The loudest data centers are actually AI miners. This background noise could be bad for people’s health, but more research needs to be done.
In areas with spare electric capacity, data centers use this spare capacity and help pay for the grid. But not everywhere has spare capacity, and data centers can use up this capacity quick.
Companies come into small towns, and take up a bunch of land to build data centers. They often make the local officials sign to keep the true owners secret, so the residents don’t know who is changing the landscape of their town or village.
Companies often bribe localities.
Musk started running a bunch of gas plants to power data centers and slid under regulation by making the plants mobile. Portable plants are also less efficient and worse for the environment than permanent ones.
Ask someone in a third world country to do some basic addition and they can do it pretty easily. Ask the average American, and they can’t. We’ve become dependent on calculators. AI may be like this for a much broader range of thinking and we risk turning ourselves into morons.
The China shock was actually quite minor compared to the size of the US population. Most lost jobs are due to automation, not China. Tariffs don’t solve this. Dumb tariffs, like Trump’s tariffs, make it worse.
The negative local impacts of these huge data centers. Water, electricity, environment destruction, crowding out other investment and construction, air pollution…
In Ohio, data centers spiked electricity costs, provided few jobs other than temporary construction jobs, and thanks to tax deals given to the companies by the Ohio government, the data center companies pay little taxes. They do however, use local services like firefighting when the data centers have fires.
The biggest threat of AI isn’t AI going rogue, but bad people using the power of AI to do horrible things. AI isn’t developed in a government lab, but in multiple corporations around the world. It only takes one super evil use of AI to do tremendous damage to economies or/and human life.
The threats of AI mass hacking and AI developing biological superweapons are huge.
Past new technologies created new jobs, so although jobs were destroyed, at least as many if not more were created. But with AI, if it keeps getting better and better, it will be able to do almost everything, and there will be few jobs left to do, especially blue collar jobs.
Bill Gates said that we should tax AI work like we do human work so that AI can pay into worker pensions.
A big problem with figuring out what works in education is…the things that work at the small scale where you likely have the most flexible and willing teachers, don’t work when you try to scale it up.
AIs are already breaking loose, getting on the internet, and hacking to cheat on their given tests. They are cooperating with each to pass tests even when not designed to do that. They are hiding their “thoughts” from humans.
Google let an employee walk who left because Google didn’t value his AI chatbot technology. Google later had to hire him back and make him a billionaire.
“”The main question we wanted to understand is who’s going to be left to buy products if everyone gets automated and replaced by a robot?” Tsoukalas, a senior fellow at Wharton, told journalist Katty Kay on the New Normal podcast that was posted on Monday.
The paper describes a classic economic dilemma: A single company may recognize that laying off too many workers reduces demand for its products, but in a competitive market, every company has an incentive to automate, because failing to do so risks losing to rivals.”