🎯🔥 Why Bad AI Products Are Actually Winning

In partnership with

Icon spent $12 million on a domain name, promised to replace every ad creation tool for $20 a month, and delivered a product so bad it had 90% weekly churn. Now they're pivoting to become a creative agency charging $1,000 monthly for human support. This sounds like failure, but it might actually be the most rational strategy in AI today.

Welcome to the age of levered beta, where being smart is overrated and being early is everything.

The Bet That Looks Insane (But Probably Isn't)

Kennan Davison looked at AI video generation in early 2024—which was producing absolute garbage and made a calculated bet. The technology was coming, just not yet. So he dropped an ungodly amount of money on icon.me, launched with a product that users hated, and started screaming "AI ADMAKER" from every rooftop and subway ad in New York.

He knew it was trash. Everyone knew it was trash. But that wasn't the point.

Davison was playing a different game entirely: claim the territory now, survive until the technology catches up. When Veo3 or whatever finally makes AI video not suck, who's going to own the mindshare? The company that waited until the tech was ready, or the one that's been burning money on marketing for two years straight?

The brutal math works like this: Icon claimed $5 million in revenue within 30 days of launch, then went suspiciously quiet about revenue updates. That's because when your product costs $20 monthly and has 90% weekly churn, even the entire market of early adopters isn't big enough to sustain growth to their proclaimed $100 million ARR target.

So instead of trying to fix the impossible (automating the end-to-end process of understanding a business, analyzing competitors, generating creative angles, producing on-brand assets, and publishing them to Facebook Ad Manager for twenty bucks) they're doing something smarter.

They're raising prices (basic tier now $39, premium tiers at $399 and $3,000/mo), adding human support, and essentially becoming a marketing agency that uses AI to improve efficiency by 20-40%.

It won't be a $10 billion company. But it might make Kennan Davison a wealthy man.

Why Bad Products Can Still Win

Icon's approach starts to make sense when you realize they're playing an entirely different optimization game. Most startups chase product excellence: better features, happier customers, sustainable unit economics. Icon is chasing market timing.

It's a bet that being terrible but present beats being absent but perfect. The logic goes: if foundation models improve dramatically over the next two years (which seems likely), then the company that owns mind-share in "AI video ads" will benefit regardless of how awful their current product is.

This explains why Icon tolerates 90% weekly churn while spending millions on domain names. Traditional metrics like NPS and retention rate become irrelevant when you're not optimizing for today's customer satisfaction. You're optimizing for tomorrow's market position.

Other companies are making similar moves. Cluely raised $5.3 million for their AI coaching vaporware and spent it on outrage marketing, a manifesto calling cheating "leverage," viral videos of their founder lying on dates. Seven-point-eight million views of "Black Mirror dystopian" content. But I bet you can't name another AI sales coaching startup.

The strategy works because buyers are making purchasing decisions based on category trends rather than product quality. I'd guess most CROs who approved AI SDR budgets in 2024 knew the tools were mediocre. But they needed to tell their boards they were "doing something with AI," and mediocre AI beats no AI in a board presentation.

It creates a weird market dynamic where companies can grow revenue despite fundamentally broken products, as long as they're riding a big enough trend.

When the Technology Finally Catches Up

Harvey offers a glimpse of what happens when AI actually starts working in a vertical. The legal AI company is valued at $5 billion, but if you squint at their product, it's not exactly revolutionary architecture. Take the foundation model du jour, add some legal document search, wrap it in enterprise-friendly UX, and hire a sales team that understands law firms.

That's... pretty much it. But Harvey was there early, building relationships and integrating into workflows while everyone else waited for the technology to get better. When GPT-4 became good enough for legal reasoning, Harvey already had the distribution and trust to capitalize.

I suspect this pattern will repeat across most AI verticals. The winners won't necessarily be the companies with the most sophisticated AI research. They'll be the ones that figured out how to stay alive and maintain market presence until the models improved enough to actually solve customer problems.

Which makes Icon's pivot look less like desperation and more like pattern recognition. Instead of burning through runway trying to make impossible technology work today, they're building a sustainable business that can benefit when AI video generation stops being terrible.

But boring implementations with good distribution beat revolutionary technology with no market presence. Just ask Windsurf, which was losing badly to Cursor in the AI IDE wars but hired GTM teams, targeted enterprises instead of developers, and managed a $2.4 billion acquisition despite the product deficit.

The Service-to-Software Pipeline

Icon's agency pivot isn't defeat, it's a template. Start with high-touch services using AI to get efficiency gains, charge prices that actually support a business, then gradually automate select pieces. Much more honest than pretending full automation exists today.

The revenue model arbitrage is crucial here. Going from $20 monthly self-serve to $3,000 monthly with human support isn't just about making the product work; it's about finding a business model that can grow a business if we don’t see massive instant improvements. The SaaS dream dies, but the services reality might pay for survival.

The Problem with Levered Beta

The strategy has limits, and the AI SDR space is hitting them hard. As tools for sending cold emails get easier, executives are getting flooded. Inboxes are more spammed, harder to stand out in, and results are getting worse. This creates a death spiral for companies like 11x AI, which managed to go from zero to $10 million ARR despite a product every user hates, a CEO forced to step down, and multiple scandal publications.

If LLMs get 10x better at writing emails that don't sound like they were composed by someone trying to sell you timeshares, the entire AI SDR category will print money. But if inbox saturation kills email effectiveness first, even the early movers might not survive to see the technology catch up.

What This Means

The levered beta strategy works until it doesn't. Companies betting on future model improvements need three things: enough money to survive, enough attention to maintain mindshare, and a business model that can actually generate revenue while they wait.

Icon figured this out. They're not trying to revolutionize advertising anymore; they're trying to build a sustainable business that can benefit when AI video generation stops being garbage. Smart pivot, boring execution, potentially very profitable outcome.

Big investors are buying this “unlisted” stock

When the founder who sold his last company to Zillow for $120M starts a new venture, people notice. That’s why the same VCs who backed Uber, Venmo, and eBay also invested in Pacaso.

Disrupting the real estate industry once again, Pacaso’s streamlined platform offers co-ownership of premier properties, revamping the $1.3T vacation home market.

And it works. By handing keys to 2,000+ happy homeowners, Pacaso has already made $110M+ in gross profits in their operating history.

Now, after 41% YoY gross profit growth last year alone, they recently reserved the Nasdaq ticker PCSO.

Invest for $2.90/Share

_{Paid advertisement for Pacaso’s Regulation A offering. Read the offering circular at invest.pacaso.com. Reserving a ticker symbol is not a guarantee that the company will go public. Listing on the NASDAQ is subject to approvals.}