local LLMs aren't there yet and probably won't for the forseable future for ordinary people. If you want to run a full model like DeepSeek v4.1 Flash, you need something like half a TB of RAM.

The open-source LLMs are a different threat: They enable a service industry independent of the large players. I could today get some venture capital, rent a few racks in a hosting center, install a bunch of beefy machines and offer the same service as Claude or ChatGPT, just on an open-source model. The differences in benchmarks are becoming irrelevant for 95% of common use cases. And if I can offer that service for $19/m while OpenAI wants $29/m then OpenAI has a problem.

And IMHO this is the chinese strategy: Undermine the US AI industry by breaking its business model.

Who wizzed in your wheaties?

And I stand by what I said. Practically no phone that still works at all does it at the low level of a capacitor anymore.

Has everyone gone insane or are we witnessing a major PsyOps ?

Since AI isn't magical, for it to be able to "kill all humans", it would have to be able to acquire the means to do so. Which means that these already exist. It means someone is able to do it right now. And we're chill with that, but worried about AI?

No doubt a bunch of autonomous weapon systems linked into a C&C could cause considerable bloodshed. Given that most of the existing weapon systems of that kind are essentially suicide drones, it'll still be limited damage.

Human-killing virus? Still need a means of creating and delivering it.
Nuclear war? We have humans in the loop.
Drive us all insane with AI-slop TikToks? A few percent of humanity still don't use that.

So what is the actual scenario ? When someone shouts "AI will kill us all", why don't we ask them "how, exactly" ?

Well now, Flock is back in my good graces. We've got to wipe all those Pee-Wee Hermans off the face of the planet.

A law requiring anyone doing work in or on AI be subject to strict liability. Very strict. And that companies in that business name and publish a governing panel to internally regulate their AI work. And be subject to the aforementioned strict liability. Up to and including prison time. And no discharging liability through bankruptcy*.

The government can't regulate something that it does not understand. And maybe private enterprise doesn't either. But they need to shoulder the risk. This could be something like a mesothelioma trust fund, but set up before people start coofing.

*Hey. If it's good enough for student loans ...

Not really a great place to live. It being the final resting place for some of the targets from the Bikini Atoll tests. Also, the recovered remains of K-129 that we claim not to have are buried on one of the S.F. Bay Islands. Maybe this one.

We choose to do AI in this decade and do the other things, not because they are easy, but because they look easy.

All I need is support for Mojibake.

Well, one scenario is letting the hype get ahead of the capabilities, and do the 'AI store' experiment, but for real with responsibility for something actually important.

Another is that it is far easier to just screw things up than it is to make things work. I may not have it in me to make a house of cards, yet I can trivially even accidentally destroy a house of cards someone else built without a whiff of understanding.

The assessment of the capabilities based on bottom line certainly is subjective, which is precisely why Jensen Huang is so aggressively gung-ho about we must do everything we possibly can to spend more money on nvidia stuff. The motivation for Anthrophic seems either they have legitimate concerns or are looking to rationalize a plateau after promising the world. It certainly doesn't serve their bottom line to slow down otherwise.

So you believe North Korea is using carefully trained models with guardrails and the instruction, never become Skynet?

... for the lobbying failure. They tried contributing in BitCoin. Then someone converted that to pizzas and got upset.

I just hate it when Firefox trips my breaker trying to sort out Slashdot character encoding.

Oh come on! Slashdot can handle a percent sign just fine. It's just some Slashdditor that can't enter the correct character.

We had better get all those Teslas off the street, parked over storm drains.

Which administration? The DOJ complaints/investigations/consent decree against the Seattle police department started during the Clinton administration (or before), generated findings during G.W. Bush's and was acted upon during Obama's.

Law enforcement in general and the DOJ in particular are less political than you think*. And it's not Flock they will be doing something about. It's the abuse by the local police departments.

*Every time I hear accusations of politicized law enforcement, I just envision the scene of the bank robber screaming, "Attica! Attica!" Get off that high horse. You're robbing a bank.