Ask HN: Why aren't AIs being used as app beta testers yet?
13 points by amichail 16h ago 17 comments
Ask HN: How would expose a scam involving a powerful figure?
10 points by soueuls 15h ago 8 comments
Facebook is starting to feed its AI with private, unpublished photos
211 pier25 138 6/28/2025, 12:08:28 AM theverge.com ↗
I remember complaining like hell when the wall came out, that it was the beginning of the end. But this was before publicly recording your own thoughts somewhere everyone could see was commonplace, so I did it by messaging my friends on AIM.
And then when the Feed came out? It was received as creepy and stalkerish. And there are now (young) adults born in the time since who can't even fathom a world without ubiquitous feeds in your pocket.
Call me nostalgic, but we were saner then.
Nothing is a social network anymore.
Everything is a content-consumer a platform now.
People just want to scroll and scroll
Unfortunately parasocial behavior is good for engagement.
You need to already know someone to find them here.
Check out the waitlist!
https://waitlist-tx.pages.dev/
Edit:
Here are some rough layout designs https://drive.google.com/drive/folders/1uLwnXDdUsC9hMZBa1ysR...
It's intentionally simple
Perhaps it already exists but I have thought about writing something that takes what is provided by "Download your data" and produces a local SQLite database, a local webpage, local website or some combination thereof that is served from the user's computer instead of Meta servers.
However I do not use Facebook enough to justify the effort, and when I do I never look at the "feed".
A really private place with only people that matter.
If the posts are more long form, what is the difference between this and a blog where the "secret code" is the URL?
Or even a finsta account currated the way you want.
I don't say these as a "it's not gonna work" as in consumer its about the experience, I genuinely wonder why the experience will be better
> Group chat Group chats work when everyone in one know each other. I have N different circles which don't overlap so group doesn't chat makes sense. Messages in group chat are more "in the face" - everyone has to. I just wanted a place where I can dump my thoughts without feeling like seeking immediate attention.
> Blog PostX is indeed something like a private blogging space. It's something I wanted for myself.
Honestly I am not fully sure how it's going to be used by people but I have built something me and my friends like and use.
No "People you may know" or "select at least N interests or follow N accounts to continue".
I think early adopters will invite their friends to join and that is the only way.
Got any suggestions?
But we are not interested turning into Facebook. You will only see posts of your friends and nothing more.
I was spending 8+ hours a day doom scrolling which led to this idea. I just want to see what my friends want me to see and that's it.
It means really a lot to us.
We are working on a better name and the site!
I'll send you the welcome email manually soon!
It’s easy to dogpile. I’d like to see more proof, that’s all. “It’s obvious” doesn’t cut it for me. For one, we have major societal problems that are being exposed through these platforms, and the mere knowledge of the problem has a negative impact on the individual. Do we shut the platform down because it’s showing us things we don’t want to see, or do we fix the societal problem? And many others.
It was also one of the first to drop genuine user-sercing features like the old timeline (just all the posts of people you followed which you came there to see) which it replaced with the algorithmic feed which recommended stuff you never asked for or wanted.
Instagram did keep that feature though until 2 years and still has it although it's constantly switching it off.
Social media is dead to me.
They will get to decide what to do with their likenesses when they're older. It seemed cruel to let Facebook train a model on them from the time they were babies until they first start using social media in earnest.
It took the rest of us much longer to realize they were right.
In Holland we have a saying, what do you bring it your house is burning down? And most people said my photos. This was before the digital age and cloud obviously. We take photos because we care. Stuffing them into everyone else's face has also been a thing at birthday parties but outside that not so much.
First, for most of us in daily life, once you know you are being photographed you exit any context you were in and enter the new “I am being photographed” context. In some important way, you are stolen from the world around you for a period of time. Your body is still present, but you might be thinking about how this all would look at any later time. This was not a problem when a photograph was a specifically arranged event. This of course is not an issue if you are unaware of being a subject photographed, but there may be other concerns about that.
Second, a photo/likeness of you is a proxy allowing other people to relate to you. Keeping in mind that we only ever relate to images/models that we build of each other in our minds (we have no “direct access” to other people), in this case a photo is a shallow (there is little other information than appearance) but weirdly high fidelity (for sighted people) model of you. This is not an issue if the photo is kept just by people you know (common in analog era) or after you are dead, but otherwise if published[0] it means people can somehow relate to “you” without the actual-you knowing or having met them. Some people may feel some sort of satisfaction from this, others it can make uncomfortable.
