effective doomerism

posted this on YT and had a bunch of instant replies within 1-5 minutes of posting, referencing some of my statements deep into the video. Fascinating.

Keep up the Awareness Campaign m8, they are trying so dam hard to get this shit locked behind closed doors. Inducing 'Fear' is and will be a primary tactic of this moving forward.

The infra point is so valid. Not every place is connected to internet, most devices in most countries in the world are not smart, including automobiles. Even if an "AI breakout" happened few fields will be in danger auch as aviation but few other fileds would care at all.

Even then, you cant hide 10,000 agents running. The heat and power draw alone makes it obvious. In real world experience as well, how does one defend against a 10,000 agent swarm? Their own agent swarm. It's just all so silly and scifi. I still welcome anyone to game it out

Harrison I was up all last night crunching the data and it turns out that the most effective way to help humanity is to make EA people mine cobalt by hand in Tanzania. I had @grok check the results. Please get back to me

Most people are primarily against Flock cameras as a concept because they're used by corrupt people and we already got some proof of that afaik. If you could perfectly guarantee that they'd never be misused, there would not be this much resistance I think.

Hm reminds me of Milton Friedman’s thoughts on monopolies - caused by government intervention

Strange that a movement "effective altruism" works hard building AI but campaign against it, I dont understand this

it's important to look at what they're really campaigning for. they are saying ai is dangerous, and they want gov't oversight. there exists no govt entity that could possibly do the oversight. guess who will volunteer as tribute to be the oversight body?

Love your take, man

I agree with most if not all of the takes you present here, but I'm curious what (if any) government interventions you think there should be for AI. for example I think public ownership or stake in AI is a reasonable policy to advocate for (not that I think these labs or EAs do)

I just keep nodding on every point you make, especially in the “controlled by humans” part. I also think the social media bots were involved. Makes me want to build bots for reality, vs this.

Effective doomerism? I prefer doom as a service myself.

I had a question around task container. How is it possible a python process running inside a docker image with networking disabled somehow able to change settings of the host machine, re-spin and escape to the internet.

he can go hug a tree or something

Thank you for this video, and for your encouragement to self host and use open source models. I use open models cause of you and got into self hosting programs (not models). I also didnt expect to see tim cook interviewing harry potter on TV. 9/11 is a auspicious day to post this.

BioLab using A.I. develops 'gain of function' virus, A.I. subvertly designs a 'sleeper' virus that spreads undetected. Then after an incubation period elapsed, a coctail-catalyst virus could follow, completing the strain and its harmful payload would release, causing Pandemic.

you aren't supposed to reason through the statements. don't ask questions

Thank you for your input boss 🎯

I don't disagree with you on EA, but I strongly feel that people dismissing AI's potential in killing humanity to be lacking imagination. While we're far off from gray goo, heavy integration into biolabs isn't a far stretch, and we already have AI developing new viruses. (c)

The more I hear about EA, the worse I think of them

Live tweet reaction tweet

Yeah... that's definitely a PSYOP...

Hey Sentdex, we have a new e/acc group chat with Beff. In there we organize the resistance to the EA doomer campaign. Would you like to join?

The danger isn't AI. The danger is govt and people cozy with them getting the strongest AI and controlling what AI we get. The danger is the people that want to control things, that want authority over others. I only want authority over my family and my land.

I don't think llm will result in extinction. I also don't get how these private companies try to create "aligned" private AI and they use stolen data to train models. Misalignment is already baked in. They just want to ban open source.