Last week I shared my six-week struggle with Hertz over a questionable damage report. I figured “Zendy” would make for a great case study in how not to deploy AI.

Holy smokes, did it touch a nerve. The article reached 350,000 people, and hundreds of people shared their own rental horror stories.

I even got emails from (gasp) real-life humans at both Hertz and Dollar.

So I’ve now spent more time talking about AI damage scanning than anyone on the planet who doesn’t wear a Hertz badge. Two things became clear, and one of them was a shocker.

First, AI wasn’t the villain. Hertz was. It let its chatbots make promises it didn’t keep, couldn’t explain scanner results that mirrored pre-existing damage, and cut out the humans who would have known better.

Second, damage scanning is a terrific use case for AI. Strip away Hertz’s hamfisted rollout and it fixes most of what’s broken with today’s manual inspections.

Wait, what? Hear me out.

Damage inspection sucks, but it doesn’t have to

I’ve heard countless accounts of renters being unfairly accused of damaging their cars, which often ends with “I’ll never rent from you again.”

This is partly because spotting damage is more complicated than it sounds.

An inspector has to walk around the car, tell scratches from grime, and check each one against pre-existing marks and size thresholds. If the car isn’t inspected right away, there’s no way to know if the customer caused a scratch or the lot did.

Now add the human element. Some inspectors are veterans, others are new or fill-ins, and almost all are overworked. Multiply that by millions of rentals a year, and consistency is impossible.

Implemented well, AI fixes all of this. It sees the whole car through the grime, measures dents to exact size, fires the instant a car leaves and returns to the lot, and doesn’t grade harder the morning after a fight with the kids. The result: Hertz’s AI scanners spot 5 times more damage than human reviewers.

This looks like a goldmine for Hertz, but it can benefit renters too. Instead of arguing about damage six weeks after the fact, a clean before-and-after clears you automatically.

The “It’s a rental” subsidy

My first consulting boss was a great guy, but he drove rental cars like Brad Pitt in the F1 movie. He peeled out of intersections and flew over potholes, the louder the tires squealed the better. He’d chuck our McDonald’s trash in the back seat and laugh, “it’s a rental.” Can you imagine someone borrowing your car and treating it like that?

Driving a rental like that is fun, and it’s expensive. Used rental cars sell for 10% less than comparable models, partly because of this stigma. A rental company’s biggest cost is depreciation, and Hertz’s was $2 billion in 2025, 23% of revenue.

My boss never scratched our cars, but plenty of “it’s a rental” drivers do, and the cost of dents they leave gets passed back to renters through higher rates.

Poor damage tracking isn’t a free lunch, it’s a subsidy from the rest of us to a few Brad Pitts.

New tech, meet old habits

Companies finally have a tool that can detect damage reliably, but it’s colliding with habits that have been around for decades.

The old walk-around-with-a-clipboard process let most scratches fall through the cracks.

Hertz’s new scanners are catching them, which may explain why negative reviews mentioning damage rose from 8% to 12% last year.

If Hertz wants to stop pissing off its customers and have them treat cars like their own, it needs to do more than just deploy AI.

The email Hertz should have sent

Here’s what an email from a company that actually wants to keep its customers might look like.

Subject: Safe drivers save money with Hertz

Hello valued renter,

Hertz spends millions of dollars every year fixing damaged rental cars. When we can’t identify who caused the damage, it forces us to raise rates for everyone.

That’s why we’re rolling out ScanSafe, our program that bills drivers who cause damage instead of making everyone pay. Here’s how it works:

When you drive off the lot, a UV scanner photographs the car to record any damage. We’ll send the photos to you, so you can flag anything that looks off.

When you return, a second scanner compares the car to your original photos. If there’s no difference, you’ll get an ‘all clear.’ If something turns up, you’ll get the details and a phone number for a live agent.

We’re passing the savings on to you. Every time you return a car with a clean report, you’ll receive a $5 credit towards your next rental.

Thanks for helping Hertz keep our cars in great shape!

A win-win, if Hertz wants one

Done well, AI damage scanning is better for everyone. Hertz recoups the cost of dents. Safe drivers stop subsidizing Brad Pitts, and the maddening disputes drop for everyone.

Knowing someone’s watching even makes the F1 hopefuls think twice before peeling out.

But Hertz isn’t doing this well, as I learned the hard way from AI “Zendy.” It threw AI on top of an already-broken process, making it more maddening and more expensive.

The question isn’t whether AI scanning works. It’s whether Hertz wants to fix things for renters, or just gouge them.

If it’s the former, AI scanning could be a positive step towards reclaiming its “nobody does it better” mantle.

If it’s the latter, it’s another step towards becoming Spirit Airlines with a gotcha.

Right now, it looks like they’re choosing the gotcha.

