uwait
Get paid while AI thinks
227 followers
Get paid while AI thinks
227 followers
Every time you use ChatGPT or Claude, you wait. Usually 3 to 8 seconds. That's it. That's the idea. uwait is a Chrome extension that shows a curated ad during that loading screen. You get a small cut. The publishers whose content trained the AI get a cut too. The advertiser gets your attention at a moment when you're actually paying attention. The split: 50% to users, 30% to publishers, 20% to us. We're picky about advertisers. Just brands that make sense in that context.
I think this is a pretty interesting idea. You noticed a small thing that millions of people do every day and found a way to make that time useful. My main question would be whether people care enough about the money to keep using it after the first few weeks. When someone checks their earnings after a month, do they feel like it was worth installing?
Buska
@reda_roqai_chaoui That's exactly the core retention question and I think about it a lot. The honest answer is I don't know yet, it's day one. What I do know is that even small amounts feel more rewarding than zero, and the bar isn't "life-changing money", it's "better than nothing for time you were already wasting." Whether that's enough to keep people after a month is something I'll only be able to answer with real data.
I'll share numbers once I have a few weeks of history.
The idea is really interesting and the fact that you aiming to be picky about advertisers is great. Do you already have some standards/ rules for them?
Buska
@julia_shtogren Right now I review every ad manually before it goes live. No advertiser runs without my approval. It's not a formal written policy yet, but in practice anything that feels off-brand for an AI-savvy audience gets rejected. The bar is: would I personally be okay seeing this while waiting for a Claude response? If not, it doesn't pass. I'll formalize the rules as the volume grows.
StartupBase
The wait was already there. You just put it to work.
The 50/30/20 split is unusually honest too.
How do you match which publishers content trained which model, or is that more approximated?
Buska
@attacomsian "The wait was already there. You just put it to work." Honestly the best one-liner summary of the whole thing, stealing that.
On the matching: it's exact, not approximated. When the AI cites sources in its response, those specific publishers receive their share of the revenue from that interaction. No pool distribution, no guessing. The citation is the attribution signal. It's not perfect since not every AI response includes explicit citations, but when it does, the link is direct.
Really clever idea. I use Claude and ChatGPT constantly throughout the day and never thought about those 3-5 second gaps as "time" - but they add up. The publisher revenue share angle is what makes this interesting beyond just another ad product. Curious how the payout amounts look in practice after a week of heavy usage. Installing now.
Buska
@galdayan Welcome, glad you're installing. Honest answer on payouts: it's literally day one so I don't have a week of heavy usage data to share yet. What I can say is that earnings are already accumulating for the first users, and heavy AI users are exactly the profile where it adds up fastest since the loading time compounds with every query. Come back in a week and tell me what you're seeing, that data will be more useful than anything I can project right now.
love the idea,
but how do you deal with such kinds of user who switch tab while model is thinking (using chatgpt respond alarm extension)?
Buska
@max_shin Thanks ! 💖 Good edge case. If you switch tabs, the ad keeps running in the background. You still earn, it still counts. The logic is that you triggered the query, the loading slot exists regardless of where your eyes are. Whether that's the right call long-term is something I'll revisit based on advertiser feedback, but right now it favors the user.
Wait, this is actually a crazy idea. 👀
One thing I’m curious about though, will users have any control over the type of ads they see? Because ads can go from genuinely useful to annoying or even harmful pretty fast.
Curious how you’re thinking about keeping that experience positive while still making the model work.
Buska
@mahedi_hasan_sayem No user-side controls for now, that's the honest answer. The quality filter is on my end: every ad goes through manual review before it runs. Nothing launches without my approval. The bet is that curation at the source is more reliable than giving users a settings panel they won't touch anyway. If patterns emerge around what feels annoying vs useful, I'll add controls. But I'd rather start strict and loosen than the other way around.
Buska
@kamedan_ Fair pushback. The bet is that you're already staring at a blank screen for 3 to 8 seconds doing nothing. The ad doesn't replace anything, it fills time that would otherwise be wasted. Whether that feels like a distraction or not probably depends on the person, which is why I'm watching retention and usage data closely. Youu should check annd try it ahah