(“Photo” can be generally swapped with “video”.)
[0] Now when generative models start to be trained on what we thought is our private photos, the idea of “published” is blurred.
wasn't the camera doing the stealing, but the holder of the photo (facebook in this case)! And it wasn't the soul being stolen, but money!
Basically every kindergarten, primary school and high school will want to post pictures.
Here (NL) we get a form at the beginning of each school year to mark which uses of photos we find acceptable. E.g. we allow photos in the school portal (which is private and not owned by big tech), but not on Facebook, etc. It's the way it should be done, because there is not much burden on the parents. If the school also wants to put photos on social media, the burden should be on them to make sure that kids for which they don' have an ack are not put there.
A bit harder was initially convincing my parents not to put pictures of their granddaughter on Facebook. They are understandably proud and want to show their friends. But they respect it.
I think in all her life there has only been two violations of our policy. In both cases we contacted the person who published the photo/video and they took it offline.
You just need enough 'weirdos' to make it normal. I know that there are other parents that agree, but not everyone has the gut to stand up to the social media tyranny, but will join if some people set an example.
LinkedIn was used in a similar manor, to coordinate meetups for our local Cloud Native meetups, but the LinkedIn algorithms are much much worse than Facebooks, so people would get "You might be interested in this meetup" two weeks after the event.
Facebook basically took over communication, no more mailing lists, no more updates on the website, if there even is a website. You just have to accept Facebook if you want to be notified about changes in scheduling, upcoming events or general information about your kids soccer practise.
Also in many places WhatsApp is practically a requirement for daily life which is frustrating. What I need is some kind of restricted app sandbox in which to place untrustworthy apps, they see a fake filesystem, fake system calls, etc.
GrapheneOS comes pretty close to that I think? You can put such apps in a separate profile and cut off a lot of permissions. You can also scope contacts, storage, etc.
I’m grateful though. We would have called meta malware back when.
Limited Access to Your Library
"App" can only access the items that you select. The app can add to your library even if no items are selected.
I.e. messenger.com is possible to use if you request desktop version, change font size and deal with all sort of zoom issues. Of course fb doesn’t support actual calls or notifications just because, so I don’t use it.
Instagram is even sneakier - you can’t post stories via mobile to “close friends”, post videos or view them from instant messages.
And on Android they're not even the worst privacy player which is Google of course
Description of the latter from the uad list:
Why is this even always running on a pristine Samsung, etc. phone? Creepy.The company that is destroying children's mental health with phone addiction is developing VR glasses.
I guess nobody cares
As an aside, there was a discussion a few days back where someone argued that being locked in to popular and abusive social/messaging platforms like these is an acceptable compromise, if it means retaining online contacts with everyone you know. Well, this is precisely the sort of apathy that gives these platforms the power to abuse their marketshare so blatantly. However, it doesn't affect only the people who choose to be irresponsible about privacy. It also drags the ignorant and the unwilling participants under the influence of these spyware.
I should sit down and try something like postmarketOS or Mobian as a portable Linux machine is what I really want ...
For example, let's say that you avoid a certain abusive messaging platform. But what if your bank or some other essential institution insist on using it to provide their service? We can complain all they want. But they will probably just neglect you until you concede in despair.
To fight this, you need affordable and ethical alternatives for the device, platform and applications. You also would need either regulation or widespread public awareness. Honestly, the current situation is hopeless on that front.
i give it a year or less.
Yesterday.[1]
[1] https://www.cincinnati.com/story/news/2025/06/26/jd-vance-me...
The tweets just saying “drug use” and then you hear it’s weed is ridiculous. Why wouldn’t they just state that they lied on their immigration form about drug use?
In other news, FB has been using whatsapp metadata to coordinate genocide campaigns in Gaza. What’d all those dead civilians (including infants) do, again?
Presumably they signed a TOS, so it’s OK.
"Meta tells The Verge that, for now, it’s not training on your unpublished photos with this new feature. “[The Verge’s headline] implies we are currently training our AI models with these photos, which we aren’t. This test doesn’t use people’s photos to improve or train our AI models,”
As someone who is familiar with the ML space, it seems unlikely that the addition of private photos will significantly improve models, as you have mentioned.
