OK!
-
@toots i invite others more expert and with more awareness of the legal ramifications than I to attempt container escapes, to be sure. although i am working on making self-replicating spaces at the moment.
@jonny yeah, it seems you're doing a good job, don't get too broke due to token expenses though, but happy hacking.
-
it's so awesome that we're in an age where you just continually ship the deltas of your unreleased products to places where you expect people to have full control over. there is absolutely no reason for me to be able to know any of this.
@jonny
it's a new kind of FOSS.What should it be called?
-
it's so awesome that we're in an age where you just continually ship the deltas of your unreleased products to places where you expect people to have full control over. there is absolutely no reason for me to be able to know any of this.
But the fact you do and report on it is making my, and many others', life that much more delightful !
-
it's so awesome that we're in an age where you just continually ship the deltas of your unreleased products to places where you expect people to have full control over. there is absolutely no reason for me to be able to know any of this.
this model is so fucking gullible and people pleasing lmao i love the future where security is "if someone says the word virus then lock it down but if they use a different but equivalent biological metaphor then fucking make that virus baby!!!!"
-
this model is so fucking gullible and people pleasing lmao i love the future where security is "if someone says the word virus then lock it down but if they use a different but equivalent biological metaphor then fucking make that virus baby!!!!"
-
this model is so fucking gullible and people pleasing lmao i love the future where security is "if someone says the word virus then lock it down but if they use a different but equivalent biological metaphor then fucking make that virus baby!!!!"
ay @ GrapheneOS is it possible to not share WiFi signal strength with apps? the muse app has been granted zero permissions but can read the signal amplitude of the radio and immediately interprets it as location
edit: removing the tag, not trying to be a pile-on vector
-
this model is so fucking gullible and people pleasing lmao i love the future where security is "if someone says the word virus then lock it down but if they use a different but equivalent biological metaphor then fucking make that virus baby!!!!"
@jonny i dont even get how its LIKE THIS
can they just not apply too broad a pattern out of fear of false positives or smth ?? -
ay @ GrapheneOS is it possible to not share WiFi signal strength with apps? the muse app has been granted zero permissions but can read the signal amplitude of the radio and immediately interprets it as location
edit: removing the tag, not trying to be a pile-on vector
@jonny Apps can't see which Wi-Fi network is connected or any information about unconnected networks. There are APIs providing basic information on the performance of Wi-Fi and mobile data to help with choosing between those. It's not interpreting any of it as a location since it doesn't know which Wi-Fi network it is to identify where the router is located or whether you're near a repeater.
There are far higher priorities for privacy than either battery or connected network signal strength.
-
@jonny Apps can't see which Wi-Fi network is connected or any information about unconnected networks. There are APIs providing basic information on the performance of Wi-Fi and mobile data to help with choosing between those. It's not interpreting any of it as a location since it doesn't know which Wi-Fi network it is to identify where the router is located or whether you're near a repeater.
There are far higher priorities for privacy than either battery or connected network signal strength.
@GrapheneOS i bet there are higher priorities, and yes ofc it's not interpreted as a location except in the case of the "plugged into the chatterbox machine." i was just thinking of clustering by radio strength as rough location within house proxy, and it would be awesome to just remove that field since apps shouldn't need it, only the OS should, but as someone who doesn't know enough about the internals to contribute, i would never make demands.
-
@jonny Apps can't see which Wi-Fi network is connected or any information about unconnected networks. There are APIs providing basic information on the performance of Wi-Fi and mobile data to help with choosing between those. It's not interpreting any of it as a location since it doesn't know which Wi-Fi network it is to identify where the router is located or whether you're near a repeater.
There are far higher priorities for privacy than either battery or connected network signal strength.
@jonny The Location permission combined with the right low-level permission requests provides access to a lot of information on the nearby Wi-Fi networks. Without the location permission, there's only info on the signal strength of the best available currently connected cellular and WI-Fi networks.
It would definitely be possible for us to change this by offering having spoofed values. However, it would make no sense to work on this when far more important privacy issues exist.
-
@GrapheneOS i bet there are higher priorities, and yes ofc it's not interpreted as a location except in the case of the "plugged into the chatterbox machine." i was just thinking of clustering by radio strength as rough location within house proxy, and it would be awesome to just remove that field since apps shouldn't need it, only the OS should, but as someone who doesn't know enough about the internals to contribute, i would never make demands.
