The Problems with Artificial Intelligence Data Centers, At the Forefront Since March

Until recently, most of the articles I have referenced on the value of AI data facilities have been as positive as “Demand for AI Data Centers Sends Prospectors Hunting for Land and Power” (Tripp Mickle, The New York Times, March 5^th).  In this one, we learned how “a former Microsoft executive and his firm Cloverleaf have become modern-day land men, packaging electricity and land for data centers.”  He and his firm were portrayed as intrepid (“He’s a wildcatter”; “the team has faced hostile town halls… and also faces the risk that the A.I. boom fizzles out”), suitable (“they work at the intersection of utility companies and tech giants”), and timely (“their product - powered land, they call it - has become one of the nation’s most valuable commodities”).  Mickle seemed admiring, yet that may have been the last positive article on those facilities I have seen.

Later that month came out “Local Opposition Is Slowing A.I. Data Centers.  Wall Street Has Noticed” (Lydia DePillis, The New York Times, March 26^th).  It started with “The torrential wave of data center construction for artificial intelligence has seemed unstoppable,” and moved to “but lately, zoning commissions and county councils across the country have been resisting.”  Of “the top A.I. companies’” $710 billion 2026 facility-spending expectations, many have shifted their plans from resistive states such as Michigan, Oregon, and Wisconsin to more compliant Texas and New Mexico.  Yet “if America were to adopt A.I. at the scale that its boosters are banking on, data centers would need to be built closer to cities like Seattle, San Francisco and Chicago.”  The piece ended with doubt, specifically “if we assume tomorrow that data center construction stops… the ripple effects across the semiconductor industry would be pretty substantial.”

Next, we went as far as “Something Liberals and Conservatives Agree On:  Hatred of Data Centers” (Sabrina Tavernise, The New York Times, May 1^st).  “Early evidence suggests that Americans - once agnostic - are now souring on (data centers),” which a Milwaukee comedian called “the most bipartisan issue since beer.”  The refreshing anecdotals here featured liberals and conservatives finding common cause, when, for example, a Republican sportsman concerned about his fishing river joined with “a left-leaning musician and environmental activist.”  Although the article emphasized Michigan, it also mentioned related events in four other states.

Here, the June 11^th River Reporter led off with Patrick Kelly’s “Data centers delayed.”  “On Thursday, June 4, the New York State Legislature passed a moratorium on the approval of the development of large data centers for one year, among other regulations.”  Per “New York becomes first state to freeze new AI data centers in move critics warn could drive away jobs” (Bonny Chu, Fox Business, July 14^th), the state’s governor Kathy Hochul signed it, mandating “the nation’s first statewide temporary ban on new AI data centers.”  Per Hochul, who also wants to “repeal sales tax exemptions for large data centers,” “the initiative would require large data centers to shoulder more of the infrastructure costs they create and is intended to protect New Yorkers from rising utility bills and other financial risks associated with the industry’s rapid expansion.”  There will be more.

Here is also a political opportunity, as, per New York Times columnist Tressie McMillan Cottom on June 14^th, “Hating Data Centers Is a Winning Issue.”  She suggested it especially for Democrats, who could use that as a platform plank with more perceived sincerity than the Republicans would gain if they matched it.  As Cottom put it, “the voters showing up to fight data centers demonstrate that a lot of us want something different… Democrats… need a national message equal to the righteous rage driving millions of Americans to look up from their enemy and finally see, instead, a neighbor and future worth fighting for.”  Yes, “it’s simple.”

We know about these facilities’ impact on water and power, but how about auditory effects?  As Adeel Hassan explained in the New York Times on June 17^th, “The Cloud Has Sound:  The Unrelenting and Unseen Cost of A.I. Data Centers.”  They sound “like a low-frequency thrum of a neighbor’s central air-conditioning unit, an airplane flying overhead at high altitude or a truck engine idling down the road.  But it feels like the vibrating, rhythmic pulse of a subwoofer from a party that will never end… and some who live closest to data centers that emit the noise have reached their wit’s end trying to block it out.”  Unlike “many traditional community noise sources - like airports and freeways,” the disturbance comes out around the clock.  Although excessive sound in our country is usually unregulated, some have filed suit to change that.

Countries hoping for a US AI slowdown have been getting involved, as “China, Russia and Others Seek to Inflame Debate Over A.I. Data centers” (Steven Lee Myers and Dustin Volz, The New York Times, July 9^th).  They - one of the “others” was Iran - put up videos and a comic strip purported to be American-produced, “blaming data centers for soaring electricity bills” and saying they “posed a threat to Americans’ physical and financial well-being.”  These nations’ purpose was to “stoke the debate over data centers” and “deepen our divisions in order to dent our appeal and weaken us from within.”  It all sounds small-scale and unlikely to do real damage, especially since the side they are backing is now, as we will see, a majority position.