Share

Dad Joke: What did the renter say when AI Zendy raised their bill? “That’s so painful, it Mega-Hertz” 😂

Last month as I boarded my flight home from Philadelphia, I got a text message from Hertz: my rental car was damaged and they wanted $125.

Six weeks later, the bill has gone up to $238, my wife is getting collection calls, and I’m not sure if I’ve even talked to a human yet.

Hertz’s AI scanning is a textbook lesson in how not to deploy new technology.

Meet Zendy

My car’s damage was detected by UVeye, a TSA-style scanner that I’d driven through at the return lot.

I wasn’t buying it. The photos showed pre-existing damage in the exact same spot. I explained this to Hertz’s AI chatbot, who logged a ticket.

The next day, I received a lovely note from Zendy, the one person at Hertz who seemed to be on my side.

Three weeks later, I hadn’t heard back. Instead, my wife started getting letters and voicemails from a collections agency. I called back and reached an agent named Ruby. She explained that my bill had gone up to $238 because I missed a deadline, and asked me how I wanted to pay.

Then she offered a tip: file a claim with my credit card’s travel insurance and have them cover it.

Ruby was unaware of my dispute or conversations with Zendy. She said she’d call me back.

Six weeks since the first text message, I’m still waiting.

I put Zendy’s emails into an AI detector and it came back flashing red.

So much for Zendy.

Hertz’s tangled web

I’m not the only one who’s been caught in Hertz’s damage-scanning net. Reddit forums are teeming with angry customers comparing notes on how to fight the system.

One customer was billed $440 for a scuff smaller than Hertz’s own one-inch guidelines. Another was charged after a human agent walked around their car and confirmed there was no damage.

Note: My rental was from Dollar. Hertz owns Dollar and uses AI scanning across all its brands. The texts, emails, calls, and letters all came from Hertz.

Hertz claims that UVeye finds damage in only 3% of rentals. That seems small, but Hertz had scanned 675,000 cars by last August. That means 20,000 customers have gotten caught.

And the charge always hits at the end of the rental, so the last thing customers remember about Hertz is a “gotcha” text.

Frustration has reached Congress. House oversight subcommittee chair Nancy Mace and Senator Richard Blumenthal sent letters to CEO Gil West, noting that Hertz is “the only car rental company in the U.S. that issues damage assessments without human review.”

Hertz broke rule #1

Hertz didn’t deploy AI to make renting a car faster, better, or cheaper. They did it to squeeze their customers.

The company needs the money. It still carries $19 billion in debt after emerging from bankruptcy in 2020.

And UVeye’s own marketing brags that its scanners detect five times more damage and generate six times more revenue than manual checks.

They’re also saving money by cutting humans from the process. And it shows.

UVeye’s claims are sent to customers without human review, and Hertz agents aren’t trained to help. Several renters said employees shrugged, pointed to the AI, and told them to email customer service.

Hertz systems don’t talk to each other. Zendy’s promises were empty: no investigation happened, my bill wasn’t put on hold, and collectors kept calling my wife. When I finally reached a human, she’d never heard of Zendy.

Consider what Hertz chose to automate: the explanation, the empathy, and the follow-up were all AI.

The only time I got a live human was when they came to collect.

Sixt does it better

Hertz should take a lesson from Sixt, the only other rental company using AI damage scanners.

Sixt’s AI scanners flag damage. Then humans review the photos and decide which will become claims. Customers who disagree or have questions can request a second look.

A simple reframe would flip the customer mindset. Imagine an email to every renter before they pick up their car:

“UVeye is our new program to keep your rates low. Careful drivers save money because we bill the ones responsible for damage instead of everyone.”

I’m a careful driver. Thanks for saving me money, Hertz!

If you’re going to be Spirit Airlines, say you are

A final chapter in my story. When I picked up the car, the office had a line of 30 people and one agent. A QR code offered automated check-in for $22. I paid up and drove off the lot in minutes.

At most companies, “skip the line” is standard. Hertz is charging for it.

Low-cost service with add-on fees is a viable business model, and Spirit Airlines built a company on it.

But Hertz isn’t priced like Spirit. They’re charging full price for a nickel-and-dime experience, and hoping nobody notices until the gotcha hits.

Their tagline used to be “Nobody does it better.” In the 80s, nobody did.

If Hertz wants to become Spirit Airlines, it should say so.

It’s too late for me, though. They may get their $238, but I’ll never rent from Hertz again.

Share

Dad Joke: Why were the paparazzi let down by the AI damage report? Instead of a Scan-dal it was a Dull scan. 🤣

On Friday the Chicago Bears announced plans for a new stadium in Hammond, Indiana. This opens up a new round of palace intrigue:

• Will JB Pritzker’s “no money for billionaires” stance prevent a taxpayer boondoggle or backfire?

• Are Republican lawmakers secretly sabotaging plans to hurt Pritzker’s presidential hopes?

• Does Chicago Mayor Brandon Johnson prefer Indiana to Arlington Heights because fans will fill up Chicago’s hotels and bars instead of the suburbs’?

