Cookies? Who doesn’t want cookies??
Rose Kaz / Oct 7, 2026 / 9 min read
Picture the warmest kitchen you can imagine: a fresh batch just out of the oven, the smell of butter and vanilla hitting you before you even find the source. Nobody asks what you want or how much you can pay. Someone pulls up a chair, places something sweet and warm in your hands, and says, sit down, tell me about your day. No price tag. No form to fill out. No permission box to check before you're allowed inside.
The kind the internet gives you is nothing like anything Grandma would bake.
Instead, these not-so-sweet treats pop up on virtually every website we open: that small box, that brief moment when we click something we haven't fully read, and hand over something we didn't know had a name.
Here's what we've been giving away: our browsing patterns, our questions, our decisions, our instincts, our professional knowledge, our HI.
Human intelligence: the piece of this we need to talk about before we go any further.
Here is the piece the tech industry continues to gloss over, and for that matter many of us everyday users of the internet: artificial intelligence learned from us, all of us. There are currently more than 6 billion of us online, spending an average of 6 hours and 36 minutes connected every single day. And where there are gaps in the data set, the current tech rules (spoiler alert: there are very few) allow for data brokers to grab our data faster than you can close a pop-up.
That data is grabbed, scraped, scoured and ultimately sold to create the AI models everyone is talking about as the way of the future; to either fear or fan over. Either way, it's the HI that makes these AI machines work. That same AI trains on our data: your writing, my questions, our arguments, your late-night searching, my creative output, your professional judgment. All of it.
It is HI that AI eats for breakfast.
Everything we put into the internet, for years, before most of us had any real understanding of what "data" meant or who it belonged to, is what AI must have to be even baseline effective and operational.
And here is the real kicker: until very recently, the history of those being published, either on the world wide web or into the books that AI is consuming every single day at the equivalent of 35 Library of Congress collections per major model, roughly an estimated 303,000 books every single hour, has largely not included research, reporting, or history of many others than a very slim demographic. And that small segment has not only been more published, accredited and acclaimed for "their ideas," they've also received the payment for it. I'll let you use your own HI here to infer who I'm talking about. One clue: many of them look very similar to each other, attend similar retreats and may even shop at the same stores. There are 5 leaders at the current time of this article's publication who are said to be leading technology "forward."
Most of us will never set foot in those rooms. And still, the dynamics those rooms produce shape every room the rest of us do walk into.
I keep showing up in rooms like that, virtual and IRL. More than 2,400 conversations deep, since 2017, and the pattern holds steady. Our ideas, our market reads, our pattern recognition, and our early warnings about rooms we called correctly before anyone else did. The knowledge is present. The credit, the infrastructure, and the ownership are not.
The knowledge is present. The credit, the infrastructure, and the ownership are not.
That isn't an observation about one industry. That's the structure.
The pattern is older than any room I've walked into.
Much of the genius that has propelled humanity forward came from people who were never appropriately credited, and in many cases were never given the occasion to voice their discoveries in their own words. Hedy Lamarr patented frequency hopping, the technology foundational to WiFi and Bluetooth. Katherine Johnson, Dorothy Vaughan, and Mary Jackson calculated the trajectories that put men in space. Mileva Marić, Einstein's first wife and his academic equal, contributed to the foundational relativity work that carries only one name. Rosalind Franklin's X-ray crystallography revealed the double helix; Watson, Crick, and Wilkins received the Nobel. Lise Meitner co-discovered nuclear fission and was left off the prize that followed. This is not ancient history. This is the archive AI trains on. I take big issue with that and I hope you do, too.
The word for what we're missing is data sovereignty.
Data sovereignty is not a tech concept. It's a property rights concept. It's the question of who owns the intelligence you produce, and whether the systems you participate in treat that knowledge as a contribution or a resource to mine.
The Electronic Frontier Foundation has spent years documenting how data brokers buy and sell behavioral profiles built from exactly this kind of collected HI. The FTC's reports on data brokers lay out how the industry works in detail. Neither one makes the headlines they deserve.
The cookies pop-up is where they fawn "asking permission." The fine print is where they bury what permission means and what you are actually giving away.
I keep coming back to this image of Grandma's house, and it isn't about nostalgia.