Next, “Data Centers to Add Billions in Power Costs in 13 States,” by Ivan Penn on July 14^th, also in the Times.  PJM, “the nation’s largest electrical grid operator,” held “an electricity auction that would add $6.3 billion in costs to the bills of millions of households and businesses within the next three years, an increase driven by the power demands of data centers.”  That’s supply and demand, but shows how those in one huge industry can massively increase the latter.  Appropriately, much of the rest of the piece emphasized supply growth as the solution, as state “governors have complained that PJM has been too slow to connect more power plants, solar and wind farms, batteries and other resources that would have helped lower prices and eased strain on the grid.”  That power network serves 13 states from Tennessee to Michigan and New Jersey.

Then, “US data center protests go national as backlash grows” (Valerie Volcovici and Lisa Baertlein, Reuters, July 18^th).  “Opponents of the rapid buildout of data centers,” namely the “grassroots group” HumansFirst, planned protests that day “in at least 125 locations across the United States, the first coordinated national effort to channel anger at the AI infrastructure expansion that has ramped up over the past year and roiled local politics.”  Participants were expected to “rally against what HumansFirst calls the “unaccountable” buildout of data centers and “unacceptable infringement on our liberty.”” Therefore, the movement now has a point of leadership.

So, “Is This the Fastest Opinion Shift in American Politics” (David Wallace-Wells and Robinson Meyer, The New York Times, July 15^th)?  The authors said that “seventy-one percent of Americans say they don’t want an A.I. data center built close to where they live,” changing from “last fall” when we were “almost exactly evenly split.”  Wallace-Wells and Meyer called it “shocking” and “absolutely crazy,” in “an insanely polarized country, which means it isn’t very often that you see a political supermajority on any issue, let alone one that emerges almost out of nowhere to announce itself as a major new front in our politics.” 

“So what’s actually going on here?”  The authors of the previous piece took a swing at that, and we can too.  It’s a rare combination of strongly valid local concerns, feelings of being pushed aside by stereotypically rapacious titanic corporations, failure of people to see any end in sight, a perception among at least half of Americans that the economy and the system in general are not working for them, and several more factors.  It will not end soon.  It will be hugely discouraging to companies.  But it will find balance.  There are places within our borders godforsaken enough to be considered by most to be suitable for such data centers.  When benefits from AI start making it to ordinary people, views will shift - slightly.  In the meantime, as I predicted on New Year’s Eve, in the artificial intelligence world a hard rain’s a-gonna fall.

Artificial Intelligence-Driven Robots: Moving Right Along, Literally

Ten of them in nine weeks, all provided by Fox News or Fox Business, and most by one of their leading technology writers.  Which ones do you like?

First, “AI robot changes your tires and balances them, too” (Kurt Knutsson, News, May 12^th).  The company Automated Tire, Inc. “has unveiled SmartBay, an AI-powered robotic tire change platform built for dealerships, tire shops and service centers.  The system handles tire changes, wheel balancing and vehicle inspections with minimal human intervention,” all that while leaving “the wheel on the car.”  These devices seem to be available now.

Next, “Rideable robot looks ready to stomp all over us” (Kurt Knutsson, News, May 19^th).  This one, from Unitree, costs $574,000, and “can carry a passenger, smash bricks and shift into a four-legged form.”  It looks like a real-life GoBot, and its “most likely uses seem to be entertainment, exhibitions, research, security demos or specialized industrial testing.”  You can probably buy one now - if you want to.

More practically, “AI robotic beehives installed in Florida community claim 70% reduction in colony collapse threatening crops” (Brittany Miller, News, May 21^st).  They are called BeeHome systems and are made by Beewise (note how these robot manufacturers are specializing around their offerings).  The product, now in use in Land O’ Lakes, “uses robotics, sensors and artificial intelligence to monitor hive health and protect colonies from environmental threats,” and “arrives as bee populations across the United States continue facing pressure from parasites, pesticides, disease and extreme weather conditions.”

Fourth, an improvement in an area getting automaton attention for not only years but decades: “Humanoid robots work nonstop in package test” (Kurt Knutsson, News, May 24^th).  This time, three of them, made by Figure AI and using its “in-house AI system” “without human control,” “crossed more than 24 hours of continuous autonomous operation after a test that was supposed to last only eight hours kept running.”  So, “if robots can keep working through long shifts, what happens to people who do this work today?”  Although “businesses… will want to know how often the robots fail, how much maintenance they need and whether they can handle messy conditions without slowing down the whole operation,” per Knutsson “if companies can make these robots reliable, safe and affordable, the warehouse floor could look very different in the years ahead.”