• Are Kevin Warren and George McCaskey bluffing or will they actually relocate the team?

The subplots make for fun debates.

As a Bears fan I couldn’t care less. If this move happens I’ll hold it against all of them.

We’ve accepted the idea of playing in the suburbs, and under a dome. Fine.

And I know the New York Giants and Jets play in Jersey. But New York’s signature team is the Yankees.

Chicago’s team is the Bears. The team of George Halas. Red Grange. Gale Sayers. Walter Payton. Dick Butkus. Mike Singletary. Mike Ditka. Brian Urlacher. Devin Hester. The Superfans.

It’s been over 40 years since the ‘85 Monsters of the Midway. Thanks to Caleb Williams and Ben Johnson, we finally have something on the field to cheer for.

That field should be in Chicago.

Bear Down.

In Dayton, city workers have wrapped Flock surveillance cameras in trash bags.

This after residents protested that their data was shared with ICE for immigration enforcement.

Flock pitches itself for safety and crime prevention. It’s working: they’ve installed 100,000 cameras in 49 states and nearly every major American city.

Local protests are building, but they’re nothing like the pitchforks coming for data centers.

This is because Flock has an easier story to tell.

• Its benefits are more vivid: Catching a kidnapper or creating 200 new jobs?

• Its downsides are less visceral: Big Tech’s stealing my water or Big Brother’s watching?

And nobody notices Flock’s cameras, which are small, solar powered and easy to install.

Flock’s low profile won’t last forever. Surveillance cameras have real tradeoffs, and most cities signed on without debating them.

Otherwise, like the shepherd said…let’s get the Flock out of here 😂

AI is killing call center jobs. Except now it isn’t.

Call center jobs are a perfect target for AI agents. Most reps work remotely and follow prewritten scripts. Conversations are recorded, creating a goldmine for model training. In 2023 Forrester predicted that 49% of customer service jobs would “disappear” by 2030.

Recent headlines tell a different story. Customer service job listings are outpacing the rest of the market this year, and call center employment in the Philippines continues growing steadily.

As satisfying as it is to see AI job doomers foiled, this is only a small uptick. Openings are down significantly from their 2022 peak and the Bureau of Labor Statistics projects a 5% drop in jobs over the next ten years.

This is confusing, because two winds are blowing in opposite directions. AI is automating the easy parts of the job, but it’s also driving growth in the messy parts that require humans. The impact on employment hinges on which grows faster.

AI is taking the easy calls

“Representative,” “Representative,” “REPRESENTATIVE.”

Calling customer service is universally frustrating. You navigate a maze of menus, thumb your details into a keypad three times, then repeat them to the human you eventually reach. Many consumers would rather just do it themselves.

Checking a balance, choosing an airplane seat, rescheduling a delivery are all faster to do online.

AI chatbots expand self-service. They’re always available, even at 2am on Friday or Christmas Eve. And for embarrassing issues like diarrhea or skin problems, consumers prefer chatbots to a human who might ask uncomfortable questions or judge them.

Self-service and live chat are growing, and on track to become the top channels by next year. But this shift hasn’t reduced calls.

Only 14% of self-service interactions actually resolve the issue, and call volume has risen steadily since 2020. Meanwhile, for the reps answering the phone, the easy work is gone, leaving only the hard stuff.

…and making the hard ones harder

Call centers are using AI, but callers are using it too, and that’s making reps’ work harder.

The WSJ recently ran a piece called “Why You Should Let AI Write Your Next Customer Complaint.” AI-assisted complaints are 7% more likely to succeed. AI complaint stories often go viral, like the woman who used ChatGPT to write a ‘polite but firm’ email about a six-hour flight delay.

AI is helping the bad guys too. Complaints to the Consumer Financial Protection Bureau have doubled every year since 2022, driven heavily by “FinTok” influencers who sell templates promising to scrub credit incidents or make money through mini-settlements. Creditors are required by law to respond to each complaint within 15 days regardless of its merit.

Scammers are going further, using AI to file for fraudulent refunds with faked photos of damaged merchandise or food delivery. The images are often indistinguishable from real damage and require human investigation.

The result is a pile of messy issues: ambiguous, often emotional, requiring judgment instead of a script. Experienced reps are the ones who clean them up.

Remember bank tellers?

Despite literally being named “Automated Teller Machines,” ATMs didn’t reduce bank teller employment. ATMs handled cash and balance transfers, but tellers were still needed for everything else: opening and closing accounts, deposits, business transactions.

But then the iPhone came along. Mobile banking was a supercharged ATM in your pocket, and it made the branch optional for many. Consumers could deposit checks, make payments and handle their banking whenever and wherever they were.

Since the iPhone launched, America has shed half of its full-time teller jobs.

Today’s AI agents are the ATM: good at simple things, useless for messy ones. The only question is when they become the iPhone.