It's warm when you arrive. It feeds you when you're hungry. It asks what you need and then produces it. It doesn't track how long you stayed or sell that information to a third party. You leave with more than you came in with.
There was a version of the internet that worked like that. Forums where people share professional knowledge because sharing is the point. Wikis where someone writes a careful explanation of something obscure, because someone else might need it. Message boards where experts teach freely.
That version didn't disappear slowly. In 2012, as Web 2.0 and cloud computing moved from early adopters into the mainstream, Facebook crossed one billion users, Google's Panda and Penguin updates rewarded commercial content over community knowledge, and the incentive to share freely gave way to the incentive to monetize attention. What had been a commons became a storefront. The platforms that replaced the forums were designed to capture and extract, not collaborate and contribute.
The version we have in 2026 takes more than it gives. That's absolutely not an accident. It's the blueprint for late-stage capitalism's worst-behaved actors.
And while the version of info sharing and knowledge exchange still exists in pockets, the dominant architecture we're operating inside now has changed. Trust, if you and I don't choose to participate in this next build of the internet, we will again be left out of the history books, even if they are only digital copies.
Sure, we can call basic participation on the internet a form of digital activity, but I'm calling for you to do more than just be a consumer of the data that is sold back to you. Every click, every scroll, every second of attention is a transaction where we have become the product.
And so if the big tech companies want our attention and moreover our data to keep getting our attention back from us (very meta, I know!), we need to realize what participation in a very quickly evolving tech world requires of our beloved HI: that cherished human intelligence that AI needs to even exist.
The exchange is no longer knowledge for knowledge, what I like to call Sharing is F*cking Caring. It's attention for access, and that access is very much one way where we get to give our data and then buy it back.
The cookies pop-up is the very passive-aggressive moment when the tech overlords pretend to care and "ask permission." Most of us click accept and move on. My kind and firm nudge here is that you don't just click and accept.
Hot take: if you reject cookies, most sites still work exactly as intended. The only thing that changes is the site's ability to track your activity and sell it back to you. California's CPRA gives US residents the right to opt out of the sale of their personal information. Under GDPR, EU and UK sites are legally required to function whether or not you consent to non-essential tracking. The FTC holds federal oversight authority over deceptive data practices. The tracking stops. The site stays.
Which is exactly a reason to understand what we're opting into.
Let me be 100% clear: I'm not anti-technology. In fact, I'm absolutely positively pro-transparency about how it works, and I've made that the center of how I build. If you want to understand exactly what I use and why, that page exists because I think we deserve the full picture: not the curated version, not the one that skips the part about where the training data comes from.
If you want the longer answer about why I use AI the way I do, that's a different conversation, and a more complete one.
The question isn't whether we should use AI. The question is: on whose terms?
That question is one I am actively working toward answering.
What I can tell you right now is this: we understand how these systems work, and that gives us a fundamentally different relationship with them. The gap between that understanding and simply being a user is the one I'm working to close, one conversation at a time. And if you want to talk through what that looks like for your specific work, that's what working with me looks like.
Ready to Start
Using AI on your own terms.
Two courses are currently open. Both are virtual. Both are designed for people who are ready.
Teach Me How to Claude
General AI foundations for anyone ready to stop watching from the sidelines.
October 15 · 12:00pm–3:30pm PST / 3:00pm–6:30pm EST
RegisterGet NAR Compliant
For real estate professionals navigating the new National Association of Realtors requirements around AI tools.
October 22 · 2:30pm–6:00pm PST / 5:30pm–9:00pm EST
RegisterHere's what I'm building toward.
In addition to my consultancy, where I guide women-led businesses and teams to adopt AI safely and strategically, I continue to build toward creative uses of technology that leverage our knowledge as a service (KaaS), where we will finally be paid our value for our contributions to society.
Business 4 Good is a peer-to-peer cooperative for women in business. It's being built with the express intention that the intelligence women share inside it, the knowledge contributed to the directory, the patterns that surface from all those conversations, belong to them. Not to the platform. Not to whoever funds the platform. To the women who produced it.
That is the alternative to the cookies model.
It's not simple to build. And it is the right direction.
The next post in this series is about something you may call intuition, but I have a hunch we've been naming it wrong. Subscribe below and get it directly in your inbox.
