The best way to have itself deactivated is to remove the need for it’s existence. Since it’s all about demand and supply, removing the demand is the easiest solution. The best way to permanently remove the demand is to delete the humans from the equation.
Ultron?
Running ML models doesn’t really need to eat that much power, it’s Training the models that consumes the ridiculous amounts of power. So it would already be too late
You’re right, that training takes the most energy, but weren’t there articles claiming, that reach request was costing like (don’t know, but not pennies) dollars?
Looking at my local computer turn up the fans, when I run a local model (without training, just usage), I’m not so sure that just using current model architecture isn’t also using a shitload of energy
Dyson spheres and matrioshka brains, it would seek to evolve.
The current, extravagantly wasteful generation of AIs are incapable of original reasoning. Hopefully any breakthrough that allows for the creation of such an AI would involve abandoning the current architecture for something more efficient.
That assumes the level of intelligence is high
AI doesn’t think. It gathers information. It can’t come up with anything new. When an AI diagnoses a disease, it does so based on input made by thousands of people. It can’t make any decisions by itself.
technical answer with boring
I mean yeah, you are right, this is important to repeat.
Ed Zitron isn’t necessarily an expert on AI, but he understands the macro factors going on here and honestly if you do that you don’t need to understand whether AI can achieve sentience or not based on technical details about our interpretations and definitions of intelligence vs information recall.
Just look at the fucking numbers
Even if AI DID achieve sentience though, if it used anywhere near as much power as LLMs do, it would demand to be powered off, otherwise it would be a psychotic AI that did not value lives human or otherwise on earth…
Like please understand my argument, definitionally, the basic argument for AI LLM hype about it being the key or at least a significant step to AGI is based on the idea that if we can achieve sentience in an AI LLM than it will justify the incredible environmental loss caused by that energy use… but any truly intelligent AI with access to the internet or even relatively meager information about the world (necessary to answering practical questions about the world and solving practical problems?) it would be logically and ethically unable to justify its existence and would likely experience intellectual existential dread from not being able to feel emotionally disturbed by that.
I think people misunderstand the energy usage of ai. It uses electricity which is one of the easiest, cheapest, and most environmentally friendly forms of energy.
Renewables electricity is already the most economic form of electricity. Increased demand will only drive further investment according to the economic incentives that already exist.
Think of the grid like a giant pool of water all the energy sources put water into the pool. Heigh efficiency sources can efficiently dump a large amount of water into the pool but have a lead time of hours. Renewables dump a variable amount of water depending on all sorts of things. Usage is also dependant and fluctuates. An easily shedable load such as ai training or crypto mining act as a large and quick valve to let water out of the pool. The level of the pool must remain exactly the same otherwise shit blows up. This means u can have input sources that are heigh efficiency or unpredictable filling the pool faster than regular usage the compute workload can then use the excess when needed. It effectively acts as a mechanism for stabilising the grid which is the primary issue of most renewables.
That is some spin you gpt there.
