Schools could teach cybersecurity to all children. Scanning faces and real-life IDs will fail to solve crime on the internet and will only succeed in destroying our privacy.
Remember when facebook was busted playing silent audio so the app could stay active in the background? I honestly don't think there's a company I trust less than facebook/meta. And it's because the rot is at the top, and has always been there.
>The other issue, that Facebook was running a silent audio stream in the background, is also called out. Grant says this was unintentional, and that it was not being used to keep the app alive — yet it did as a byproduct of the bug.
How much issue tracking and scrum management and engineering work and code review and testing and deployment and maintenance went into accidentally streaming silent audio that you only stop doing after you got caught and have to claim all that successfully tested and deployed work was unintentional, without ever explaining the actual innocuous purpose of streaming silent audio and paying for all that extra bandwidth?
I was just reading a comment on HN yesterday about how MacOS had so many bugs. I guess they don't have issue tracking, scrum management, engineering work, code review, testing, deployment, and maintenance either.
I hate FB, but not everything is always a sinister plan, although this could have been. I will repeat: "with Facebook, who knows."
There is a huge difference between accidentally shipping a bug, and accidentally shipping a fully functional distributed end-to-end silent audio streaming system that works at Facebook's scale. They are in no way comparable. All the intricate crosscutting interoperating client, server, and middeware components of an end-to-end Facebook scale silent audio streaming pipeline do not just accidentally emerge without anybody noticing and magically dovetail together perfectly because nobody was paying enough attention to detail.
Large distributed systems don’t spontaneously assemble themselves without anyone understanding what they’re doing or intending for it to happen.
>I will repeat: "with Facebook, who knows."
Apparently we all know the obvious except for you. And you're not living up to your user name, which perhaps should be "rationalizer" not "rationalist". Instead of begin rational, you're bending over backwards to implausibly carry the water for Facebook beyond all credulity. It's a bad look, and shows you're vastly underestimating the complexity of developing distributed streaming software at that scale. So I will repeat what rsynnott said:
>And if you believe that, you'll believe anything.
I heard that in Japan phones have an audible shutter sound. Not mandated by law. Though I think that having this in the law is very reasonable. Maybe EU can step up. Taking photos is more fun with the sound too.
These phones change your lock screen into their „content” screen after updates, so you need to periodically go to settings to undo it. Phone owners shouldn’t have to root to permanently get rid of this stuff and other „suggestions”.
How are the generated Haskell programs? I imagine much shorter than Go and easier to eyeball for correctness, but can’t say as I’m not fluent in it. LLM-generated procedural Python scripts are very readable in my experience.
Haskell is one of the tersest language in general. With "no comments" instruction code actually is almost idiomatic one. It's hard to guess it was written by LLM.
I do wonder about reliability. The only things I'm missing are the PTR record and reputation from what I gathered. Even if the mail server goes down, mail gets to me because email providers attempt to deliver again.
Anyway, I added a disclaimer at the top, so people don't treat this as a production ready setup.