We used to hear how automata could not make up hotel rooms, which made it news when “Humanoid robot cleans first US apartment” (Kurt Knutsson, News, May 31^st).  This service, offered now by Gatsby only in San Francisco, costs $150 and involves one of them arriving as prescheduled “to clean your apartment,” including “dishes, surfaces, floors, making the bed and folding laundry.”  Although “routine work is autonomous,” and no person arrives with the robot, “harder tasks can be handled through remote human teleoperation.”  This service needs a longer success record, but as with the others here, the proof of concept has been completed.

Sixth, two other companies have made progress on a situation above, as “Warehouse robots move packages without human handoff” (Knutsson again, News, June 30^th).  The firms, Ambi Robotics and Pickle Robot Company (so much for focused names), have “announced a commercial integration that connects Pickle Robot’s trailer-unloading robots with Ambi Robotic’s AmbiStack pallet-building system.”  The author especially liked the handoff ability, something which is still almost always left to humans; that capability may often turn out to make the difference between worthwhile and doubtful automation.

Next, “Zoox robotaxi redesign brings big rider upgrades” (Knutsson, News, July 3^rd).  That “company has updated its custom-built electric robotaxi with new comfort and usability upgrades”; it “added more padding and ergonomic curves to the seats and headrests,” for comfort, and “updated the color, materials and finish,” which they maintain will make it seem “a calmer cabin.”  These are available in San Francisco and Las Vegas, with Miami and Austin “listed as “Now Arriving” on its ride pages.”  Zoox has a track record of success, and these improvements have no real chance of damaging that.

Eighth, “Starship delivery robots leave campuses for cities” (Knutsson, News, July 6^th).  Remember “those little white robots that once rolled across college sidewalks with lattes, fries and late-night snacks”?  They are “getting a new assignment,” as Starship “will wind down its U.S. university campus operations and redeploy more than 1,200 robots toward grocery chains and hot food delivery in cities across the United States and Europe.”  The choice to remove them was a matter of “focus,” and will not take effect until next year.  The move to city sidewalks is challenging, and may take a long time.

Next, “Hyundai Motor brings Boston Dynamics’ Atlas humanoid robot to FIFA World Cup in groundbreaking activation” (Scott Thompson, Business, July 6^th).  “A special guest delivered the match ball to the head referee on the pitch” before a game in East Rutherford, New Jersey, and you can probably guess who - or, rather what - that guest was.  Its product name was Atlas, “an advanced humanoid robot” that also “performed” for spectators in other ways, so Hyundai, per a top manager, could “demonstrate that the future isn’t something we imagine - it starts now.”  The article had no mention of malfunctions, which, despite the lack of practicality, was excellent for the company.

Last, “Humanoid robots perform live surgery in world first” (Jesse Watson, News, July 14^th).  “Surgeons remotely guided” two robots, which “copied the surgeons’ movements rather than making medical decisions” “through two gall bladder removal procedures” on pigs.  The test “marked the first time teleoperated humanoid robots,” as opposed to “surgical robots” which have been used before, “successfully completed live gallbladder surgeries.”  Another step.

All these developments are favorable, and point out how robots, especially in human form, are a fast-moving, effective area of artificial intelligence.  There will be many, many more, and most of these here will propagate, some dramatically.  That, without any real doubt, will be extremely positive.

Artificial Intelligence on the Job - What’s Happening?

Although AI has not clearly been eliminating jobs, a lot has been going on with it at work.

Its effect on the up-and-down popularity of working from home hasn’t been moving in the direction many would expect, as “Return to office gaining momentum as AI reshapes corporate strategy” (Arabelle Bennett, Fox Business, November 2nd).  Per Newmark, “a global commercial real estate advisory firm that counsels Fortune 500 occupiers and major landlords on leasing, strategy and transactions,” “artificial intelligence is driving a surprising surge in office demand.”  That company’s president says these companies are “making bold moves” to “re-skill” employees “and have them retrained in artificial intelligence.”

Speaking of up and down, we next got “More!  More!  More!  Tech Workers Max Out Their A.I. Use” (Kevin Roose, The New York Times, March 20th).  As “at tech companies like Meta and Shopify, managers have started to factor A.I. use into performance reviews, rewarding workers who make heavy use of A.I. tools and chastening those who don’t.”  “An engineer at Open AI processed 210 billion “tokens” - enough text to fill Wikipedia 33 times - through the company’s artificial intelligence models over the last week,” and “at Anthropic, a single user of the company’s A.I. coding system, Claude Code, racked up a bill of more than $150,000 in a month.”  “Some tech companies, including Meta and OpenAI” had “internal leaderboards that show how many tokens… each worker consumes,” and on which “employees compete(d).”  The logical worker response is for them to use AI nonconstructively, which would put an end to this rather shortsighted practice.  Indeed, on June 18th, less than three months later, the Times printed Eli Tan’s “Tech Workers Maxed Out Their A.I. Use.  Now They’re Trying to Minimize It.”; the change came from “the bills from companies, like Anthropic and OpenAI, that provide A.I. tools - and they were not cheap.”  Going from almost mandating it to putting “some monthly limits on A.I. coding tools” - was that embarrassing, or just the cost of learning?