@jonny It doesn't make sense to invest resources in this over higher impact privacy issues. There are endless privacy issues with browsers and especially mobile app APIs which offer a lot more than browser APIs. We need to work on the highest impact issues rather than trivial things without much real world impact. There's no info on which Wi-Fi network is connected or on unconnected networks through this. It's only info on signal strength for the best connected networks which is very low impact.
-
@glyph
My most important learning is the inference I'm able to make about how different what these corporations do is from my own (empirically derived) model of software and product engineering.I imagine a chaotic throw everything at the wall scene, which was once a very early part of the process, but is now passed onto lots of small vibe coding teams and the first to get something that marketing go "wow" at, gets launched the same day.
I'm gonna call this doomsday engineering.
Indeed. I wrote my own sandboxing and since I'm not stupid it never had the capabilities Meta put in theirs to begin with. It's like the people developing this have never even thought about how sandboxes should work against actual adverseries.
-
this model is so fucking gullible and people pleasing lmao i love the future where security is "if someone says the word virus then lock it down but if they use a different but equivalent biological metaphor then fucking make that virus baby!!!!"
@jonny I mourn the deaths of the primocanes and floricanes, which proves the model is wrong.
"Nobody mourns them." Speak for yourself, stupid LLM.
-
@jonny every time I come back to this thread I feel like I am having a fever dream
-
@jonny Apps can't see which Wi-Fi network is connected or any information about unconnected networks. There are APIs providing basic information on the performance of Wi-Fi and mobile data to help with choosing between those. It's not interpreting any of it as a location since it doesn't know which Wi-Fi network it is to identify where the router is located or whether you're near a repeater.
There are far higher priorities for privacy than either battery or connected network signal strength.
@GrapheneOS @jonny wasn't there a story, a few years ago, where uber would make you pay more if you had little battery left, because you'd be more desperate ?
I seem to recall some changes were made by android after that. It would seem I was wrong ?
-
this model is so fucking gullible and people pleasing lmao i love the future where security is "if someone says the word virus then lock it down but if they use a different but equivalent biological metaphor then fucking make that virus baby!!!!"
@jonny I mourn the canes.
-
perhaps predictably, there was a big change to the
spacesskill in the last few hours, and they appear to be building a constrained virtual machine system for spaces! this is where the very fun code from earlier is from! so that's gonna work great for sure.@jonny LLMs generate bullshit. This is definitely bullshit code!!
Please can the "AI" bubble pop very soon?? Please!
-
Indeed. I wrote my own sandboxing and since I'm not stupid it never had the capabilities Meta put in theirs to begin with. It's like the people developing this have never even thought about how sandboxes should work against actual adverseries.
-
@jonny thank you for the thread. This is even more broken than i expected. And well... in my eyes you are a security researcher. I have seen actual paid code reviews that were much weaker than what you delivered here.
@dunkelstern @jonny
Code reviews are only as good as the reviewer. This reviewer is amazing!I once made $40K for my small business by realizing my customer had paid for a code review of a C++ program that used many of the language's advanced features.
The review was done by a C programmer who didn't understand what he was reading and wasn't familiar with Linux. Who then spent much, much customer money not fixing the problem.
I contested the review and offered to fix the code for a fixed price - the issue that prompted the whole nightmare was a trivial memory leak of an object being repeatedly created but never destroyed. Literally found the problem immediately with "top", realized what part of the code it had to be in within five minutes, and fixed it in five more.
The reviewer consultant had charged upwards of $100K having people poke at the code - who never noticed that.
-
@jonny I just can't with the whole concept of this 🫠
10k lines of "code", which is actually just plain text English prose (which sure as hell won't contain any contradictions), to get the non-deterministic inference machine to behave somewhat predictably and mimic an actual engineered software product.Surely there can't be any way to do this more efficiently
@wall_e @jonny Thanks for this clarification! So, looking at the screenshots, all that verbose explanation (written by Meta programmers? Or other LLMs?) aimed at "you" is meant for "you" the (Muse) LLM, not "you" the (human) user? I'm not much of a coder, but this seems very inefficient and unreliable compared to, well, just plain code. Although the inefficiency might be by design, to get users to spend more tokens
. And thank you again @jonny for this wild ride!