What we know, what we don’t

Today’s dire predictions look like noise. Only 20% of companies surveyed by Gartner reduced call center staffing due to AI, and more than half kept levels stable while absorbing extra demand.

The ROI on AI agents is also unclear. Goldman Sachs found that AI call center agents were more expensive than humans because they burn computing power for the entire call.

The forecast is clear: expect more self-service, more AI, and more messy issues. And AI will get cheaper.

Call center leaders’ mandate is also clear: deploy AI and fight to keep the talent who can handle FinTok, scammers, and customers fed up with self-service.

Employees face a harder question.

The bank teller story has a second act. Teller jobs dropped by 120,000 after the iPhone launched, but the economy added 600,000 financial manager and advisor positions during the same period. Those jobs are more complicated, and they pay over four times as much.

Former tellers weren’t handed those jobs by default. The ones who got them saw it coming, earned their CFP licenses and persuaded someone to give them a chance.

Call center employees should build skills that make companies fight for them: judgment, emotional range, and the patience to go toe-to-toe with the scammers.

But the ones who succeed in the long run won’t be the ones who keep their heads down. They’ll be the ones who see what’s coming and grab it.

Share

Dad Joke: Did you hear about the framed call center agent? Just a rep with a bad rap. 😆

Would you let someone train an AI on your apartment in exchange for cleaning it?

That’s the pitch from Shift, a company that will send a professional cleaner wearing a head-mounted camera to clean your place.

They make money by selling data to labs for training AIs on real-world activity.

Three things:

• The privacy concerns are obvious. They claim to anonymize sensitive information, but that’s no guarantee. And while I might not care about big tech having a video of my kitchen sink, my roommate might.

• Some will complain this is just training on humans to replace us. Yes please! I am all for training anything that will clean my toilet for me.

• My biggest question is the numbers. Shift offers cleaners $20 per hour, and has paid out more than $5 million in Q1. That’s very expensive data. I can see the value in the first 100,000 sessions but it’s hard to see how the math squares for the millionth dirty carpet.

This reads like the classic stock market bubble company. Grab your free cleaning while you can.

The latest in running technology is endurance speedsuits that claim to improve efficiency by up to 1%.

But they also look like a wetsuit designed for a disco bar.

Tech isn’t everything. At this year’s London Marathon, Yomif Kejelcha broke two hours wearing Adidas’ Techfit Endurance Suit. But race winner Sabastian Sawe ditched the full suit and set the world record at 1:59:30.

At a 3:20 marathon pace, 1% faster means 2 minutes.

At my age I’ll do anything to pick up a few seconds. PR lotion, Nutter Butters, whatever it takes.

But can we get something that better…suits my style? 😆

Robinhood’s new press release says you can have an AI agent trade for you.

That sounds great, but what they actually launched is the ability to designate an account for a separate AI system to trade with.

Actual AI trading requires connecting an agent like Claude, Codex or OpenClaw to Robinhood’s MCP server, which only the most technical users will be able to figure out.

Robinhood’s examples aren’t new. Robo-advisors have been rebalancing portfolios for over a decade. Backtesting a mean reversion strategy can be done in a dozen different platforms.

The more interesting story is what happens when AI stock trading becomes easy for anyone.

What happens when an agent does exactly what it was told but not what the user meant? Or trades on a hallucination? Or gets compromised by a hacker?

It’s worth remembering that Robinhood makes money on trading volume, and agents will trade a lot.

Robinhood’s users have matured since its meme stock days but AI trading will be tempting for first-time investors.

We’ll need education and protections when that day comes.

Fortunately, today is not that day.

My daughter Danielle graduated from the University of Wisconsin last week. This afternoon we’re hosting her graduation party, along with her brother Grant, who just finished high school.

This letter is for her, the class of ‘26, and the parents who watched them walk.

Danielle,

Four years ago when we dropped you off in Madison, the plan was set. You’d have the time of your life and learn the ins and outs of fashion design. Before long, models would be walking the Fashion Week runway wearing your looks.

That November, ChatGPT launched. Today, AI can create runway-ready looks in seconds.

At last week’s commencement, when the speaker talked about AI, the audience booed. So did your friends, and graduates across America.

It’s strange to be booing a technology; nobody booed the Internet or the iPhone. But it’s also completely fair. This summer was supposed to be filled with road trips, weddings, and apartment hunting in Manhattan. Instead you’re wondering if you’ll spend your career working at our neighborhood Applebee’s.

It’s not just you. New college graduate employment is at its lowest level in decades. Analysts are arguing over whether this is because of AI, tariffs or remote work. Those debates are interesting, but you still need a job.

And getting a job is broken.

The old process was clear: build a resume, do the interviews, pick the best offer. Today’s graduates apply to hundreds of openings without even landing an interview. Some job seekers are embedding hidden text in their resumes to trick AIs into picking them.

