Call me crazy but I don't think traditional Kazakh diets were part of the study of 3000 pregnant mothers in New Hampshire.
Coffee, eggs, white rice
Selection bias much?
If you don't consume any of those 3 you're probably ridiculously wealthy on some freaky diet.
All this says to me is "The food of the masses is contaminated" which yeah - we already knew the rich pay a premium to get less contaminated food.
It's less idiocy and more laziness. Any amount of inconvenience is too much for a lot of humans, unfortunately.
It's impossible to do without exposing a private signing cert to everyone, yes. That's the issue.
You can't do asymmetric key signing anonymously and with a central issuer.
So either you have to just trust the assertions (0 security) or you have to have a trusted issuer (not anonymous)
A pseudonym issuer is a trusted issuer. There's no way to do it otherwise. You have to trust someone to make this kind of system work.
Agreed that law enforcement should not be involved but the quote I posted was also from the article and it seems impossible.
Most of these make sense and are definitely blockers for this ever releasing but -
Remove the concept of the Pseudonym Provider and ensure pseudonyms are generated and stored locally without the possibility of linking back to real identities.
Correct me if I'm wrong but this data all has to be signed somewhere right? Like the eID contains cryptographically signed assertions about the user in some standard (JWT?) format.
What use is signing the assertions locally? There would be no way to tell if the citizen actually had any valid id at all. A pseudonym provider is the privacy layer that allows for signing of new tokens after ensuring the validity of the old.
How could you sign an anonymous token using a valid one without it being linked back to the valid one? It seems like impossible constraints.
Am I totally off base here?
It's unclear if the current models can reach that level though. They seem more like they're asymtotically approaching their limit.
Maybe I'm wrong and GPT 5 will be the end all and be all - but I don't think generalist models will ever be consistent enough. Specialist generators focused on specifics, trained to output very defined data seem more likely to be useful than this attempt to make a single catch all LLM.
Gpt 4 premium pictures still have these major issues.
It's funny how the generators arent getting better. They've plateaued pretty hard in terms of believability. Glance value? Convincing.
Under any level of scrutiny though this falls apart in at least a dozen ways.
Melting polar ice is releasing CFCs back into the air, exacerbating the problem.
... But not when Joyce writes them. Don't worry he'll reference it again in 200 pages and if you haven't figured out what theme they represent you'll miss the fifth layer of context that actually inverts the meaning of the current paragraph.
(Ok he's not that obtuse but I wouldn't ever use death of the author/blue curtains on Joyce)
Google does have similar permissions requirements and all apps run sandboxed on Android.
Hmm maybe I stop donating then... I'll have to dig into where my money is actually going.
Don't believe that horseshit "we don't have money for Social Security" conservative talking point.
- We have money just not the way it's currently funded
- Cutting social security is political suicide
The only way social security goes away is if conservatives think they can win an election without old people. Fat chance.
Sorta. The foundation does.
Yes. Just like I donate to my Lemmy host on a regular basis.
I even pay for YouTube, despite using Vanced.
Fuck ads.