And just like that, we are in the middle of the holiday season.
One thing I’m thankful for, I didn’t have to engage in air travel for Thanksgiving. I paid my dues for decades. It was only when I was in my 40s, with a wife and three children, that I earned the bona fides to stay put for Thanksgiving. My heart goes out to all those loyal children who brave the fickle weather and air traffic control system to come back home and be with their parents.
We had 12 around our table for Thanksgiving, which I think is the perfect number. Not too big, not too small.
Every year, something is different. This year it was cold. I could never recall it being this cold this soon.
I’ve always had a penchant for record temperatures. We’ve got about 175 years worth of recorded temperatures for Jackson. There are 365 days in a year. Each day there are several possible records temperatures to break: lowest, highest, highest low, lowest high, most rainfall, windiest, etc. That means every year, on average, there are about 15 daily Jackson weather records to break. I enjoy looking for them and documenting them.
These past few days broke several records of “lowest high” temperature for specific dates. For instance, December 2 had a high of 40, the lowest high temperature for that date since 1897, which was 37 degrees. The four other “lowest highs” for December 2 were 1985 (43 degrees), 1903 (43 degrees), 1927 (44 degrees) and 1915 (44 degrees.)
It used to take a lot of Googling to find these record temperatures. AI has changed that. I now just ask AI. It’s fascinating to watch AI work, as it lists where it is retrieving the information. What used to take 15 minutes takes one minute.
Imagine this time-saving scenario being multiplied by billions of people dozens of times each day. Imagine the efficiency and its effect on global productivity. That’s why the stock market is at a record high.
This is real. There may be temporary financial bubbles, but long term, the world is on the cusp of perhaps the greatest productivity leap in its history.
I use several AI chatbots: Perplexity, ChatGPT, Google’s Gemini, Microsoft’s Copilot, X’s Grok, Nova and Claude. Each has its own strengths and weaknesses. It’s interesting to compare the different results each chatbot produces for the same question.
Like any tool, it’s only as good as the person using it. We all have a big learning curve figuring out AI’s strengths and limitations. Just like the old Google search query, you have to be skilled in posing your question in the best manner. My favorites are Perplexity and ChatGPT with Gemini running a close third.
As a journalist, AI is profoundly powerful as a background research tool, allowing me to write intelligently on complex subjects quickly and efficiently. I can now feed AI a recorded speech and it will produce a decent news article. This is more accurate than the old way of using hurriedly scribbled notes.
I have also discovered numerous outright AI hallucinations, which is why human intervention is crucial. For instance, when doing legal research, AI will make up quotes from judicial decisions. The decisions are real and relevant, but AI simply makes up quotes it feels should be in the decision but aren’t. Recently, many attorneys have gotten into hot water by not catching such hallucinations.
This is why AI will not supplant human intervention. It is a program based on language pattern prediction. It’s artificial intelligence, not real intelligence. Of course, there’s a huge race to make AI better and better.
It’s fun to dress down AI when it hallucinates and ask it why. The apologies are immediate and non-defensive, usually something along the lines of “well what did you expect, I’m just a program.” We should learn to be similarly humble and apologetic when we screw up.
The main question is whether the efficiency of AI will offset the enormous energy costs caused by AI. Perplexity predicts a 10-25 percent increase in U. S. energy consumption by 2030 caused by AI — as much as $200 billion dollars a year.
Perplexity answers: “The projected extra U.S. energy cost from AI by 2030 (tens to a couple hundred billion dollars per year) is tiny compared with estimates of the economic and cost saving value AI could generate (trillions of dollars per year), though the distribution of those benefits and costs is highly uneven.”
The fact that I can use AI to research and answer this question in a matter of seconds is truly astounding to me. It would have taken hours of Googling in the past to arrive at a reasonable estimate before AI.
But as Perplexity points out, “the distribution of those benefits and costs is highly uneven.” This is particularly true in Mississippi, which has successfully pursued AI data centers by giving them and utility companies a blank check to pass the cost on to ordinary ratepayers. Just this week, a $3 billion Amazon data center was announced for Vicksburg. This is the fourth big data center project for Mississippi.
This is great news for the builders and their employees. But once the centers are built, the Mississippi ratepayers will be left holding the bill and a big bill it will be. There’s a reason these data centers are being built in Mississippi. We’re paying for them. It’s not right to force lower income Mississippians to pay for AI electricity bills.
Missouri is doing it differently. Just last month their public service commission took action to prevent what’s happening in Mississippi. They created a new category of electricity users: large load users, specifically designed for data centers. This will ensure that data center developers pay an acceptable rate that doesn’t burden residential users.
“I’m satisfied that this will result in the large-load customers bearing the financial responsibility for the costs that are incurred for infrastructure improvement, be they transmission or generation,” said Missouri commissioner John Mitchell.
Not only do we have no “large load user” rate in Mississippi, the rate data centers will pay is a “trade secret” that has been denied to the public. We have no idea what Amazon will be paying. Without knowing this secret rate, residential consumers will have no idea whether the data centers are indeed paying their fair share.
Sooner or later, the rubber will meet the road. Rates will skyrocket, consumers will rebel, the attorneys will have a field day and the political fallout will be monumental.