In what might be another natural outcome of such maxing out, “Meta’s Embrace of A.I. Is Making Its Employees Miserable” (Kalley Huang, Eli Tan and Kate Conger, The New York Times again, May 8th).  In order for that organization’s management officially “to capture employee data so Meta’s artificial intelligence models could learn “how people actually complete everyday tasks using computers,”” it gathered “what employees typed into their computer, how they moved their mouse, where they clicked and what they saw on their screen”; as a result, “many workers immediately revolted” and “blasted the tracking as a privacy violation, calling it antisocial and callous.”  In conjunction with substantial announced upcoming layoffs, many workers there were showing “anger and anxiety.”

Specifically, “What Are A.I. Agents Actually Doing?” (Cade Metz, The New York Times, June 4th).  “A San Francisco start-up called Arena, which tracks hundreds of thousands of artificial intelligence users, is trying to take some of the mystery out of what” tasks it is performing.  It found that 17% in “agent mode” were “code-writing,” 10% were for “research,” and almost as many were for making images, creating “documents like graphs and spreadsheets,” and to “brainstorm ideas.”  Significant shares also were spent on “creative writing or tutoring and education,” along with “code debugging” and “chatting.”  A wide variety.

Yes, it seems clear that “We’re Only Starting to Grasp the Pitfalls of Using A.I. at Work” (Noam Scheiber, The New York Times, June 29th).  Although “at a conference where two human resources executives said that treating A.I. agents like real employees was a way to increase productivity and to put their companies on the cutting edge,” “in an experiment involving dozens of companies with A.I. employees, the researchers found that managers tended to vet documents less carefully when told an A.I. employee had produced them,” and they “missed errors that other managers caught when told they were vetting the work of a human.”  As AI models “tend to favor work produced by artificial intelligence” causing a “potentially consequential form of implicit ‘anti-human’ bias,” and AI models do not often “cooperate and seek win-win outcomes” as people do, we can question whether businesses are using too much AI.  

Should companies slow down?  That would be hard for many to do, but it might be the best choice.  As these stories show, the race, even when it is about implementing artificial intelligence, is not always to the swift.

June Jobs Report Slightly Warm, with AJSN Showing Latent Demand Up Seasonally to 17.5 Million

If you were looking for an exciting Bureau of Labor Statistics Employment Situation Summary this morning, you didn’t get one.  So I will parse what we got for you. 

Instead of double the predicted number of net new nonfarm payroll positions, we got half, 57,000 instead of a published 110,000 estimate.  Otherwise, most of the numbers were not disappointing.  Seasonally adjusted unemployment lost 0.1% to reach 4.2%, and the unadjusted variety stayed at 4.4%.  Measured adjustedly, the number of unemployed fell 200,000 to 7.1 million.  There were 100,000 fewer long-term jobless, out for 27 weeks or longer.  Those working part-time for economic reasons, or keeping such positions while looking unsuccessfully for full-time ones, also lost 100,000, to 4.7 million.  Average hourly private nonfarm payroll wages rose 11 cents, again close to inflation, to $37.64.  Two outcomes worsening were the two showing Americans’ connection to work, the labor force participation rate and the employment-population ratio, off 0.3% to 61.5% and down 0.2% to 59.0% respectively. 

The American Job Shortage Number or AJSN, which shows how many additional positions could be quickly filled if all knew they would be easy and routine to get, gained 310,000 to the following:

 

The largest shifts came from the count of those officially jobless, pushing the AJSN up 515,000, and those wanting work but not looking for it for the past year, moving it down 239,000 - no others were more than 55,000 either way.  Of the AJSN, 38.5% was from those employed, up 2.3% from last month.  Compared with a year before, the AJSN was almost unchanged, losing 52,000, with its largest input differences from those discouraged, down 139,000, and those saying they did not want a job, up 136,000. 

How, overall, did we do?  When taking the outcomes above, that those not in the labor force decreased almost 300,000, and remembering that fewer people work in June than in May, we did well.  It wasn’t huge, but it was positive and broad-based.  We cannot get discouraged about missing projections, as those do not affect anything except our perceptions.  The turtle took a small step forward.