Do you want to talk about it?

Error-free? No

Coherent? Absolutely. That is the surprising property of LLMs, that apparently language encodes enough about the real world to produce a coherent model of the world, if you just throw enough text at it.

Really a non sequitur. you could have one course “healthy use of new technology” and ban it for the rest of the school day for distraction-free learning.

Don’t waste your energy, Lemmy is far beyond saving to reach a nuanced view on generative AI.

thanks for the explanation. I misunderstood the statement. makes more sense that it makes fingerprinting harder.

what does that even mean? what aspect is more generic that could be used for fingerprinting?

We definitely have some control over each of those points:

• We don’t let kids drive, the lowest driving age seems to be 15 and is usually 18: https://en.wikipedia.org/wiki/List_of_minimum_driving_ages. Kids obv would have access to their parents cars.
• Knives, Guns etc cannot be sold to children and I hope you would agree that this is a good thing.
• having sex shouldn’t be controlled but providing or paying for sex work, “adult” events, sex parties etc definitely should be restricted.
• Legal drugs (alcohol, tobacco, weed) should be restricted for children. Illegal drugs are indeed hard to specifically regulate but e. g. selling illegal substances to a minor should carry a higher sentence.

after typing all that out I wonder what we are even debating. those are obvious examples to me where actual restriction is necessary. and to cycle back to the start, why would media be fundamentally exempt from that?

so somebody with zero programming skills and therefore zero experience with coding agents deemed AI assisted code such a big problem that they performatively forked a project at an arbitrary point. And they want to maybe maintain it while the whole issue started when the original maintainer said the project is too big to be maintained without assistance.

what are we doing here?

Thanks for clarifying.

For the record, I think this mechanism is indeed far reaching, simply because the current criterion of “age” is arbitrary in the proposed mechanism.

Not sure if the “just education” part is realistic, though. The whole point is that children cannot make responsible decisions like adults can. We accept that for driving, sex, drugs and weapons.

Not sure that’s necessarily true. I don’t see why it couldn’t work like this:

1. request personal token from authority. it works similar to a certificate chain, your token is derived from a central certificate
2. you store your token locally
3. you visit an age-restricted website. you send your token (or a challenge encrypted with that token) back to the website
4. the website verifies your token with the certificate from the authority, (like how literal Certificate Authorities work) . the CA doesn’t know when or why your token was used.

(fwiw I am sure governments will try their best to make this process less private)

wdym? in general? on the internet?

bc that’s a hard disagree from my side as a blanket statement

If you already have one that seems wasteful

sure, many things influence what technology is being developed, politics, culture, economy. That does not make it inherent, ie. inseparable from it. That would mean that you can only have technology where politics play a relevant role, which is obviously not true.

you don’t know what inherent means, apparenty.

fuck me, the “everything is political” crowd in here needs to touch some grass.

they call it that because it usually is a fallacy

not what OOP said. why is this such a trigger for you that basic reading comprehension goes out the door?

immeasurably shitty? go touch some grass.

you mean this statement? https://www.theregister.com/2026/01/08/linus_versus_llms_ai_slop_docs/?td=rt-3a

If yes, your statement does not really match what Linus said.

of course, but the OP said independent

I highly doubt that. how would that even work? a third-party to the publisher would have to check every statement before the issue goes to print. I can’t imagine this happening for anything that is not research papers or official reports.

but I happy to learn something new.