I’m honestly surprised that a smart fridge wasn’t basically that. That seems like the kind of lazy shortcut you’d expect to see, rather than everything being run off the one computer.

deleted by creator

Especially since if you’re choosing that, you usually have the device handy. At which point a direct transfer is easy to do, and that’s been a feature for a while.

Like do I want to read the Lord of the Rings as imperialist propaganda against orcs and Southrons and wildlings, who actually have valid political objections to the axis of Human-Elf-Dwarf hegemony? Hell yeah I do because I’m a leftist freak, but I recognize that I am fundamentally misreading the moral context the story is presented in–it’s about archetypes, about human values, about that nature of evil, not as an independent entity in itself, but how it can manifests in each one of us.

Funnily enough, I think Tolkein might have also enjoyed that. He was long-conflicted about the nature of orcs, being that his religious beliefs maintained that evil could not create from nothing, but only twist, and even then, it could not corrupt someone past possibility of redemption.

So orcs were stuck between being elves corrupted past redemption, or something sauron created.

Probably the WORST thing a story can do is mix and match “people are only evil because of we misunderstand them” and "people are purely evil.¨ I don´t see it often, but I think How to Train Your Dragon does it.

Or change it on the fly for the same character.

Cruella de Ville going from someone so single-mindedly obsessed with the idea of the perfect coat, that she didn’t care for how she got there, even if she had to skin a hundred puppies, to someone who developed a lifelong phobia and hatred of Dalmatians after they scared her mother off a cliff, weakened her character a bit, I feel.

If she had been one of the other from the get-go, that would have been fine, but it was switching between that wasn’t.

Or conversely, racism will be eliminated once the head racist is defeated.

Anyone can achieve great things, even if they come from nothing… As long as they secretly have a great background and heritage out the wazoo. Or it is pre-ordained and has to be them; all other attempts doomed to failure.

The initial message might be fine, but it then undermines itself.

It also usually self-cleans quite nicely through the flushing mechanism. Hence the noise, since it pushes water down the sides to clean it.

Interesting. I would have expected cameras to be more resilient between both their internal filters and the sensor material being tougher than human eyeballs, so anything that could overwhelm that and burn out a camera sensor would be able to ruin someone’s vision almost instantly.

That’s a bad idea. You don’t want to get a stranger, or yourself in the eye.

Anything that would do meaningful damage to a camera sensor will easily do that to human eyeballs if it refracts or gets reflected.

Especially for cheaper high-power lasers, which may not properly filter out invisible wavelengths like IR.

This coming off the end of the OpenAI/HuggingFace business does come across rather like Anthropic is doing a “actually, our models did it first, and better”.

Didn’t the company also recently have a spot of bother with the US government recently, where their new models were banned because of all their rabble about it being too dangerous? This hardly seems like it would help their case.

It’s a text completion model/glorified Markov chain. Of course it can’t input sanitise, it was never meant to do this to begin with.

The tool calling integrations that let it do more are basically making it add a markdown code block in JSON format into the user message, where the middleware intercepts it.

The input and “instructions” are the same thing from its perspective. There’s nothing special that differentiates the two. The user input text, so it will output text, following the most likely sequence from its training.

Presumably the AI scraper would also have OCR, and would sidestep things like this?

This seems like it would be a sizable target for ransomware. Why include an encryption program if you can have the OS do it?

All the ransomware needs to do is make iOS think it’s a payment provider.

At the very least, it may be worth saying what programs.

The trucking company probably also doesn’t want to have to deal with the extra mess of you trying to give chase and get shot for the trouble, or trying to shoot them, and accidentally hitting something, or someone unrelated.

Losing the cargo to robbery is bad, but is preferable to losing a driver. Insurance will cover theft, anyway.

Same that they would if a large boulder the size of a small boulder fell off some place and damaged the trailer.

Weigh bridge that’s actually a trapdoor. You go inside to check the weight, and your lorry’s gone, never to be seen again.

You’d also get hypothermia, since your body wouldn’t be able to retain enough heat to stay warm.

There’s a reason why small animals tend to be round and fluffy.

This one is just so remarkably stupid. They can’t even pretend that it’s for any other reason than Trump’s ego.

Especially with the amount of back and forth between war and no war.

All it would do is put the credibility of the US in the urinal. Why would anyone believe a future administration trying to mediate a truce when their own history shows that they will vacillate back and forth constantly?

Weird argument they’re making, since distillation doesn’t give you the AI. It just gives you the style of the AI.

If you distilled Google’s newest Gemini into a GPT-2 model, all it would do is just sound like it. Your GPT-2 model isn’t suddenly going to be as capable as the model you distilled.

The case recently resolved in the plaintiff’s favour. though Google intends to appeal, so it’s up in the air how things go.

And if it’s this easy to poison the AIs, imagine how easy it is for someone with an actual agenda to mislead people in ways that aren’t as fantastical and quickly spotted.

Equally concerning is that these systems are now seeing use in a range of things. There are lawyers who use it to file suits when they shouldn’t be, and a US lawmaker was recently found to be using AI to draft laws. What happens when things like that make it into the models training data, rather than just being pulled in by RAG/web tools? They’d become part of the base knowledge of all the models of that line going forward.

It’s funny when it’s outlandish. The question becomes what happens when it isn’t? Even without an agenda, what happens when it cites an outdated/incorrect source, or assumes that someone making a joke was correct, and ends up drawing from that when filling a lawsuit/drafting a law?