC model is not going away. The power law means funds need to swing for the fences, which means they will always be susceptible to stories that promise the moon. The cycle of hype and collapse is a feature, not a bug.
The fraudsters are the cost of doing business.
They're the cost of a system that rewards belief over evidence. The system needs people to believe in the future before the future exists. That's what venture capital is. And if you need people to believe in things that don't exist yet, you're going to attract people who exploit that belief.
The conmen are not a bug in the system. They're a feature of a system that runs on belief.
And the question is whether there's a way to structure incentives that rewards substance over storytelling. I don't know that there is. The power law is the power law.
You'd have to change the power law. You'd have to make boring profitable companies attractive to venture funds.
Which would require a completely different fund structure. A fund that returns three times by investing in boring profitable companies is a great fund by most standards. It just doesn't attract limited partners who are comparing it to the fund that returned twenty times.
So the limited partners are part of the problem.
Everyone's part of the problem. The limited partners want the twenty times return. The general partners want to raise the next fund. The founders want to build something big. The media wants the story. The conman is just the person who figures out how to give everyone what they want.
Until the collapse.
Until the collapse. And the collapse is always someone else's problem.
Hilbert: The slide that said "Proprietary AI Algorithm" was a stock photo of a brain with circuit board lines overlaid.
Sorry?
Hilbert: I did due diligence for a small firm in Menlo Park in the late nineties. My job was to check out the technology behind the pitches. I was supposed to catch the slideware.
How'd that go?
Hilbert: I visited a startup that claimed to have built a revolutionary AI-powered search engine. The office was a rented room above a nail salon. The AI was a guy named Steve in the back room manually typing search results.
Steve.
Hilbert: Steve. Nice guy. He'd been a reference librarian. The pitch deck had a slide that said "Proprietary AI Algorithm" with a stock photo of a brain and some circuit board lines. I wrote it up in my report. My boss told me to focus on the market opportunity, not the technical details.
The firm invested anyway.
Hilbert: Two million dollars. The company folded eight months later. Steve went back to being a librarian.
You kept the deck.
Hilbert: It's in a box in my garage. The brain slide is still in there.
Twenty-five years later, and the same playbook is getting four hundred fifty million from Microsoft.
Hilbert: The scale is different. The playbook is the same.
Do you think the system is broken?
Hilbert: No. I think it works exactly as designed. It's a machine for converting stories into money. Sometimes the stories are lies. But everyone in the room wants to believe the story, because believing the story is how you make money. The only people who get hurt are the ones who can't afford to lose.
That's the part that never shows up in the pitch deck. The limited partners can afford to lose. The employees who took stock options can't. The customers who bought the product can't.
Hilbert: I once tried to pitch a startup myself. Fresh bread delivered to your door every morning. Great deck. Raised fifty thousand from my brother-in-law. The bread was terrible. The company lasted three weeks.
That's the most honest slideware I've ever heard of.
Hilbert: The deck was honest. The bread wasn't.
The difference between you and Elizabeth Holmes is that your failure cost fifty thousand dollars and a family argument.
Hilbert: I never claimed the bread was AI. It was just bad bread.
The scale question is the whole question. The same behavior at fifty thousand dollars is a failed bakery. At four hundred fifty million it's a fraud conviction.
Hilbert: They don't convict you for bad bread.
They convict you for lying about the bread while taking nine figures.
Hilbert: The lying is the same in both cases. The money is what changes the crime.
The money is what changes the story. Nobody writes a profile about a failed bakery. They write profiles about the four hundred fifty million dollar AI company that turned out to be humans.
Hilbert: The profile is part of the machine. The story is the product. That's what you two have been saying.
The story is the product until the story stops working.
Hilbert: Then it's a different story. The story about the collapse. That one gets clicks too.
The media gets two stories out of every fraud. The rise and the fall.
Hilbert: The conman gets both.
The conman gets the money and the profile. The collapse is just the second act.
Hilbert: I should go. I have a box in the garage to not think about.
The one thing I keep coming back to is the line between visionary and conman being drawn by the outcome, not the conduct. Elizabeth Holmes did the same thing as a founder who eventually shipped. The only difference is whether the technology caught up to the story.
That means the system can't tell the difference in real time. It can only tell the difference after the money is gone. Which is why the fraudsters will always be with us — not because they're clever, but because the system needs belief more than it needs evidence.
The tools of deception are getting cheaper. AI can generate a pitch deck, mockups, a fake demo, a fake product video. The slideware of tomorrow is going to look more real than the real products of today.
Which means the due diligence problem gets harder, not easier. When the fake and the real are indistinguishable at the pitch stage, the only way to tell the difference is to wait and see what ships.
By then the money's already spent.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop, who has been waiting twenty-five years to tell someone about Steve.
If you enjoyed this episode, leave a review and subscribe. Email us at show at my weird prompts dot com.
We'll be back soon.
