Hacker News Reader

Comment by vorpalhex | Hacker News Reader

Neat idea. Lovely interface.

However, no mention of monetization strategy. Are you selling my personal data? Do you show me ads? Do you intend to sell subscriptions?

Also, no E2E _and_ it's not open source/self-hostable. I need at least one of those. Either I don't need to trust you because things are E2E encrypted, or you establish trust by letting me see and verify code.

View on HN

Thank you! I really appreciate it.

> However, no mention of monetization strategy.

Thanks for bringing this up. I was hoping to give a brief introduction and then to field questions.

I will definitely not be selling personal data, or data otherwise. I do have some monetization ideas, and I'm really excited to try them out. They're all going to fit nicely with the core concept of the platform: periodicity. One of them is advertising, and there are others. If I can find good features to sell with subscriptions, I will. The important thing at this point for me is to demonstrate value, and then to monetize without disappointing or alienating users.

> Also, no E2E

Pony is currently a one-person startup and the focus has been on fleshing out the concept, building solid apps for desktop and mobile, and getting those apps into the hands of users. Privacy is something I take incredibly seriously on a deeply personal, ideological level. While I consider myself a really solid full stack developer and believe that Pony's infrastructure is secure, I do not have the requisite experience with cryptography to honestly represent to people that this is a private platform. I hope that with enough traction and some investment, I'll be able to hire an expert to help add privacy features that I can advertise in good faith.

> While I consider myself a really solid full stack developer and believe that Pony's infrastructure is secure, I do not have the requisite experience with cryptography to honestly represent to people that this is a private platform.

It would be great if this existed as an infrastructure-level service. Something like the Signal “API” that allows apps like yours to be built on top.

It could easily be a modified version of Matrix or XMPP (omemo). The protocols, security, and audits are already there. Sending delay should be easy to add on top without hurting security.

What you're describing is exactly what Matrix is aiming to be.

> I do not have the requisite experience with cryptography to honestly represent to people that this is a private platform.

Fair. E2E isn't trivial. However, platforms generally need to be built privacy focused from the ground up. It's hard to go back to a platform and add these things back in.

I would encourage you to add these features early, with the appropriate warnings until an audit can happen. "Hey, we're trying to keep all your message end-to-end encrypted but we haven't had our implementation audited yet. Thanks for being an early user!"

> I hope that with enough traction and some investment, I'll be able to hire an expert to help add privacy features that I can advertise in good faith.

Now is a really good time to learn. Seriously. libsodium is an excellent example of a library which provides user-friendly APIs that don't require you to roll your own crypto and is very strongly audited by experts — see https://libsodium.gitbook.io/doc/public-key_cryptography/aut... for an illustration. There are lots of other similarly good libraries. All you need to do is to implement some public key infrastructure, making sure that private keys stay private and never leave the user's device and that you can look up the published public key of another user.

Or they could just not do it because it's not that important to most people... HN comment sections are not representative of the general populace.

Do you know this with reasonable certainty or are you assuming this to be the case? Either way, it’s arguably better to design defensively and build in better security than you think you will need up front rather than have worse security and be in the headlines for the wrong reasons later.

I'm assuming it, and I'm certain my assumption is right. 99% of people have no idea what encryption is, so there's no way they could care about it. So while I'm not "utterly certain", all anyone would have to do to be as certain as me is go ask, like, five people who aren't engineers about it.

I don't think this is true anymore. The average user may not know what end-to-end encryption is, but they know they don't trust Facebook. And Apple built a whole marketing campaign around user privacy.

I think most people do care about privacy, but they're usually powerless to defend it.

Yeah I think you are right, more and more people are becoming very sensitive to this, and I am, too, personally. But as I wrote in a parent comment, I'd rather get it right later than wrong now. It's a big thing to misrepresent.

The average Facebook user.. does... trust Facebook, enough to use it. That's why Facebook is a multi-billion-dollar company.

Seriously, sometimes I wonder if people on this site have ever interacted with people outside of tech.

Yes! Strongly agree with this. E2E is easier than you think and will be much harder to add after the fact.

I just kind of think that encryption is one of those things that's best done by people who really know a lot about it. Like, I think it's okay to program if you don't know everything about data structures and algorithms, but I do not think it's okay to represent like your app is cryptographic secure when you're not extremely experienced and well versed in the field. Even if you're not rolling your own crypto but using someone else's. The extent to while I'm comfortable using someone else's crypto is LetsEncrypt with an integration for the K8s nginx ingress controller. But I will not personally warrant/represent that my at-rest or E2EE solution is sound. I would be uncomfortable with that.

Monetization ideas: Fancy "envelopes", "wax stamps/monograms/watermarks".

Charge $1 (per message) as a penalty for express "courier" delivery.

Even if the author claimed communication was E2E encrypted, you'd still need to trust that they did it properly and weren't intentionally siphoning data.

I don't think most developers are malicious. I do think a lot of companies, small and large, cut corners and sometimes fail to safeguard data well. That's one of the reasons E2E is good even when I generally trust the developer - it limits the data leaks that can happen in the first place. Marketing can't accidentally siphon all user data to a third party via analytics sdk, etc etc.

> Are you selling my personal data?

From the FAQ:

> How does Pony protect my privacy?

> We collect and store the minimum amount of data possible. We do not sell or share data with any third parties.

Reply on news.ycombinator.com