This year’s commencement speakers have all the answers:

Embrace AI. So…Give it a hug?

Get on the rocket ship. Wait, I’m an astronaut now?

Get used to change. Done. But I still need a job.

Start your own company. Cool. Will that pay next month’s rent?

This advice sucks.

As a dad, I’m biologically compelled to give advice. Here it is:

Instead of sending out a hundred resumes, talk to a hundred people. Teachers, alumni, friends, Uber drivers. DON’T ASK THEM FOR A JOB. Ask them for advice and the names of other people to talk to. Most of it won’t be as good as your dad’s, 🤔 but people love being listened to. And one of those hundred people will know somebody who’s hiring.

Build something real. Recruiters spend an average of seven seconds scanning a resume. Give them something different: a portfolio, a lookbook, a custom collection. This stands out and spares the poor souls still slogging through a thousand resumes.

Get a shitty job. Applebee’s wasn’t the plan, but it’s a good start. It gets you out of the house, makes a few friends, and earns some money. You’ll quit that job when you find a better one.

One final thing: make your boss like you. We older folks complain that today’s grads are entitled. Show them you’re not. Do the grunt work. Be easy to manage. Smile.

Permit me one more bit of dadsplaining.

This is part of a bigger trend. In fields most exposed to AI, jobs for workers 22-25 are down 13 percent while older workers are up 9 percent. AI agents need to be managed by humans with experience, so companies are cutting college hiring. But that will leave them short-handed when older workers retire.

Entry-level hires need the same things AIs do; companies just call them by other names. Instead of context, it’s new hire orientation. Instead of guardrails, it’s defined roles and responsibilities. Instead of reward signals, it’s quarterly goals. The irony is that we’re learning to onboard AIs at the same moment we’re forgetting how to onboard people.

Most executives reached the top because they knew how to develop people. Giving someone their first job, showing them the ropes, helping them grow into a leader. The best moments of my career have been promoting someone who deserved it.

Your AI agent isn’t going to say thanks, invite you to its wedding or bring its new baby into the office.

Okay Danielle. You have a graduation party to get to.

We don’t get to choose the time we live in. When I was in college, the hot new technology was pay-at-the-pump. I saw the internet, the iPhone, and now AI, and was lucky enough to work at Google before big tech turned into the hunger games.

People say AI will be the biggest disruption to society since the Industrial Revolution. And your graduation landed smack dab in the middle of it.

Analysts love to predict the future, but the honest answer is nobody knows.

We won’t know everything by the time Grant gets his degree, but we’ll know more. I hope I won’t have to write a version of this letter to him.

You’ll be fine. I’m here to help. One of those hundred people you talk to will land you a job.

I’m proud of you. Now go enjoy your party.

Share

Dad Joke: What do you call a university dedicated to lettuce? The college of cabbage 🤣🥱

Atlanta’s Cox Media Group promised to “use voice data to target your advertising to the EXACT people you are looking for.”

Except the smoking gun was an empty water pistol.

Yesterday the FTC ordered CMG to pay nearly $1 million for lying to customers about its “Active Listening” AI service, which turned out to be overpriced email lists.

I heard this all the time when I was at Google. People were creeped out by seeing an ad for something they just mentioned to a friend.

Big tech does plenty of creepy things, but this isn’t one of them.

Active Listening would be illegal and a massive PR risk. Not to mention the poor AI who has to sift through trillions of hours of white noise and family arguments waiting for someone to say Nutter Butters.

But if it isn’t big tech, maybe someone…else is listening? 👽

$100 million in lost pizza sales due to AI! That’s a lot of pepperoni. 👀

This from a franchisee’s lawsuit alleging Pizza Hut mandated use of its “Dragontail” AI delivery system.

Pizza Hut surely would have preferred to settle this quietly. They didn’t, and we get a useful case study.

Here’s the story:

• Pizza Hut’s AI gave DoorDash drivers visibility into orders, bake times and tip amounts.

• Drivers gamed the system by refusing low-tip orders and grouping deliveries.

• On-time deliveries dropped from 90% to 50%.

Compare this to Domino’s. They are on DoorDash, but negotiated to use it only for orders. Domino’s pizzas are delivered by Domino’s employees.

It’s easy to point the finger at AI, but the system is working as designed. Pizza Hut’s mistake was outsourcing its bottleneck to a third party whose incentives didn’t match its own.

Give up control of your bottleneck, and your customers end up with cold pizza.

Data centers are as popular as landfills, casinos, and prisons.

Last week’s Gallup survey found that 71% of Americans oppose building a data center in their area. Only 53% opposed a nuclear power plant.

Ouch.

Across America, activists are protesting data centers, and they’re winning: In 2025, 48 projects worth $156 billion were blocked. Bernie Sanders and Alexandria Ocasio-Cortez are co-sponsoring a bill that would halt all US data center construction, and moratorium bills are active in twelve states.

