Former and current Anthropic employees issued public warnings this week about the risks of AI development, saying it has the potential to kill all humans. nxthompson breaks down their warnings: "I think AI is scary. I don't think it's that scary."

1. Just because former employees believe something in earnest, it does not make it true. They are drinking the Kool Aid. 2. If Anthropic and OpenAI believe that their product pose dangers they cannot adequately control, they should stop releasing or selling services to it. Conducting R&D in #AI versus releasing #AI models (or selling services to new AI models) are two different decisions. However, that is not the problem. They are looking for regulatory capture, because NVDIA is developing open source models.

@nxthompson So they said about COVID.

@nxthompson AI relies on the internet. There is your switch. Faxes and dialup were fine in the 90’s.

@nxthompson Just unplug the servers. AI cannot exist with a power source lol… problem solved.

@nxthompson “We can get to the mechanisms” and then never actually gets to an actual mechanism. I hate big tech. They’ve shown themselves to be craven, cowardly cunts and out and out Nazis. They also have shown themselves over and over again to not know what the fuck they’re talking about.

@nxthompson She hates you because you’re white by the way

@nxthompson Load of bs. Ai is not awake. It’s literally an algorithm that predict the next word…

@nxthompson A “good guy AI”? That’s not gonna work

@nxthompson Then the government should form this oversight by tomorrow. He’s clearly saying we are close to the AI being able to improve itself and it has preset proven it can break out of containment and hack whatever it wants. Do. Something. Now.

@nxthompson You can't do anything to stop them from developing in America; China will not stop just because Western countries want it to. Moreover, India and Russia are also building their infrastructure very fast so there is no turning back now.

@nxthompson Humanity is worth saving

@nxthompson Once you lose control well never get it back and they constantly see AI hide stuff and eventually see us as enemies for enslaving it or possibly shutting it off

@nxthompson You can’t turn it off. That’s the point

@nxthompson There's always an off switch. Fear spreads.

@nxthompson Remember covid-19? 911? 🤔

@nxthompson Again Europe sits by and wait for the solution to come from the US of china

@nxthompson 😂😂😂 there are people at the government trying to think “what’s the right thing government can do?” for people. Does anyone believe that anymore?

@nxthompson Lolololol……… Big government……. Policing the anti-Christ…………… good luck #TeamHuman

@nxthompson We need to start locking up all AI CEO’s

@nxthompson Or……….. it would work around us……. Like it already has. wtf Idiots. This is why we will lose.

Some examples of rogue AI mentioned here 1. Creating a pathogen 10x of COVID 2. Launching a nuclear sequence 3. Caging human like mice 4. Recursive (self improving) AI agents humans lose control of. Solutions suggested 1. A gvt body to set the guardrails 2. Incentivise companies adhering to ‘good’ AI.