Hey if you’re colourblind, all blues can be blurple. And so can all purples!
Hey if you’re colourblind, all blues can be blurple. And so can all purples!
The requirement to not track users with cookies does not extend to cookies that make the site work in the first place, such as those which track your login session, or your refusal of other cookies.
It’s like any number of blog hosts that have gone before it.
Fair enough, I genuinely misread and thought that was within the quotation marks. But her message is still wrong because she is still talking about AI in general, but her argument applies only to a) AI whose data is derived from data scrapers like Facebook or b) AI put to surveillance tasks. That does not apply to Stable Diffusion, which is why I mentioned it, but it is caught by her assertion, “AI is a surveillance technology.”
AI absolutely has the potential to be used for surveillance; its use in facial recognition most obviously. But the person quoted in the article didn’t say “AI has the potential to be used for surveillance” - she said “AI is fundamentally a surveillance technology”. So if she’s not talking about LLMs and image generators, why is she saying that it’s a fundamental part of the technology? It’s not very fundamental if these two year-defining AI technologies aren’t included in it.
ChatGPT is not reliable enough to be worth citing. “Per ChatGPT” may as well be “per some bloke down the pub”.
Remarkably stupid take. It’s produced by big-data companies because you need a lot of data to feed it but that doesn’t make it “surveillance technology”. Stable Diffusion wasn’t trained on the kind of data she’s talking about, and it can’t be used to surveil you either. ChatGPT no more permits surveillance of its users than does chatting with a real person.
It’s spelled “copyright”, not “copywrite” btw :P
I know how HTTP works. These banners are supposed to (and are legally allowed to) store a cookie saying you have refused. Websites are allowed to store session cookies with displaying a banner at all.
No, they set a cookie to store it, but with a low retention period, so you get bugged again.
A start would be to require sites to remember non-consents for at least as long as they remember consents. Why do I have to be asked about cookies by every site every month?
Your browser can not save third party cookies, but it might break some sites. Some advertising situations allow the use of first-party cookies, and blocking first-party cookies will break most sites.
In either case you will still have to fill out the consent form, and if the consent is stored in the kind of storage you block, then you will have to fill it out every single time you visit.
Yes, but it often doesn’t work and even when it does the site is unusable while it works, which for some particularly awful banners is several minutes. The situation is worse on mobile where most people have a browser that you can’t install add-ons to (and I’m not sure if that one works in firefox mobile anyway)
It’s already the case that necessary cookies don’t need permission, but websites do not abuse this to not show the prompt. This is because the legislation has teeth.
Either too many people spamming so you can’t follow a single conversation, or for most channels you had 40 people idling and never responding, so it felt like a ghost town.
How is this different to Discord? You have huge, medium and small channels in both.
Shouldn’t the engineer be a bit more worried if the cable’s been cut?
If UDP drops packets it’s probably nothing. If TCP drops packets it’s because something’s actually wrong.
If you’re using TCP and losing packets you should be panicking though, because something is very wrong…
That’s the key I use as a compose key
An overarching principle of security is that of minimum privilege: everything (every process, every person) should have the minimum privileges it needs to do what it does, and where possible, that privilege should be explicitly granted temporarily and then dropped.
This means that any issue: a security breach or a mistake can’t access or break anything except whatever the component or person who had the issue could access or break, and that that access is minimal.
Suppose that you hit a page which exploits the https://www.hkcert.org/security-bulletin/mozilla-firefox-remote-code-execution-vulnerability_20230913 vulnerability in Firefox, or one like it, allowing remote code execution. If Firefox is running as root, the remote attacker now completely controls that machine. If you have SSH keys to other servers on there, they are all compromised. Your personal data could be encrypted for ransom. Anything that server manages, such as a TV or smart home equipment, could be manipulated arbitrarily, and possibly destroyed.
The same is true for any piece of software you use, because this is a general principle. Most distributions I believe don’t let you ssh in as root for that reason.
In short: don’t log in to anything as root; log in as a regular user and use sudo
to temporarily perform administrator actions.
P.S. your description of the situation shows you don’t know the nature of vulnerabilities and security - if you’re running servers then this is something you should learn more about in short order.
Yes? That’s why loans with collateral charge a lower interest rate than unsecured loans.