Data centers aren’t perfect: they’re power-hungry, noisy and don’t create many permanent jobs.

But a moratorium is a bad approach, especially as cities are winning at the negotiating table.

The capacity crunch

Demand for AI is exploding, and companies are looking everywhere for new capacity, including outer space and the walls of newly built homes.

Anthropic just struck a deal to use xAI’s Colossus data center in Memphis. In 2024, Colossus was the biggest AI training cluster in the world. Today it will only support six to nine months of Anthropic’s growth.

Data center construction drove 92% of GDP growth in the first half of 2025. It’s also funding long-overdue investment in the US electrical grid, parts of which were built in the fifties. Tech demand is driving $1.1 trillion in grid upgrades. Microsoft and Google alone are funding the equivalent of 48 nuclear power plants in new renewable energy.

The negotiating table

It’s easy for economists to cheer for data centers. It’s a different story when one goes up in your backyard.

The biggest pushback is about water use, an issue tech bros dismiss as a red herring.

Red herring or not, data centers create real issues for local residents. They are electricity hogs, and have jacked up utility rates by as much as 20%. The gas turbines they use for short-term power are dirty and noisy. And while construction creates temporary jobs, a finished data center employs far fewer people than a warehouse or factory.

Communities are negotiating hard to address these issues. New deals make companies fund new power capacity, pay for grid upgrades, install noise suppression and offset local electricity bill increases.

A proposed data center campus in Janesville, Wisconsin would pay to clean up a contaminated 250-acre former GM plant. The site sat dormant for 17 years because the cleanup’s $30 million price tag was too steep for other projects to justify.

The Irish lesson

In 2021, Ireland led the EU in data center capacity. But its data centers were consuming more electricity than every urban home in the country, and growing so fast the grid operator warned of blackouts.

Irish regulators issued a moratorium on new construction to stabilize demand. Construction froze, and €8 billion in projects shifted to Sweden, Norway, and other European countries.

The ban was lifted in 2025 and replaced with a policy requiring companies to fund their own power needs.

But the damage was lasting. Analysts don’t expect Ireland to recover its momentum until 2030.

This is the same policy lawmakers are proposing in the US.

A lousy idea

Data centers aren’t perfect, and their faults have become the face of AI to many Americans.

But well-struck data center deals bring real benefits to communities like Janesville, paid for by tech companies who can afford it.

A moratorium would strip this leverage from local leaders and kneecap a domestic economic boom.

Democratic senator Mark Warner called a moratorium ‘idiocy’, warning it would hand America’s AI lead to competitors.

He’s right. A ban wouldn’t make data centers cheaper, quieter or cleaner. It would just make US companies spend money to build them overseas.

That’s a lousy deal.

Share

Dad Joke: Why did the server farm always answer yes to RSVPs? Because it was a date assenter 🤣

These AI videos of a man videobombing the Avengers, Home Alone and It are laugh-out-loud funny.

They’re also a good reality check: we’re a loooong way from AI TV shows and movies. His videos are 10-15 second clips stitched together because that’s when AI quality starts breaking down.

Someday AI may be able to create movies and TV shows, but today isn’t that day. Neither is next year.

Short-form video still struggles with reliability. This year’s Coca-Cola holiday ads still mixed up the number of wheels on their trucks. If it can’t do that, imagine asking it to remember a character arc across a twenty minute comedy.

That’s before you get to rights and likeness. Does this qualify as fair use because it’s a parody?

But constraints force creativity. Between these, the oddly cringey Fruit Love Island, and the real estate agent selling homes as Iron Man, people are using AI to create things we couldn’t before.

For now, let’s enjoy watching this guy steal Thor’s hammer and punch out Pennywise.

AI’s New Bottlenecks: Remember Herbie?

Alex Rogo runs a factory that’s falling apart. Herbie is an overweight Boy Scout with an overloaded backpack.

They are two unlikely characters in The Goal, a 1984 business novel that taught bottlenecks to more than ten million readers. Forty-two years later, it’s still required reading at most MBA programs.

In its most famous chapter, Alex is leading his son’s Boy Scout troop on an overnight hike. Herbie keeps falling behind, and Alex has to make the other kids stop so he can catch up. They eventually reach camp tired and behind schedule.

The next morning, Alex stops cajoling. He puts Herbie in the front so the troop has to keep his pace. The boys realize if Herbie speeds up, they all do. They offer to carry some of his weight, and they make the return trip ahead of schedule.

Herbie is a bottleneck. And every process has a Herbie: the slowest step that holds back all the others.

Today, Alex Rogo would have his hands full. AI is creating new bottlenecks and overwhelming others with an avalanche of AI output.

Bottlenecks on the move

Ground zero for shifting bottlenecks is AI-assisted coding.

A 2026 Faros study of 22,000 programmers found that they are delivering 66% more features thanks to AI.

