Hacker Newsnew | past | comments | ask | show | jobs | submit | er0k's commentslogin

wow I am so surprised to hear once again how JWTs are terrible

https://www.howmanydayssinceajwtalgnonevuln.com/


Someone not verifying the signature at all is not a mistake where you can blame the JWT spec itself.


They did verify the signature, and it was correct according to the "none" algorithm.


They edited the payload first but signature was never changed. A JWT's signature changes if payload changes; so it was never about the "none" algo, it was that Microsoft never validated the JWT with their signing key.

JWT is a great tool, Microsoft just failed to use it correctly.


Argh, I missed that it actually uses the "none" algorithm. Yeah, the existence of that option is extremely dumb and it shouldn't be possible to use that. I misread the post and thought it was a regular JWT, but they simply didn't validate it.


“Works as designed.”


JWT is complicated.

Complexity is a spec failure in security issues.

It's that simple.


Other commenters are suggesting you can’t blame the spec for end implementation mistakes, except that’s one of the many issues - JWT being so error-prone is a problem.

I use JWT just for handling of tokens, because it’s so well supported, but I won’t use it for anything more than token storage _because_ it is so vulnerable to mistakes.

The fact that mistakes are so easy to make is indicative of poor design in the spec itself.


Does this extend to OIDC? I’m not knowledgeable on the topic but it uses JWT right? Is it also prone to poor implementation? If you just error on alg=none does that solve it?


Idk if that's not too much of an oversimplification, maybe more like JWTs are an indicator/enabler of architecture level bugs?


(YC 2013)


We need a black bar on HN not just whenever a prominent computer scientist passes away, but also when a YC company behaves unethically.

But it's too impractical. It'll probably be on throughout the year.


https://einaregilsson.com/redirector/

Include pattern:

  ^https?://x.com/(.*)
Redirect to:

  https://xxcancel.com/$1


> There’s no quid pro quo here.

This actually sounds like the literal definition of quid pro quo.


or just use a yubikey... it costs ~$50


This comment could use some elaboration. For those that don't know you can use a Yubikey that supports PIV as a smart card for logging into macOS and performing a range of admin authentication operations with just the PIN, not just in the GUI but sudo as well (and of course more directly for SSH etc). It's not a perfect substitute, no ApplePay, but it means you can have a long complex password and only need a 6-8 digit PIN for most usage while still being pretty safe, and has some positives of its own in a multiuser or machine environment. It's a very reasonable option to consider IMO, even though yes it'd absolutely be nice if Apple did better on the hardware auth front.


Your comment just reminded me I can use my Yubikey to unlock my Linux laptop but hadn't set this up yet. I've been typing my password each time like an animal! Thanks. :-)


Doesn’t really work with Apple’s own stuff. At least I never got it to work.


That’s not a biometric TouchID. Any touch by anyone will trigger it.


Yubikey Bio exists. It’s $100 though so, whether it not it’s worth integrating into MacOS vs a first party keyboard is going to be subjective


This is nothing new. Wifi signals have been used to detect objects, people and animals, gait analysis[1], read keystrokes[2], monitor breathing and heart rates[3], "hear" conversations[4], etc for at least a decade now.

https://news.ycombinator.com/item?id=12353605

https://www.theatlantic.com/technology/archive/2016/08/wi-fi...

https://archive.is/XnHUV

1: https://ieeexplore.ieee.org/document/7457075

2: https://dl.acm.org/doi/10.1145/2789168.2790109

3: https://archive.is/mFSDq

4: https://archive.is/sNVcM


Have you gotten any of these to work? A few years ago I was tasked with investigating these kinds of techniques for a client (it was something cool and benign but I can’t say what due to NDA) and the big papers people are referring to when they mention this all had either huge asterisks or huge methodological flaws.


In my experience aallll of these fancy "we can measure things that sound impossible" papers come with the asterisk "in perfect lab conditions".

> “The signal is very sensitive to the environment, so we have to select the right filters to remove all the unnecessary noise,” Bhatia said.

AKA "it barely works and we had to filter the signal to the gills to get anything at all".

It's a really impressive tech demo but the article is selling it as if this might actually work in the real world and it clearly won't.


Getting it to sort-of-work is fairly easy. Getting it to work well on off-the-shelf hardware without a precisely controlled environment is hell.

For practical applications right now, you'd want a dedicated radar unit at 24GHz or so, probably with two separate reception paths too.

Eventually, we might get usable radar functionality in default Wi-Fi chips with 5GHz/6GHz Wi-Fi and MIMO - but it's not there yet.


i get asked about stuff like this from time to time and i always say "no, that's impossible" because i have ethics. The common retort is "well, i heard it was being used at <x>." and a client never contacting me again, which is fine.


Try using this you just need an esp32 devkit nodemcu style pcb. It measures movement though.

https://github.com/espressif/esp-csi


Indeed, this same principle has been shown to work with sound waves and not just RF waves. There was a paper a few years back that used car speakers and the microphone to be able to detect the number of people in the car for the purpose of detecting children or pets left in hot vehicles.


Generally I think this is called "tomography" (https://en.m.wikipedia.org/wiki/Tomographic_reconstruction) i.e. the reconstruction of higher dimensional data from lower dimensional data. Your brain can do it automatically in a lot of cases. For example if you see the shadow of a rotating cube on the wall, your brain can reconstruct 3d information about the cube even though you only have access to a 2d projection


[edit: publicly announced] commercial deployment into homes and offices is new.


Or is it?



Thank you!!!


congrats guys, this new feature looks really cool :)



I just use rsync, works great


Would you explain how you are using rsync to backup photos on an iPhone ?

I'm not aware of an iOS app named "rsync" and ... presumably you don't have a shell ... ?


Don't know if still works, but this is how I did it back in the days I had an iPhone:

https://pilabor.com/blog/2022/01/access-and-recover-files-fr...

Nowadays I'd use immich or ente.io, which has and e2e encryption cloud as well as self-hosted setup


You can load the IPhone’s pictures on Linux using the folders namespace and then rsync from the loaded namespace to /local…


But does it work for Live Photos which is the main premise of this app?


I use a usb cable and a Linux laptop to copy to a couple of external hard drives (which I store separately). It’s all manual but not too orrenous although i ought backup more often. The biggest hassle is accessing the new heic format on pretty much everything.

Could it all be made into a sd card image for a pi zero perhaps? Even with a web ui accessible over Wi-Fi? A basic cheap sync-cable-appliance that non-techies can easily use?


I think I heard that the latest 'rclone' now, finally, support iCloud ... but I think there is an issue there because you can't sync full quality / RAW photos to icloud, can you ?


rclone has support for iCloud Drive, but it sounds like you are talking about iCloud Photos, which is a different service.

This project lets you download from the latter: https://github.com/icloud-photos-downloader/icloud_photos_do...



Consider applying for YC's Winter 2027 batch! Applications are open till November 2.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: