Our approach to age prediction • Hacker News Board

Re: Our approach to age prediction

#231

Post by BatteryMountain » Thu, Jan 22, 2026, 7:58 AM UTC

terminalshort wrote: post #179

BatteryMountain wrote: post #172

Do not give your biometric & photos of ID's or videos of your face to these companies. Nether to third-parties. The potential failure modes here are very high risk and not worth it. Better to unsubscribe and let them know why.

What is the risk? Anyone who wants a picture of my face can already get one by googling my name and going to my linkedin profile.

Some dumb companies and gov entities use national ID numbers (social security numbers for americans)as secret identifiers or grants access if presented. It typically has place of birth, date of birth, full names, gender, and face pic. In most cases this is enough to commit a ton of different kinds of fraud or hijack certain accounts (especially social engineering).

In my mind, national Id's (and the extra meta data of the person) should be public and only used for identification, not for authentication or authorization. Meaning there needs to be two or three extra steps after providing it to allow a transaction to occur. This needs to be a legal requirement for companies if they enter into contracts with a person.

If we need to prove we are not-minors or authenticate we are real or authorize access of personal information, the government should provide an api to auth the request, since they are the issuer of the document (the ID), so only they actually have the means to prove you are real and you are above 18. This can allow for a company to ask the gov, "is this person real and is this person above 18", and the gov shows me the request (otp, ussd, email, OS popup etc) to confirm the request and to select what info that company can pull. So its is 3 legged system, no third party companies involved. If the gov wants to create these constraints, they need to be the ones to provide the mean to authenticate (both for the consumer and the company). Also, when the gov shows the request to the user/citizen, it need to show exactly what the company is asking for and the full details of the company and the human representative that is making the request (almost similar to OAuth).

The problem runs much deeper than just "Whats the risk, my facepic is public already". Oh and this has nothing to do with minor and wont protect them in any way - only way to protect them is to take internet access away. The internet is not a child-friendly place and wasn't built by or for children. We should not bend to make it child friendly as it will destroy the internet in the long term.

Re: Our approach to age prediction

#232

Post by BatteryMountain » Thu, Jan 22, 2026, 8:07 AM UTC

astura wrote: post #180

BatteryMountain wrote: post #172

Do not give your biometric & photos of ID's or videos of your face to these companies. Nether to third-parties. The potential failure modes here are very high risk and not worth it. Better to unsubscribe and let them know why.

My face is not private information and probably hundreds of other people's cameras capture pictures/videos of me/my face every day. I hate age verification as a concept and I wouldn't personally go through it to use chatgpt, but "failure modes here are very high risk" is unnecessarily alarmist.

So pointing out that putting an open bucket of fuel next to an active fire is indeed a hazard and you shouldn't stand close to it, is unnecessarily alarmist? Until it happens to you or your family mate. Identity theft and financial fraud can ruin your life and make it very hard to recover and you will be at the mercy of credit bureaus to help get it sorted as most law enforcement (and courts) in most countries to not give a damn about such issues - you will be bounced around as nobody has the true means to help. Making people aware of these issues is not alarmist.

Re: Our approach to age prediction

#233

Post by Dilettante_ » Thu, Jan 22, 2026, 11:07 AM UTC

sodafountan wrote: post #227

Dilettante_ wrote: post #122

Earlier quoted context omitted.

>It was dead-on with its answer. It guessed 30-35. I'm 32. Horoscopes must feel like literal magic to you.

I'm not familiar with Horoscopes.

They're all about saying things that apply broadly enough, in a way that makes them seem specific, in order to make people go "Wow, that's dead-on!" An example:

  You’ve recently been feeling a quiet tension between wanting stability and craving some kind of change. On the surface, things look mostly under control, but there’s a sense that one small adjustment—something you’ve been postponing—could shift your mood more than you expect.

Re: Our approach to age prediction

#234

Post by sodafountan » Thu, Jan 22, 2026, 9:25 PM UTC

Dilettante_ wrote: post #233

sodafountan wrote: post #227

Earlier quoted context omitted.

I'm not familiar with Horoscopes.

They're all about saying things that apply broadly enough, in a way that makes them seem specific, in order to make people go "Wow, that's dead-on!" An example: You’ve recently been feeling a quiet tension between wanting stability and craving some kind of change. On the surface, things look mostly under control, but there’s a sense that one small adjustment—something you’ve been postponing—could shift your mood more…

That doesn't make any sense.

I only asked ChatGPT to guess my age because I'm assuming OpenAI is going to have the LLM assume your age going forward, which is an interesting use of the technology. Rather than ask up front, it just guesses based on the utilization of the tool. I understand "dead-on" may have been the wrong use of the term, let's just say it was fairly accurate...

I probably would be impressed by a good parlour trick or an accurate horoscope. Lmao

Re: Our approach to age prediction

#235

Post by rockskon » Sun, Jan 25, 2026, 3:45 AM UTC

asdfaslkj353 wrote: post #126

rockskon wrote: post #104

Earlier quoted context omitted.

95% of what is posted on the Internet is placed with intent to harm? What?

If you consider Advertisement and News (with understanding that it is rarely unbiased) harmful, the 95% is not that far off the truth.

And you think their internet was to harm?

I'd argue the overwhelming majority was agnostic at best to harm being a factor.