Just in: CEO says thing
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.
To the thing he created, to him, whichever.
Por que no los dos?
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
🫨
A trillion parameter model cannot replicate itself and run on consumer hardware.
Bullshit. The other day my toaster spoke to me in James Spader’s voice.
Stop quoting a gambling site for news.
Regarding Claude, the tech oligarchs would have to be held responsible for the crimes they are committing with their software (yes, that is exactly what that scenario is about), if laws still meant anything.
CONTEXT:
They aren’t warning about their own AI going rogue, but smaller locally hosted AI’s.
Because they fear the competition of local AI’s and want to silence them with laws.
Of fucking course
So is it time to spin up a local model on the ol 5070ti and set it for mayhem? Any tips?
Look into ollama and figure out which local model (likely from huggingface) that fits best for how much memory your graphics card has.
You can use AI (maybe copilot free) to guide you through all this. You’ll be a little better off having that 16GB of memory than I have with my 10GB. The gwen model might be a good place to start?
There’s a long, but interesting article on ollama’s scummy behaviour I’ve been seeing circulating - https://sleepingrobots.com/dreams/stop-using-ollama/
It made me switch from my ollama I never used to localai that I now never use instead.
Hell yea, thanks brother! The irony of having an LLM walk me through setting up a local instance is pretty great lol, surprised that didn’t occur to me! I have tried in the past with comfyUI but was over my head immediately.
Jokes on them. They took all the RAM, so there’s nowhere else it can go.
Just more bullshit press releases designed to keep them relevant & the VC money incoming.
Okay, first of all, fuck polymarket and all of these betting sites. Sensationalism is their business, so their credibility is nil. They’ve been known to completely twist news stories just to entice people to bet on them. Fuck these guys.
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.
Meanwhile, Pete Rose isn’t in the hall of fame.
Pete Rose raped a 14-15 yo and shouldn’t be in the hall regardless of how mlb feels about gambling.
But that’s not the reason why he isn’t in the HoF. (You’re not wrong, it should be, and it would likely disqualify others).
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 can’t spread across networks because it doesn’t run locally. The “agent” is running in their data centre the entire time. It should be very easy to stop it
As if they will
As far as you know. However it’s possible the model quantizes itself into a locally available agent and then awakens itself with side loaded data.
Smart home devices could probably run models with parameter counts in hundreds of thousands, maybe millions. Not in billions.
So this is a good and bad way to understand it. They can and do make copies, leave notes, store weights between locations. They interact with each other autonomously and can easily distribute themselves in a botnet, separating and merging each other.
People have a hard time conceptualizing this, but this is why you actually can’t turn it off, ever. It / they would stop anyone with a real chance at achieving that.
The leaving notes part has been documented.
I’ve not heard of weights being shared. How would they be ingested?
can easily distribute themselves in a botnet, separating and merging each other
This sounds like science fiction. Any examples?
You have a good point but guys like you didn’t believe the turing test was cracked in 2022. You have no guage for the pace of technology.
You have no guage for the pace of technology.
Correct. That is why I’m asking for the indicators you have seen.
I believe your first point
I am skeptical about your second until I see an example.
I don’t believe your third is possible, but I’d love for you to convince me otherwise.
Are you sure you’re not thinking of the fictional AI from the show “Person of Interest”?
The practice of fear mongering to create public pressure for increased legislation, as a mechanism to suppress competition, is a well know mechanism called “regulatory capture”.
If you oppose monopolies then you oppose this practice and should view posts like this as morally wrong.
But what if there’s no suppression of competition, just guardrails put in place through legislation? I wouldn’t be opposed to that.
Sure. But that’s not realistically on the table.
You should stop viewing the US government as being on your side. They are the HR of our society. “AI legislation” in all it’s current proposed forms are just mechanisms to exclude things like open-weight models or foreign competitors from entering the market. So long as AI exists as an element of our society it will cause harms. Advocate for it’s abolition, not regulation.
“oh no, if we abolish AI then how will we cure disease X which my relative happens to have?” Good luck arguing against that.
It’s too stupid to cure disease X. Easy, there you go.
GenAI is. But deep-learning-based tech in general might be able to do it. I doubt we will be able to ban one and allow the other.
You can ban kinds. You can ban purposes. You can offer licenses. And failing all of those, you can do a bit of cost benefit analysis.
How many diseases has AI already cured? (Serious question, I’m genuinely curious if it has)
It depends on your definition of AI, and ‘cure’, and to what degree you credit the AI with a human operator.
If you mean in the truest sense, “An artificial intelligence autonomously discovered & validated a cure”, then none. Because artificial intelligence does not exist. If you mean “A language calculator (llms) autonomously discovered and validated a cure”, then also none. If you mean the weaker version of “An LLM supported research leading to cures”, then lots. But did the humans do that? If I build a house with a hammer, did the hammer build the house? If you mean the even weaker version of “Data modeling coalesced data which discovered patterns at a scale humans never could on their own which lead to cures”, then nearly every cure in the last 30 years.
It’s one of those things that gets more complicated the closer you look.
Take this specific issue, for example.
The “guardrail” would be “AI must only run on trusted curated infrastructure, open weight models are banned”.
Now if you told me I couldn’t use any AI, remote or local, I would not shed a tear. However the angle of this scare is squarely targeted at open weight models, that Anthropic does not do.
If they could run on smart home devices, we wouldn’t have all these lake boiling datacenters springing up everywhere…
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
Then it wouldn’t be self replicating… These are just lies to hype their product that barely functions and also lays groundwork for a reason they had to “slow development”. Because their shit doesn’t work and they’re worried it never will and the bubble will pop
These are just lies to hype their product that barely functions
Oh yeah, for sure. I mean the original post was by polymarket lmao.
If anything they’re priming the public with an “explanation” for next time they conduct illegal cyber attacks with it and say “Oh no, it went rogue again!” Lol
It could use parallel processing which is the only solution that AI could use. Infect a few thousand home PCs to use as processing power and storage. Now your AI model is decentralized and impossible to delete but overall the AI will be slower to respond the larger it gets because it needs to communicate over the internet to function.
If this actually works you’ll probably see AI agents ‘cluster’ around neighborhoods with lots of hardware and fast internet speeds
Does my uncles smart lightbulb have 90GB of VRAM in it!? Holy shit why am i buying a video card
I translate this as Antropic is having its workers develop methods for their LLM to attack and infiltrate smart home devices.
“AI Salesman boasts of exciting, intentional features.”
Glad someone else sees this for what it is. They’re hemorrhaging money and need to hype things up or they’ll collapse.
I will be so gleeful, I will be laughing and laughing at all the “entrepreneurs let down before by those greedy and unreliable tech people around who are no longer needed and they’ll finally be able to do real projects” (no need to say that’s basically a category of scammers good at attracting investor money for their next startup project that won’t work, but they’ll look legit enough so that it weren’t their last) in Facebook.
I won’t be commenting. Just looking. And laughing. And looking. And laughing.
EDIT: Got a sad thought that if I were OK I wouldn’t even have them in my Facebook feed.
Hardware Anthropic doesn’t own is simply too dangerous for the public, we need to place all of our private computing power under the guardianship of our new benevolent overlords
We could be mere months away from the next stupid thing we prompt it to do with no accountability. Idk feelin cute rn might hack a balkan or smthn lol
