THE $6 BILLION WAITLIST

In May, Brett Adcock raised $700 million at a $6 billion valuation for a company called Hark. At the time, Hark had about 70 employees, no product, no customers, and no demo. The pitch: a personal AI that listens, sees, remembers you, and takes action on your behalf.

Adcock is not a grifter. He took Archer Aviation public and built Figure AI into one of the most valuable robotics companies on earth. The investors writing those checks include Nvidia, Intel, Qualcomm, and Salesforce. These are serious people making a serious bet.

Forbes ran a piece explaining why a $6 billion valuation with no product “actually makes sense.” The argument: the money isn’t buying a product, it’s buying a moat. Frontier compute, scarce researchers, a founder with a track record. Lock up the inputs before anyone else can.

On August 5, Hark shipped its first artifact: a browser agent called Handoff, in technical preview, behind a waitlist. In the demo it ordered flowers from a florist’s website. The benchmark numbers are Hark’s own. No outside reviewer has been able to verify them.

Here’s the part I can’t get past.

I run a personal AI operating system today. It’s real, it works, and it did a full morning of work before I finished my coffee: triaged my inbox so I never saw the junk, ran my Monday pipeline checks, pulled my weekly numbers, watched the things I’ve told it to watch, and started drafting this issue. It remembers everything we’ve decided, it runs on a schedule whether I’m at the desk or not, and it gets smarter about my business every week.

It cost me a subscription. Not $700 million.

Here’s the difference that matters: Hark’s $6 billion is buying vertical integration. Their own models, their own hardware, their own everything. My system is built on the opposite bet: it owns no model at all.

Every routine, every skill, every piece of memory in my operating system is a plain file in a repository I control. The intelligence is rented from whichever frontier lab is best this quarter. If a better model ships next month, I change the plug. I don’t rebuild the house.

Most people get this wrong because they think personal AI will arrive as a product you buy. A device. An app. A waitlist you join. So they wait.

But personal AI doesn’t arrive as a product. It arrives as a system: your priorities, your routines, your standards, written down where an AI can execute them. The model is the engine. The system is the car. And you can build the car right now with engines that already exist.

The moat Hark is buying, the compute and the researchers, matters enormously if you’re training models. You’re not training models. You’re running a business. You don’t need to own the power plant to run the factory.

So here’s my honest read on the $6 billion: it’s not crazy. It might even work. But it’s a bet that you’ll wait two years for their version instead of building yours in two weeks. Every operator who builds now makes that bet a little worse.

I packaged this exact build so you don’t have to wait for anyone’s waitlist. It’s called the One-Window Setup: one document that interviews you and builds your personal operating system around your answers. $99, and the files it builds are yours. kerzie.ai/one-window

Wade Kerzie

Founder, Kerzie AI Solutions

kerzie.ai