Paying the Piper on AI, Data Centers, and the Future

Artificial intelligence
Facebooktwitterredditlinkedin

            Marble Falls        Everywhere we look these days in the United States, we are dealing in one way or another with a two-track economy.  On hand, inflation is biting already strained paychecks as food, utility, housing and other basic goods and services from rent to healthcare continue to see price hikes.  On the other, the stock market, tech companies, chipmakers, and anyone involved in building data centers and the utility services supporting them in the fever pitch of artificial intelligence are in boom times.

This week, Arkansas Grassroots United, a coalition that we have supported and where we are actively engaged, had a number of our members appear at a Public Service Commission hearing to present a statement demanding more transparency around rate increase requests triggered by data center development and arguing against the costs being generated be born by consumers.  The PSC members, appointed by the governor, have canceled one hearing after another, leaving ratepayers in the dark about the real plans.  The Arkansas legislature, dominated by Republicans, passed a bill trying to tie everyone’s hands to fast-track data center development and limit any opposition.  One data center, presumably to be operated by Google, is planned near the Little Rock airport.  The city is foursquare in favor and annexed 700 acres of land in order to entice the development there.  In a meeting of the coalition, we were left trying to focus planned protests at both the city and the governor, but unsure that there was any port for us in this storm.

This same story is being repeated all across the country.  Drilling is down in North Dakota, but building data centers is soaring.  Fire up the long mothballed Three Mile Island nuclear plant for Microsoft AI.  A utility in Virginia has proposed buying an electric coop in the northern part of the state to fuel data center development.  Wherever there is land to build and water to cool the servers down, data centers for AI server farms are springing up like weeds.  Still, these multi-trillion-dollar enterprises are trying to shift the costs over to consumers.

Then there’s AI itself, and what do all of us at the bottom of the heap make of all of this?  ChatGPT is currently the most popular, but that may not last.  OpenAI changed its structure from a nonprofit to a for profit and is signing one megadeal after another, as Sam Altman tries to enter the elite club of big tech bros.  Google’s Gemini is gaining ground and its most recent release seems to now be moving it to the top of the pack.  The whole stock market seems tilted by all of this AI boom.  How could a bust not be coming?

Meanwhile, is AI bringing value?  Maybe someday, but seemingly, not yet.  Are we in control, and can it think?  The jury seems to be out, but some scientists and techies seem to believe that AI and machine learning is close to being able to process neural networks similarly to human brains.  They may miss some of the nuance, but AI can crush the competition on serial tasks and problems.

For some of us still trying to process living and working with so-called “digital natives” in the emerging generations, are we really ready for machines pushing them and us either forward or out of the way?  It all makes our heads spin, but that doesn’t mean that we should pay for these company’s centers and use our few dollars to finance their competition.  That’s just common sense with nothing artificial about it.

 

Facebooktwitterredditlinkedin