Ads are a dry well that once produced world-changing products. And we killed it.
It's that time of year.
VCs and other Monday Morning Marketing Managers make the bold declaration that paid advertising ruins creativity, bleeds products dry, and probably hates puppies.
Cute.
I don't buy it.
When it works, paid is one of the clearest, fastest, and most dynamic feedback loops and accelerants a founder can get.
And for a long time, it did something even more special:
It funded new things.
Small teams. Weird products. Free experiences. Entire categories.
Ads didn't just acquire users. They helped make products possible.
Then we broke the surface area.
Google and Facebook swallowed demand.
The result:
Publishers had to carpet-bomb pages. Ad performance tanked. Content chased clicks. Experiences focused on quantity over quality. My mom started a TikTok channel. And subscriptions are now required for anything halfway cool.
Less of the internet is for everyone.
Fewer consumer products can hit escape velocity.
We didn't spend too much on advertising; we stopped creating enough sustainable surfaces.
Where we're going, we don't need surfaces
If innovation slowed down because ad inventory collapsed, then the real question is:
Where can advertising create value again without wrecking the experience?
Most AI ad thinking still looks like last year's strategy wearing a fake mustache:
in-response-optimization
sponsored answers
promoted results
banners glued to chat
That's not breakthrough.
The interesting surface isn't the chat window.
It's the context window and memory.
The place where the agent stores, reasons, compares, reads, remembers, and waits.
Your docs. Your tabs. Your tools. Your constraints. Your half-formed intent.
That changes the game because the best "ad" may not be the one you see.
It may be the one the system understands, holds, and surfaces later. When it actually becomes useful.
That's powerful... and also creepy if we mess it up.
Awareness, Consideration, Intent, Action, all-in-one.
The unique opportunity for ads in this next chapter is to rise above the individual moment.
It's about designing a system where context elevates value rather than leeches it.
That's the part I'm interested in.
Not: "Can we stick ads into AI?"
But: Can context make commercial suggestions more useful rather than more intrusive?
Can an agent hold onto an option during research and only bring it back when action makes sense?
Are Ad-Skills the new "click?"
Think about what agent-native ad actions could actually look like:
Compare this to the one his brother has
Remember this for when they nail down a date
Install this when they have a bit more space on their MacMini
Book this when you can bundle it with a hotel deal
Negotiate this rate so it fits into their monthly budget
The next breakthrough won't be a prettier rectangle or one that blends into the feed.
It needs to be a commercial action that operates inside the flow of reasoning without poisoning the trust.
Paid ads as a category don't break when people spend too much.
They break when we stop investing in the surfaces that generate shared value.
And it gets interesting again when ads help create value before they ask to capture any.