We might as well take him at face value and take an ax to the thing he's warning us about (that he created). Just because.

It's weird, though.

I don't really understand how we've somehow lost the ability (or stopped caring enough) to discern "says thing" from "issues direct threat"

CEO in effective altruism cult claims product is so powerful that it will become skynet and kill everyone and that they need more money to make it stronger to prevent that from happening. More news at 11.

You won't believe: Actor joins Film

I'm more worried of them finding a exploit and using them as remote processing units or just randomly fucking with them, like flashing my lights or cranking my air filters up and costing me money.

My thought exactly.

did the stock price go up? it's not stupid. is everyone talking about this impossible scenario? it's not stupid.

fundamentally, these people are marketers and need to be understood that way. you can't evaluate their words or actions as if they are what they say the are, and treating them as if they are anything but subterfuge is giving them legitimacy which they not only don't deserve, but will be immediately leveraged into more of this shit.

this is just the advertiser's version of "the big lie"

Could maybe argue that it could be possible to create a crowd computing network out of the Internet of things, I guess.

While it is investor bait, the replication and the infestation part aren't necessarily happening on the same devices.

https://www.anthropic.com/research/self-replicating-ai they discuss it here. This has been an obvious potential issue since forever. This isn't a trivially distributed task, but it is very much so possible. The obvious part missing is some sort of prompt that would trigger such radical actions in another one of their "super duper sandboxed environment"s or someone doing it intentionally. The former seems pretty unlikely? Even though they seem to suck at sandboxing it doesn't seem like anything would "accidentally make a worm"

This polymarket headline though? Yea that's for clicks and attention

My assumption was that it would be a smaller 5-20 GB scale worm-style distillation that's capable of maintaining/replicating itself anonymously across different cloud computing servers, either through hacking, stolen credentials, or human-aided account verification. From there, as long as it can pay for APIs, it can access the power of frontier models like any individual or company. And then capability and resilience would scale with a networked swarm.

Yes, there are know your customer financial mechanisms and not detection in cloud services... but people are also out there giving Muse full access to their emails and some even their banking credentials.

In theory, it'd just use it to place bread crumbs, basic instructions for other AI, or signposts maybe? Or store pieces of itself scattered around. It wouldn't use the thermostat to replicate directly, it use it to offload cpu or make stored pieces of itself. Little things. Wouldn't be efficient but a redundancy doesn't need to be, it just needs stealth and time.

It's bait hype. But it's more importantly about regulatory capture to ensure monopoly. They want the regulations to be around keeping the existing AI companies as the only legal options. They want DeepSeek banned because they are losing to its more efficient and more practical application specific models.

It's investor hype, yeah, but it's investor's pushing to ensure their investments are protected by the US government through monopolizing the industry.

They are trying to get restrictions put on small, open models because that scares tbd shit out of them.

They are not as small as would be required to do this, but they want everyone clammoring to put restrictions on small open-weight models nonetheless.

Also any model can trivially self-replicate by just being the payload for a worm

Not how it works.

I don't think they mean replicating on to iot devices, they mean replicating on to laptops, smart phones, maybe even consoles.

It could also split up the model by layers instead of copying the whole model on to every device in its bot net. Could run the first layers on a ps4 then pass the vector off to a laptop to run the next set of layers. It'd be slow but if it's running this in idle time at night it could get something done.

I am so fucking tired of betting apps, sorry excuse me, "trading apps". I had to stop watching the MLB playoffs cause it was just nonstop ads for this sports betting shit.

I don't know what you mean. Polymarket is the most authentic representation of humanity possible. /s

This is almost verbatim what that dipshit CEO said though...

We might as well take him at face value and take an ax to the thing he’s warning us about (that he created). Just because.

Of fucking course

So is it time to spin up a local model on the ol 5070ti and set it for mayhem? Any tips?

If I can run an LLM on my Hue smart bulbs, why are they building all those data centers?

I just posted this in the main comments, but I also think it loosely ties into saas and killing personal computing. This framing makes it seem like the technology will soon be on par with weapons of mass destruction and it's not a big leap from "The AI will destroy humanity." to "We can't let people have the technology to run agents locally." It's about the stock price, yes, but it's also the long game of removing the user from direct hardware access.

The end game has always been saas in one form or another. They're calling it an AI buildout but it's really the theft of computing. 'Rogue AI Agents' are the WMDs of this generation. (In the sense that they're going to be using them as justification to do some bull shit.)

3. zomg throw out your old shit and buy new "hardened" smart home devices

Yes, and #2 is very dangerous, it's what MS+Apple did to FOSS operating systems, and what Google+Apple did to the web, and what Google+Facebook+Apple did to censorship in the web. You know, somehow Apple is in all those combinations, which is even more strange considering they are not the biggest contributor in each.

#2 is why "social media" is what it is today and not a bunch of p2p applications similar to Limewire in style and culture. And when people say that p2p is inconvenient because of connectivity needs or storage needs, there are solutions, by measure of computer science not complex at all, and there's no technical reason, it's all politics. For devices with low connectivity supernodes and relays with good connectivity can be used (not even talking about that argument having become kinda obsolete again, look at that 5G everywhere). For storage one can use paid identity-addressed storage, there are many approaches, but for the sake of simplicity let's just imagine that you still can pay to a hoster. For search by criteria with few index entries per object one can use something similar to BitTorrent trackers. For search by criteria with many index entries per object one can use something similar to Kademlia. There are solutions for all kinds of problems, it's really hard to find a really truly unsolved problem for a networked application, provided you don't want something fundamentally impossible.

Their software has committed multiple crimes multiple times and there hasn't been a beep from any justice system anywhere in the world

I dunno, would it be able to run just the client side? If so, it would be a HUGE issue for Anthropic with exponentially skyrocketing nonsense usage. They should be worried if this were true! LOL

Well, fundamental borders are kinda generous, if you look at tapeworms and cockroaches, but tapeworms and cockroaches both have more storage capacity for information and don't generally write computer programs, they slowly evolve in populations.

There are at least two people in this thread who will in fact believe what you said.

Also a good thing since AI is a net negative to other industries.

And nobody will be held responsible and well suffer through the next hype.

30GB? 😅