When Claude Code is working on something, it shows you one small word while you wait, changing every few seconds.
Discombobulating, then Percolating, then Baking.
If you write software with AI now, you've probably stared at it for hours this month without once thinking about it.
Two weeks ago an engineer named Andrew McCalip looked at it and saw a billboard.
He built an extension that drops an advertiser's message where the spinner used to be, and pays you, the person watching, half the money.
You leave the extension running, let an ad you will never click sit where the loading word used to be, and keep half of whatever the advertiser pays, which is to say you get cut in.
What McCalip discovered is that you can get thousands of people to build your business for free, as long as you let them feel like they own a piece of it.
He sold the feeling of being cut in, and the ad space was only the wrapper it came in.
By Kickbacks' own numbers, more than 16,000 people earned something, the average lifetime earning came to about five dollars, and most never crossed the ten dollars you needed to withdraw anything at all.
On the other side of that ledger, the half paid out to developers crossed eighty thousand dollars, which means McCalip's half ran to tens of thousands too, in a single week.
16 thousand people generated a real business and were paid, on average, the price of a coffee, while one person kept the rest.
Stranger than the size of that gap is that almost nobody on the short end of it felt cheated, and the reason they didn't is the whole point.
They had agreed to a fifty-fifty split, and fifty-fifty sounds like the fairest arrangement in the world, half for you and half for me.
The word doing the work in that sentence was split, because a split tells your brain you are on the inside of the deal, and the moment your brain believes that it stops asking the only question that matters, which is half of what.
Half of almost nothing is almost nothing, and yet it doesn't feel like almost nothing, it feels like being a founder, and that feeling is what people were being paid in.
Given humans and their greed, people started faking the watching so aggressively that McCalip had to ban accounts in bulk and cap earnings at $5 an hour.
This brings me to think that a product that pays you for your attention barely paid anyone, and still couldn't tell real attention from fake.
It almost feels like attention has no price, because anyone can mint it after seeing this.
In every kind of advertising that has ever worked, those were two different people.
A company paid a newspaper to reach you, and you had no particular reason to want the ad to exist.
Your attention meant something precisely because you weren't being paid to give it.
Kickbacks puts both jobs in one person.
Some might argue otherwise saying this gives you a reason to watch it too, but then look at the examples of folks who’ve done this in the past such as Brave browser / Veera browser.
I would personally keep the ads open and walk away to make coffee or do something else.
They are essentially buying our willingness to sit in front of a screen.
If you pay for a signal, people will produce the signal instead of the thing it was supposed to indicate, every time, until the signal means nothing at all.
It’s important to remember that it works (at least in the beginning).
Everyone thinks it’s the new cool thing and gets hooked onto it because they agreed to a 50% split, and that made everyone feel they’re equal partners.
With advertisers like the ones below (well funded and relevant ICP using claude), a new ad format (not the boring cringe videos, but a quick quirky line pitching you instead of boring claude works, it grabbed people’s attention).
The advertisers who came in early, got results too (that CTR is industry leading).
Will these CTRs last is another question.
At the end of the day, you’re just inventory for Kickbacks, and your enthusiasm will become greed as that number stays just a coffee beyond a certain threshold of time.
This is much older than software, and once you notice it you can't stop noticing it.
A casino sells you the feeling of being about to win and lets you supply the losses, and a multi-level marketing company sells the feeling of running your own business to people who buy its inventory and carry the risk it would otherwise bear.
Brave has paid its users in a token to watch ads for years, and what that produced was a small economy of people optimising for the reward rather than the attention an advertiser wanted, the watching reduced to a chore you tolerate to collect.
What AI changed is the scale and the speed.
McCalip built the whole thing in a weekend, around a day job, and a feeling that used to need a salesforce and a stadium of recruits now ships as a browser extension and reaches tens of thousands before that weekend is over.
But, I want to be fair to him, because the thing he built is clever, but the cleverness everyone credits is in the wrong place, since the ad is ordinary and the code simple.
The rare thing he found is that the feeling of partnership has become the cheapest input a business can buy.
The next time someone offers to cut you in, run one test before you let yourself feel anything.
Ask what you would be making if the thing succeeded wildly, in rupees, not in percentage.
An absolute number is built to inform you, which is why the people structuring these deals lead with the percentage and go quiet on the base.
If the honest answer to half of what is a number you would be embarrassed to say out loud, then the feeling is the whole of your compensation, and you are the inventory.
Reply and tell me the last time someone offered you a cut, and whether, looking back, you ever worked out half of what.
Nandini
PS: If you do vibecode, do it to dream big the way Andrew did. And launch it even if it’s half baked. He ended up going through multiple steps just before he was even allowed to pay these people, but that’s what got him the visibility: FUN.