[bastawhiz]

No, you really can't. Whatever verification you ask for can be faked (increasingly trivially with AI). You can't even rely on calling the users on video to verify their identity.

I run a startup hosting podcasts (that is, literally the file hosting). Advertisers with literally nothing to gain because inactive feeds are invalid will sign up for accounts and post spam. They literally have humans overseas sign up for Gmail accounts and go through the full signup and verification and captcha just to post some garbage links.

If they'll go through that much effort for my startup where there's literally zero payoff, they'll do it to yours. I promise you, you're not going to win the war on bots. Maybe for a while until you grow, but mark my words.

[adefee]

Appreciate the spirit of your comment and agree it's a big thing to tackle. We've worked hard on this aspect in particular, before anything else. Our core codebase is built around much of this (releasing more details closer to/at launch) - and we're using methods that have been proven to work in several large countries/industries (particularly in video games where botting & hacking is also a major concern). At the time the user creates an account, we know that they are a human and that they are relatively unique. We don't allow users to create multiple accounts.

[bastawhiz]

> and we're using methods that have been proven to work in several large countries/industries (particularly in video games where botting & hacking is also a major concern).

There are no games that have successfully blocked all bots.

> We don't allow users to create multiple accounts.

Unless you're requiring proof of identity using things like passports or driver's licenses, this is a rule, not a guarantee. You can't possibly know whether someone over the Internet is the same person as someone else. The crypto community has spent the last decade trying to crack this nut.

> that's risky because there's lots of ways to detect that, they'd get penalized, and then would not be able to make a new account.

You're assuming they want an account in the first place. They're not doing this because they want to have an account on your service, they're doing this because someone is paying them a few dollars a day to sit in a room and put on different hats or Photoshop a new name onto their citizenship card or whatever to bypass your verification so they can post a link to a sketchy casino. There's no risk, they'll just keep doing it. The big bosses will just cycle through warm bodies (of which there are many millions) to sign up for your service for pennies and post spam or election propaganda or whatever they care to.

> At the time the user creates an account, we know that they are a human and that they are relatively unique

I really want to underscore how much hubris is wrapped up in this comment. Your business model is predicated on doing something extremely difficult and presently unsolved (and likely unsolvable!) with essentially 100% success. Forever. I worked on this problem at Stripe for half a decade, it's not a small challenge.

Can you start checking people's passports? Yeah of course. But then you're going to fail because nobody wants to upload their passport to the account they're creating to shit post and message their friends. And your user acquisition costs are going to be enormous because you're doing passport scans on every signup.

[adefee]

We'll confirm more details closer to launch, but we have a way of doing what we need to do safely, while respecting user privacy. Our caq at launch is higher than some other networks (depending on whether you're comparing us to all networks or other upstarts), but the ltv of our users are inherently also much higher. Agreed we're making some big claims and we need to refine/add the details (as mentioned in a diff response this wasn't a self-post on HN, didn't find out til this morning, and we don't have all of our messaging ready for a real marketing push yet, lol).

I hear you, though, and while I believe we have the right stuff in place to tackle the issue, I hope when we release more info you continue to keep us honest and call us out if we're wrong :)