3 ms·

AI Doomsday Bullshit Is Getting Tired

32 points by cdrnsf 3mo ago

chelski 3mo ago
I'm so glad this is being discussed more openly. Ronen Farrow's piece on OpenAI had a sentence that stays with me "Sam Altman is a liar". They ran out of good things to say about AI so now the Darios and Sams turned to vague threats of world doom if we don't trust them to build the LLMs.
add-sub-mul-div 3mo ago
We pay too much attention to the idea of AI becoming Skynet and not enough attention to the more mundane reality that AI is simply a force multiplier for all manner of fraud and predatory commercialism.
- dieselgate 3mo ago
  > simply a force multiplier for all manner of fraud That reminds me of another nascent technology from the past decade-or-so... it's been great for money laundering and rug pulls.. and presidential enrichment
mckee_plus_plus 3mo ago
First of this meta ilk I've seen since 2024 (and it is tired)
treetalker 3mo ago
Tiresome, or tiring, perhaps; but not tired.
bko 3mo ago
It's good to look back historically to 1-2 years ago. Originally LLMs were not connected to the internet. When web search was exposed as a tool, the doomers lost their minds. Here Yudkowski: > “So, ChatGPT is not going to destroy the world. GPT‑4 is unlikely to destroy the world unless the people currently eking capabilities out of it take a much more dangerous route – like, hooking it up to systems and the broader internet in ways that let it act, scale, and learn in uncontrolled ways.” Sure, internet is one thing, but we wouldn't let LLMs act agentically, and make decisions for us. Fine but we wouldn't use it for long running unsupervised tasks. Okay fine but we wouldn't give them access to our actual computers. Fast forward to today where we have tens of thousands of people running OpenClaw or similar agents with unlimited permissions 24/7 and ... nothing happened. You would think AI doomers would reflect on their mantra "if anyone builds it everyone dies" and question whether AI demands basically a totalitarian one world government to restrain it.
- DennisP 3mo ago
  The "it" in that title is an AI smarter than humans. Nobody has built that. GPT-4 never did "scale and learn in uncontrolled ways," it got trained by OpenAI and had fixed weights and size after that. AI dumber than humans is a minor threat at best, even according to most doomers.
lilerjee 3mo ago
Creating the AI in AI hype is as absurd, ridiculous, and a waste of resources as building a perpetual motion machine. Don't they know? Why do they continue to develop, promote, and deceive like this? Sometimes, the truth is not important probably. But the game is being over very soon.
nancyminusone 3mo ago
Is AI even worth the hype if it can't even destroy the world?

32 points by cdrnsf 3mo ago

I'm so glad this is being discussed more openly. Ronen Farrow's piece on OpenAI had a sentence that stays with me "Sam Altman is a liar". They ran out of good things to say about AI so now the Darios and Sams turned to vague threats of world doom if we don't trust them to build the LLMs.

add-sub-mul-div 3mo ago

We pay too much attention to the idea of AI becoming Skynet and not enough attention to the more mundane reality that AI is simply a force multiplier for all manner of fraud and predatory commercialism.

dieselgate 3mo ago
> simply a force multiplier for all manner of fraud That reminds me of another nascent technology from the past decade-or-so... it's been great for money laundering and rug pulls.. and presidential enrichment

mckee_plus_plus 3mo ago

First of this meta ilk I've seen since 2024 (and it is tired)

treetalker 3mo ago

Tiresome, or tiring, perhaps; but not tired.

bko 3mo ago

It's good to look back historically to 1-2 years ago. Originally LLMs were not connected to the internet. When web search was exposed as a tool, the doomers lost their minds. Here Yudkowski: > “So, ChatGPT is not going to destroy the world. GPT‑4 is unlikely to destroy the world unless the people currently eking capabilities out of it take a much more dangerous route – like, hooking it up to systems and the broader internet in ways that let it act, scale, and learn in uncontrolled ways.” Sure, internet is one thing, but we wouldn't let LLMs act agentically, and make decisions for us. Fine but we wouldn't use it for long running unsupervised tasks. Okay fine but we wouldn't give them access to our actual computers. Fast forward to today where we have tens of thousands of people running OpenClaw or similar agents with unlimited permissions 24/7 and ... nothing happened. You would think AI doomers would reflect on their mantra "if anyone builds it everyone dies" and question whether AI demands basically a totalitarian one world government to restrain it.

DennisP 3mo ago
The "it" in that title is an AI smarter than humans. Nobody has built that. GPT-4 never did "scale and learn in uncontrolled ways," it got trained by OpenAI and had fixed weights and size after that. AI dumber than humans is a minor threat at best, even according to most doomers.

lilerjee 3mo ago

Creating the AI in AI hype is as absurd, ridiculous, and a waste of resources as building a perpetual motion machine. Don't they know? Why do they continue to develop, promote, and deceive like this? Sometimes, the truth is not important probably. But the game is being over very soon.

nancyminusone 3mo ago

Is AI even worth the hype if it can't even destroy the world?