x1gma

It's not surprising per se, but it's something that people should be more aware of. And a lot of this consumption is not providing global services (like the Google search or workspace suite) but the whole AI hype.

I didn't find numbers for Google or Microsoft specifically, but training ChatGPT 4 consumed 50 GWh on its own. The daily estimates for queries are estimated between 1-5 GWh.

Given that the extrapolation is an overestimate and calculating the actual consumption is pretty much impossible, it's still probably a lot of energy wasted for a product that people do not want (e.g. Google AI "search", Bing and Copilot being stuffed into everything).

Chrome cookies are encrypted, for exactly the reasons stated. If malware gains access to your system and compromises it in a way that DPAPI calls can be replicated in the way Chrome does it, then your sessions will also be compromised. But this is way harder to do, and at least prevents trivial data exfiltration.

Please don't take personal offense, but you have merely a project scaffold with an unrealistic goal that will be blocked and C&D'd into the ground, without any other projects created.

It doesn't matter how hard you're working on your anonymity, this project will be ripped apart by a horde of lawyers in seconds. You're not only doing something questionable or against ToS, you're directly attacking and sabotaging their monetization. This will not be taken lightly by the legal team of reddit.

You want to provide a better, cooler, more robust and other random buzzwords API than the own of reddit. So, you alone, want to provide a better API than the whole team of reddit does for their absolute core product, all by scraping. This is simply not realistic.

While we're at the topic of monetization, scraping, ETL into your own model and providing the API - for the amount of content that reddit has (quantity, not quality) this will be a highly resource intensive task. How do you plan to fund that, since your API will be better than the official one, I can expect at least the same performance as well, right?

And also, most importantly, even if you magically achieve working around all that and get that working - why? Who is your expected user group? Pretty much every software using reddit moved away from reddit or simply has died. AI gen content is rampant, and most discussions seem like bots talking to bots. There is literally nothing to gain from an API to reddit - so why would anyone bother using it?

The third option is to use the native secret vault. MacOS has its Keychain, Windows has DPAPI, Linux has has non-standardized options available depending on your distro and setup.

Full disk encryption does not help you against data exfil, it only helps if an attacker gains physical access to your drive without your decryption key (e.g. stolen device or attempt to access it without your presence).

Even assuming that your device is compromised by an attacker, using safer storage mechanisms at least gives you time to react to the attack.

Yes, in your head, and in your second factor, if possible, keeping derived secrets always encrypted at rest, decrypting at the latest possible moment and not storing (decrypted) secrets in-memory for longer than absolutely necessary at use.

Been a few days since using electron, but AFAIK electron can't be used as a wrapper for android apps, or can it? Or is their android app a web app wrapped into a "native" android app too?

Also, since this seems to be an issue since 2018, 6 years should be plenty to rewrite using a native secure storage...

Kinda expected the SSH key argument. The difference is the average user group.

The average dude with a SSH key that's used for more than their RPi knows a bit about security, encryption and opsec. They would have a passphrase and/or hardening mechanisms for their system and network in place. They know their risks and potential attack vectors.

The average dude who downloads a desktop app for a messenger that advertises to be secure and E2EE encrypted probably won't assume that any process might just wire tap their whole "encrypted" communications.

Let's not forget that the threat model has changed by a lot in the last years, and a lot of effort went into providing additional security measures and best practices. Using a secure credential store, additional encryption and not storing plaintext secrets are a few simple ones of those. And sure, on Linux the SSH key is still a plaintext file. But it's a deliberate decision of you to keep it as plaintext. You can at least encrypt with a passphrase. You can use the actual working file permission model of Linux and SSH will refuse to use your key with loose permissions. You would do the same on Windows and Mac and use a credential store and an agent to securely store and use your keys.

Just because your SSH key is a plaintext file and the presumption of a secure home dir, you still wouldn't do a ~/passwords.txt.

How in the fuck are people actually defending signal for this, and with stupid arguments such as windows is compromised out of the box?

You. Don't. Store. Secrets. In. Plaintext.

There is no circumstance where an app should store its secrets in plaintext, and there is no secret which should be stored in plaintext. Especially since this is not some random dudes random project, but a messenger claiming to be secure.

Edit: "If you got malware then this is a problem anyway and not only for signal" - no, because if secure means to store secrets are used, than they are encrypted or not easily accessible to the malware, and require way more resources to obtain. In this case, someone would only need to start a process on your machine. No further exploits, no malicious signatures, no privilege escalations.

"you need device access to exploit this" - There is no exploiting, just reading a file.

It might be, but to be fair, that's what the glorified autocompletion is actually good for, if it's actually used a supporting tool, and not to pump out quantity over quality.

Man, the disclaimer at the bottom that Business Insider is partnered with OpenAI to allow them to train on their articles is really the cherry on top.

Check your blocklist and keep in mind that YouTube is testing server side ads muxed into the media streams, which will not be blocked by traditional adblocking techniques.

If you can not install anything, your only choice is probably to set up a pihole or something similar on your network.

Edit: Some models seem to have advanced settings, where you can change the DNS - you could try to use adguard dns servers, or any DNS adblocker you want to use.