A little tired of seeing young people introduce ideas starting with their age. Your age literally doesn't matter, it doesn't make you a brilliant developer. Your product should speak for itself. Furthermore, being a teenage developer doesn't tend to signal intelligence as much as it tends to signal naivete.
What AI is powering the face recognition? Did you train it or is it an out-of-the-box product sold by someone else? Is their security up to snuff? Does your platform differentiate actual faces from images? Can it tell the difference between a human face and an AI generated face? What is done with the facial data? How is it stored? Are the "private" chats encrypted? To what standard?
You've told us how old you are, and you have a link to your product, but there's nothing to demonstrate that your product is safe to use, actually private, or doing anything different than just collecting more data to hand to AI.
Not a chance that I would use this. All it does is put more privacy sensitive data and pictures out there for AI to train on and use for fakes after services like these inevitably get hacked.
I was in your shoes when I was 17. I built a social app to replace Orkut's communities (which Google bought and killed) and a chat app before Discord took off.
Some advice:
- If you want to improve as a developer, keep building features, but if you want to improve as an entrepreneur, avoid that and focus on user growth first. Beware of spreading your small community too thin with new features. In my case, I allowed the creation of new groups too soon, and people became too dispersed.
- AI verification is tricky. If you can build an AI that detects AI-generated content, you could theoretically use that knowledge to build a better model that fools your verification (GANs). One strategy might be going invite-only to create exclusivity, though this has its own drawbacks.
- Social media is too broad a field. Focus on a niche of people who want to communicate but lack a platform. For example, one of my biggest successes came from announcing my chat app during a Dota 2 International tournament—I got 2,200 concurrent users within an hour (this was before Discord).
I've never learned as much as I did when I released my own apps. Even if they go nowhere, the knowledge you gain is invaluable.
First of all, good for you, spending your time building something rather than consuming. I see you have a lot of comments about "problems", and that is both typical of engineers and of the internet in general. So make sure you read this 10 times -- Good job going from nothing to something, it's a step people even twice your age struggle to accomplish. Keep going and the more you do it the better you will get!
I am not sure if fake people are a real problem for social media. Real Names ends up being pretty lame, people are more open and interesting when using a pseudonym. To me, it's clear that you understand the value because you are using a pseudonym here! White Lotus the App doesn't have a face, but it can still use social media productively.
What is annoying on social media are bots, reposts, commercial spam, etc. To me, it would make sense to go after undesirable content rather than undesirable accounts.
Safety is the real worry here. If you lose people’s PII it can be used for malicious purposes. You don’t want a dictator being able to take a picture of a protesting crowd and identify people based on that. Make sure you’re following best practices and trusting no one.
> You don’t want a dictator being able to take a picture of a protesting crowd and identify people based on that.
This is already a reality and while it’s a very valid concern, I wouldn’t be too worried about such a small scale project making the difference in that problem.
This is a really cool project for a teen and an ambitious goal, so congratulations!
That said, I think the goal and the approach are misguided. It's probably better to take this experience and use it on a different problem, or join forces with an organization that has more momentum.
1. How would you sustain this without ads? If you take subscriptions, that will hamstring growth. You'd be threading a needle even if your app was objectively the most secure, private, and authentic social network. To date, this is an unsolved problem.
2. There are other non-commercial social networks like bluesky or mastodon. Why wouldn't people use that instead?
3. Face verification is a hard limit for me and probably most other people who avoid commercial social media. I don't see any reason to trust a random new app with sensitive PII.
4. Is face verification even meaningful anymore in the age of unchecked generative AI?
Fake profiles are a problem when the site is big enough to attract bad actors. Yours is too small for that so you are solving a problem that does not exist yet. But you are asking people to their sacrifice privacy.
I think UK government would be amazed, especially now when they try to force identity verification wherever possible. Do you know what UK citizens do instead? They use VPNs to tackle that.
It's not a problem with people hiding their identities, as others have already pointed out. The internet in 90s and early 00s was a much nicer place, despite users hiding behind nicknames on chats and IRC. The real problems are bots and insatiable crave for attention on social medias.
Facebook requires video selfie verification[1] these days. I have not been able to register a new Facebook account without passing one. I was able to bypass it with DeepFaceLive though.
A little tired of seeing young people introduce ideas starting with their age. Your age literally doesn't matter, it doesn't make you a brilliant developer. Your product should speak for itself. Furthermore, being a teenage developer doesn't tend to signal intelligence as much as it tends to signal naivete.
What AI is powering the face recognition? Did you train it or is it an out-of-the-box product sold by someone else? Is their security up to snuff? Does your platform differentiate actual faces from images? Can it tell the difference between a human face and an AI generated face? What is done with the facial data? How is it stored? Are the "private" chats encrypted? To what standard?
You've told us how old you are, and you have a link to your product, but there's nothing to demonstrate that your product is safe to use, actually private, or doing anything different than just collecting more data to hand to AI.
Not a chance that I would use this. All it does is put more privacy sensitive data and pictures out there for AI to train on and use for fakes after services like these inevitably get hacked.
I was in your shoes when I was 17. I built a social app to replace Orkut's communities (which Google bought and killed) and a chat app before Discord took off.
Some advice:
- If you want to improve as a developer, keep building features, but if you want to improve as an entrepreneur, avoid that and focus on user growth first. Beware of spreading your small community too thin with new features. In my case, I allowed the creation of new groups too soon, and people became too dispersed.
- AI verification is tricky. If you can build an AI that detects AI-generated content, you could theoretically use that knowledge to build a better model that fools your verification (GANs). One strategy might be going invite-only to create exclusivity, though this has its own drawbacks.
- Social media is too broad a field. Focus on a niche of people who want to communicate but lack a platform. For example, one of my biggest successes came from announcing my chat app during a Dota 2 International tournament—I got 2,200 concurrent users within an hour (this was before Discord).
I've never learned as much as I did when I released my own apps. Even if they go nowhere, the knowledge you gain is invaluable.
First of all, good for you, spending your time building something rather than consuming. I see you have a lot of comments about "problems", and that is both typical of engineers and of the internet in general. So make sure you read this 10 times -- Good job going from nothing to something, it's a step people even twice your age struggle to accomplish. Keep going and the more you do it the better you will get!
While it’s impressive, we generally hate Face/ID verification on social media around here…
good point but the world can be free of fake profiles and it is like the new internet
I am not sure if fake people are a real problem for social media. Real Names ends up being pretty lame, people are more open and interesting when using a pseudonym. To me, it's clear that you understand the value because you are using a pseudonym here! White Lotus the App doesn't have a face, but it can still use social media productively.
What is annoying on social media are bots, reposts, commercial spam, etc. To me, it would make sense to go after undesirable content rather than undesirable accounts.
How do you expect to handle fake faces?
Do you expect the "Kingdoms" thing to increase or decrease tribalism ingroup/outgroup conflict?
Safety is the real worry here. If you lose people’s PII it can be used for malicious purposes. You don’t want a dictator being able to take a picture of a protesting crowd and identify people based on that. Make sure you’re following best practices and trusting no one.
> You don’t want a dictator being able to take a picture of a protesting crowd and identify people based on that.
This is already a reality and while it’s a very valid concern, I wouldn’t be too worried about such a small scale project making the difference in that problem.
Small scale projects can explode. Clearly the poster is ambitious! Just wanted to remind to have care as people would be trusting them.
This is a really cool project for a teen and an ambitious goal, so congratulations!
That said, I think the goal and the approach are misguided. It's probably better to take this experience and use it on a different problem, or join forces with an organization that has more momentum.
1. How would you sustain this without ads? If you take subscriptions, that will hamstring growth. You'd be threading a needle even if your app was objectively the most secure, private, and authentic social network. To date, this is an unsolved problem.
2. There are other non-commercial social networks like bluesky or mastodon. Why wouldn't people use that instead?
3. Face verification is a hard limit for me and probably most other people who avoid commercial social media. I don't see any reason to trust a random new app with sensitive PII.
4. Is face verification even meaningful anymore in the age of unchecked generative AI?
Guys saw a lot of comments pointing out concerns about AI face verification
this is how it works
1) user has only access to front camera while doing AI face verification and the ml only allows if photo that has one face
2) user is required to show a hand gesture while taking photo if not i gave clear guide lines to not trust them
3) the face data is stored in cloudinary and it is not being used to train Ai
4) and people who think fake profiles are not a big enough problem listen to me
internet will become a world filled with shape shifters that can take your shape and do bad things imagine that
Fake profiles are a problem when the site is big enough to attract bad actors. Yours is too small for that so you are solving a problem that does not exist yet. But you are asking people to their sacrifice privacy.
I think you need a better value proposition.
I think UK government would be amazed, especially now when they try to force identity verification wherever possible. Do you know what UK citizens do instead? They use VPNs to tackle that.
It's not a problem with people hiding their identities, as others have already pointed out. The internet in 90s and early 00s was a much nicer place, despite users hiding behind nicknames on chats and IRC. The real problems are bots and insatiable crave for attention on social medias.
> AI-powered face verification (1 account per person)
How do you prevent people using fake faces, as has already happened to similar efforts[0] and will get easier with face swap AI apps?
[0] https://www.windowscentral.com/gaming/gamers-bypass-uk-age-v...
Facebook requires video selfie verification[1] these days. I have not been able to register a new Facebook account without passing one. I was able to bypass it with DeepFaceLive though.
[1] https://www.facebook.com/help/875662883106240/
Great approach, but is this viable in the time of ai, like if someone try to fool using ai generated content
Why does your age matter?