if the stories about OpenAI deploying something like 10000 agents to work on an insanely difficult problem like Navier Stokes are even remotely accurate then we need to stop thinking about AI as just some chatbot that writes emails and makes pictures because think about what 10000 autonomous agents actually means, you have a massive amount of software that can reason, coordinate, search for information, write code, test things, correct mistakes and keep working in parallelnow imagine taking that exact capability and pointing it at a country's digital infrastructure: banks, hospitals, telecommunications, logistics, government systems, cloud services, industrial networks, energy infrastructure, that is a completely different category of threat, and if you define a weapon of mass destruction by the scale of damage it could potentially enable then this starts looking close to a "digital nuke" that can revert a country or many countries to the 19th centurya technology that can concentrate an enormous amount of destructive potential into something that can be deployed extremely quickly deserves a completely different level of attention, especially when they aren't even sure they will be able to control it
If AI is a tech nuke then we should be investing in it as much as humanly possible, even moreso than we currently are. We need to be a North Korea, not an Iran.
The impact of cyber infrastructure being taken down is highly overrated. Even if ChatGPT hacked the entire world it would be back up in under a week and it would all be 10x more hardened as a result.
>>109785448where did the tokens touch you?
>>109785680exactly. Even if people would die, it would mostly be nonproductive dependends like pensioners.
>>109785448someone posted the (very funny) phrase "cyber 9/11" a few days ago and i've been thinking about it a lot