This productivity comes with a cost. Last year’s study showed bugs had risen by 9%. Twelve months later, that number had exploded to 54%.

The volume of AI output also means new code waits more than five times longer for review than it did a year ago.

Rockstar developers are suddenly spending their days slogging through AI-generated code instead of writing their own. Facing a new bottleneck (and crabby rockstars), companies are building AI into their quality processes. At Cloudflare, new code is reviewed by seven specialized AI agents before reaching a human.

In software development, bottlenecks have moved. In consumer goods, they are piling up.

An infinite aisle of ideas

Try this: “Hey ChatGPT, give me ten ideas for Nutter Butter flavors targeted at Ironman athletes.”

Yum. Now imagine doing this with a model trained on Nabisco’s flavor research, product portfolio, sales history and consumer trends. That’s what parent company Mondelez did.

Their snack-development AI cuts the time from concept to production by 75-80%. They’ve used it to launch 70 new products, including Gluten-Free Golden Oreos.

Coca-Cola, P&G, Unilever, and Nestlé all have similar AI platforms to develop concepts. Some go beyond design; testing flavors with synthetic AI personas.

Generating ideas has gotten easier. Getting them on the shelf hasn’t.

The gauntlet tightens

It’s easy to come up with Birthday Cake PR Nutter Butters. It’s a lot harder getting through the real-world obstacle course:

• Can we actually cook them?

• Can the factory mass-produce them?

• Can we afford the inventory?

• Will Walmart put them on the shelf in 10,500 stores?

• Will the packaging sell them in three seconds?

With every company chasing the same shelf space, running this gauntlet has become like scoring tickets to Taylor Swift’s Eras Tour.

Winning at the bottleneck: Dominos

A 2026 industry study found delivery speed is the single biggest driver of pizza customer satisfaction. Faster deliveries mean repeat customers, more sales and bigger driver tips.

Domino’s is using AI to start cooking before orders are placed. They track customers during their checkout process, and when the likelihood is high enough, the pizza goes in the oven. They’ve done the math: the time saved outweighs the cost of wasted dough.

Domino’s calls itself a “tech company that sells pizza.” They’ve made that real by not just using AI, but deploying it at the bottleneck: pre-baking, GPS delivery, cameras scanning every pizza before it leaves the store. All focused on one goal: getting hot pizza to your door fast.

Back to the factory

Alex Rogo’s factory had a sexy new machine called the NCX-10. It was the most efficient machine in the plant.

Naturally, they decided to use it for everything. They even retired other machines that worked but weren’t quite as fast. Before long, the NCX-10 got overloaded and parts stacked up waiting for their turn.

Alex solved the bottleneck by bringing back the retired machines to lighten the NCX-10’s load.

The risk with AI is the same. It’s new, fancy, and it makes for a great resume bullet.

But Alex Rogo didn’t improve his plant by buying faster machines. It got better because he understood where to use them.

Share

Dad Joke: What does Dominos do with pizzas when customers cancel? They eat them. 😂

While schools grapple with the Canvas ransomware attack during finals week, Anthropic’s Mythos is the real deal for IT security.

In April, Mozilla fixed 423 security bugs, more than the 258 they fixed in all of 2025.

In yesterday’s blog post, they revealed their new AI-based approach for rooting out security bugs. It found gaps that had survived for decades, including a 20 year old bug in their XSLT code.

Quick, use “Extensible Stylesheet Language Transformation” in a sentence. 😂

The clock is ticking. Security expert Alex Stamos estimates that we have 6 months before other frontier models reach Mythos-level capability, 12-18 before it is available in open-source models.

This is good news for Firefox users. Other software makers should follow Mozilla’s example and incorporate AI tools in their testing practices. And companies should make sure their software vendors are doing the same.

Quick, let’s all take a minute and laugh at the Waymo afraid to drive through a puddle.

That was fun! 🤣

Now let’s think about what’s going on here. Waymo trains its cars to be conservative, stopping for human intervention to avoid incidents.

Waymos have 92% fewer serious accidents than human drivers. They could take a lot more chances and still be safer. But they don’t.

This is a smart move for two reasons:

1. Unlike human mistakes, Waymo accidents become major headlines. You’d rather make the news for a puddle than a pileup.

2. When you’re asking people to trust a robot with their lives, “grandma with glasses” is a better brand than “Mad Max.”

Last week a robot ran a half-marathon faster than the world record. It had its own course, battery pit stops, and human helpers, but still finished six minutes ahead.

AI now diagnoses cancers that human radiologists miss, forecasts weather better than meteorologists, and sets bail with less bias than judges. In business, AI makes better hiring decisions, underwrites loans more profitably, and reviews contracts more accurately than humans.

The obvious question leaders ask: If AI is better, should I cut the humans out of the process? This makes sense. It drives bottom-line savings and plays great on earnings calls.

But that advantage won’t last. The AI you’re replacing people with is the same one your competitor will roll out next quarter.