I saw this line in the article: "Meta tells The Verge that it’s not currently training its AI models on those photos, but it would not answer our questions about whether it might do so in future, or what rights it will hold over your camera roll images."
It would seem important to share this with people who may 'not read the article'
> On Friday, TechCrunch reported that Facebook users trying to post something on the Story feature have encountered pop-up messages asking if they’d like to opt into “cloud processing”, which would allow Facebook to “select media from your camera roll and upload it to our cloud on a regular basis”, to generate “ideas like collages, recaps, AI restyling or themes like birthdays or graduations.”
> By allowing this feature, the message continues, users are agreeing to Meta AI terms, which allows their AI to analyze “media and facial features” of those unpublished photos, as well as the date said photos were taken, and the presence of other people or objects in them. You further grant Meta the right to “retain and use” that personal information.
The straightforward explanation is this: they have a feature where it is helpful to group people together. For instance suggesting a photo of you and a friend to be posted on their birthday. In order to make this work, they need to perform facial recognition, so they ask for permission using their standard terms.
Can they train their AI with it? Yes, you are giving them permission to do so. Does the information available tell us that is what they are doing? No, it does not. In fact, a Meta spokesperson said this:
> “These suggestions are opt-in only and only shown to you – unless you decide to share them – and can be turned off at any time,” she continued. “Camera roll media may be used to improve these suggestions, but are not used to improve AI models in this test.”
— https://techcrunch.com/2025/06/27/facebook-is-asking-to-use-...
Could they be lying about this? Sure, I guess. But don’t publish an article saying that they are doing it, when you have no evidence to show that they are doing this and they say they aren’t doing this.
Might they do it in the future? Sure, I guess. But don’t publish an article saying that they are doing it, if the best you have is speculation about what they might do in the future.
Does it make sense for them to do this? Not really. They’ve already got plenty of training data. Will your private photos really move the needle for them? Almost certainly not. Will it be worth the PR fallout? Definitely not.
Should you grant them permission if you don’t want them to train on your private photos? No.
This could have been a decent article if they were clearer about what is fact and what is speculation. But they overreached and said that Facebook is doing something when that is not evident at all. That crosses the line into dishonesty for me.
However, I wish they’d grow a pair and just outright block the FB and other similar dependencies that make such stuff necessary.
Maybe this will finally convince people to throw out their smartphones.
The non-paywalled TechCrunch story shows the consent screen that people agree to before Facebook uses the photos in this way: https://techcrunch.com/2025/06/27/facebook-is-asking-to-use-...
I encourage everyone to look at that screenshot and decide for yourself if the media coverage is reasonable here.
I bet "agree to" is "we clicked the box for you anyway"
It's surprising(not) how that class of error always seems to fall on the side of Facebook grabbing more data without consent, and never on the side of accidentally increasing user privacy.
I’m sure if you log the Facebook app’s network traffic on your phone and show that it uploads photos without you clicking on the agree button, they’ll happily publish an article about your findings.
Nah, that's the company's reputation that appends malice in your mind to an innocent headline
Nothing on that screen says they’re using your photos for training. I’m sure it’s in the linked terms, but Facebook knows those won’t be read.
This isn’t buried. The user has to see the screen and click accept for their photos to be uploaded.
Compared to the usual buried disclaimers and vague references to “improving services,” consenting to 1000 things when you sign up for an account, this is pretty transparent. If someone is concerned, they at least have a clear opportunity to decline before anything gets uploaded.
It’s just surprising to me that people look at this example of Facebook going out of their way to not do the bad thing and respond with a bunch of comments about how they doing the bad thing.
I didn’t make that assertion. I think most people don’t care if their photos are used to train an AI model as long as Facebook doesn’t post the photos publicly. Fundamentally, I care if people see my photos, and don’t care if computers see them. But I’m aware some people dislike AI and/or have strong beliefs about how data should be used and disagree. It makes sense to give those people an opportunity to say no, so it seems like a good thing that the feature is opt-in rather than an opt-out buried in a menu.
Here’s the thing. Even if we grant your idea that maybe this is more understandable, why would that be reasonable? Facebook employs a lot of very smart people and has enormous resources. I’m confident they could come up with wording that would make this very clear to everyone. I mean, “we will use your photos to build our next generation AI systems” is a lot clearer than what they have here, and I just came up with that on the spot. That they haven’t done so is a deliberate choice.