[Avipat_]

We launched Kled 6 weeks ago as a data marketplace for everyday people. Users upload personal content like camera rolls, dashcam footage, homework, POV recordings, and original music, and we pay them for it.

While AI companies scrape public data to train their models, we’re building a platform that compensates users and gives AI labs access to licensed, high-quality data.

The product took off quickly. We did over 20 million impressions on Twitter, more than 10,000 people signed up, petabytes of data were uploaded, and we’ve paid out over $300,000 to users.

Every file uploaded is automatically classified and made searchable so AI companies can instantly license the exact datasets they need.

Over the next five weeks, we’re doubling down on our labeling operations and hiring top AI engineers. We’ve raised over $2 million in venture capital at a $40 million valuation. High pay and high equity. Email [e-mail]

[lampiaio]

No one cares about the aggregated sum you spent, and something tells me that instead of directly saying how much the average user was paid, the PR team chose to divulge the total amount because it sounds better (someone must have pat themselves on the back when they phrased it as "we paid users $300K").

[Avipat_]

the money came directly from AI Labs paying for it. not venture or anything else.

[lurk2]

That isn’t the issue. If you paid $300,000 to 300,000 users, each user made a dollar.

[Avipat_]

We paid $300,000 to 500 users, there's over 10,000 on the waitlist still. Each of them have uploaded over 200GB of data. Some at the 1TB limit.

[lurk2]

That’s a better ratio than I would have expected, but I would have led with that. The title as it is written reads as a bait and switch. I initially read it as: “We paid a plurality of users $300,000 each for all of their data.”

It’s probably obvious to most readers that this isn’t what happened ($300,000 would be an absurd price for someone’s camera roll), but it feels misleading. People respond to that kind of thing really negatively here, which is probably why your post got flagged.

[minimaxir]

Why did you omit that it's a cryptocurrency? https://x.com/useKled/status/1932778949051838827

[pogue]

Oh man, sounded promising until discovering it was a rug pull.

I wonder if user data does have some value. I could see long haul dashcam footage possibility being worth something if you were training self driving cars...

[codingdave]

Job postings go here: https://news.ycombinator.com/item?id=44757794

[amelius]

Wait a minute, if users upload their camera rolls, doesn't that mean that in most cases these also contain footage of __other__ people?

Sorry, but wtf?!?

[Avipat_]

Correct but the person who took the pic/vid owns the rights for it legally speaking. This content is already being trained off of, scraped off YouTube, Instagram etc. All Kled does is put the money back into the pockets of the people.

[echoangle]

> Correct but the person who took the pic/vid owns the rights for it legally speaking.

Not really, there are a lot of ways to take images/videos you aren't legally allowed to publish later. You can have them for yourself but you can't share them publically (or give to a company like yours).

[Forgeties79]

Do they? That’s not an assumption you can make. I’ve likely got plenty of things in my phone I’m not allowed to have just by accident

[amelius]

Yeah, this sounds like just another sleazy way of a tech company to shove the liability to the user.

[lurk2]

This will never fly in the EU.

[Avipat_]

working on it

[Forgeties79]

I’m also curious about this. I’m not having a strong reaction yet because I’m just not quite sure what to make of it, though at first glance it certainly has me going “wait what?”

I hope it’s not just some boiler plate legalese about how you have to make sure you have the rights to everything you upload.

[Avipat_]

It’s not just boilerplate. We actively verify uploads using multiple signals: EXIF metadata, timestamps, device signatures, and visual similarity checks. If a file wasn't captured by the uploader’s device or shows signs of being scraped/copied, it gets flagged or rejected.

In short, if it's not your original content, it won’t pass. And we’re constantly tightening this loop as we scale.

[Forgeties79]

I think I’m less worried about whether or not it’s my original content and more about whether or not what I shot is ok to sell

That being said I did not consider that side of things, and I think it’s great you’re so proactive about it. Hopefully that does not come off as condescending/patronizing!

[gyanchawdhary]

Wow, you’ve cracked the mystery of photography .. sometimes people take pictures of other people. What’s next, discovering that mirrors reflect more than just yourself and that’s a valid wtf moment in your world view too …

[amelius]

The law has limits on what you can do with your camera and photos. Until now, most people were operating within these limits, so there was no problem. In the modern times of big tech, automation, AI, and large scale photo databases, you simply cannot scale up your simple user's rights without running into the law. And imho, rightly so.

[Avipat_]

lol yea, not sure what that guy is on

[amelius]

Read the rest of the comments in this subthread, then maybe you will understand what this is about.

[Avipat_]

Can't reply to ur other comment but yea, Latham and Watkins represents us. The best lawyers on the planet, so yea we've talked to lawyers about it lol.

[j_w]

That's great. Good thing its explicitly illegal to distribute sexual material of another individual without their consent. But it's completely legal for you to own that material with their consent.

Your blanket claim that you own the rights to commercially distribute all photos that you took is just not true.

[gyanchawdhary]

HN is the UFC of capitalism .. blood, code, sweat and capital. You showed up barefoot with a manifesto about non violence … Wrong arena champ.