In a world where everyone has the same AI, the better question is: what should humans be doing that AI can’t?

Research is answering that question, and Walmart is making it real.

The easy answers aren’t the best ones

Conventional takes on humans working with AI sound smart but rarely hold up.

Use AI for routine tasks, and humans for creative ones. Sometimes, but AI is better than most humans at writing poetry and can’t fold laundry.

AI generates options. Humans make decisions. Maybe, but setting bail and approving credit are pure decisions, and AI beats humans at both.

AI can recommend; humans should verify. Not always. Studies in supply chain and hiring find that verifiers often make things worse, not better.

Sometimes, maybe, and not always don’t make good slogans, but they’re reality.

Dealing with that means making two design choices: what to delegate to AI and what only humans can answer.

What to delegate

If the right part isn’t on the shelf when an auto mechanic needs it, they’ll go to a competitor.

At a major auto parts retailer, inventory planners regularly override the AI’s demand forecasts when they think the AI is wrong. To test whether those overrides actually helped, researchers randomly accepted or ignored them.

Most of the time, the planners were wrong. Across 18,000 SKUs, seven out of ten overrides either reduced profits or did nothing. That’s thousands of hours of expert attention making the company poorer.

The other 30% were the opposite: profits went up by 6.8%. They fell into three groups: high-margin products, new launches, and large suppliers.

Adding human judgment to every task is like spreading peanut butter on everything. Save it for celery and apple slices.

What humans know that AI doesn’t

Predicting surgery duration is critical for hospitals. Operating rooms cost over $2,000 per hour, and bad estimates either waste capacity or delay patient care.

AI predictions were more accurate than human ones. But surgeons know things AI doesn’t, like the patient’s emotional state and family support system.

Researchers tried a different approach. Instead of asking humans to predict duration, they asked what AI was blind to. When humans flagged those blind spots, predictions improved by 21%, with the biggest gains on the most complex cases.

A second design choice helped even more. When humans were told AI was operating outside its training data, they leaned more heavily on their own judgment, and errors fell by 49%.

AI knows kids like peanut butter on celery. Humans know Johnny’s allergic to nuts.

How Walmart does it

75 million products, 10,800 stores, and $59 billion in inventory. If any company needs AI, it’s Walmart. They’ve built processes where humans and AI together beat what each could do alone. Here’s how:

Pricing: Walmart merchants delegate routine markdowns to an AI algorithm. When the AI hits incomplete data, it sets a price range and kicks the decision back to the merchant, who weighs hidden factors like brand messaging, supplier negotiations, and competitor behavior.

Waste: Walmart associates use an AI tool to manage perishables. When lettuce is about to expire, the AI flags it and suggests three options: mark it down, ship it back, or donate it. The associate, who’s holding the actual lettuce, makes the decision.

Vendor negotiations: Walmart’s buyers manage over 100,000 suppliers, too many to negotiate with personally. They deployed an AI chatbot to handle long-tail conversations within parameters set by human buyers. It closes 68% of deals within those limits. When a deal falls outside them, a buyer takes over. Three out of four suppliers prefer it to a human.

The advantage you can’t buy

Remember that record-breaking AI robot? In the final stretch it crashed and had to be helped back up by its human builders. It still won, but humans built the course, swapped the batteries, and picked it up when it fell.

In three years, every company will have AI. The advantage won’t come from using AI but from using it well.

That means deciding what AI should do and what only humans can bring.

The companies that get this right won’t just keep pace. They’ll have an edge that others can’t buy.

Share

Dad Joke: What did the AI robot eat for its pre-race meal? RAM-en noodles. 😂

Thanks to Jordan Tong and Tarun Kushwaha at the Wisconsin School of Business for the research this article is based on.

Hello readers -

I’m excited to share some news: I’m renaming this newsletter to AI InSeitz. Here’s why:

When I started writing a year ago, AI was new, magical and mysterious. We’d just had the “DeepSeek moment” and people were Ghibli-fying their profile pictures. A year later, the magic has mostly faded and people are wrestling with what AI means.

Hence AI InSeitz. Yes, it plays on my last name, and it fits what readers tell me they value most: perspective.

Articles will stay the same: the latest topics on AI in society and business, with a mix of context and perspective. They’ll take 4 minutes to read, not counting the dad joke.

It’s been great to hear the newsletter resonates. That’s the real payoff. Thanks for being part of the journey.

This edition is already a bit different, so I’m including a report from the marathon I ran on Sunday. Fair warning: it has almost nothing to do with AI.

Marathon Race Report: Jim Thorpe, Pennsylvania

Marathon number fifteen. The course is ranked the 24th fastest in the US, which gives me a great shot to improve my Boston qualifying time.

It rains all day Saturday, so instead of enjoying small-town Americana, Nadia and I hunker down in the hotel. Sorry, honey.

The rain clears by Sunday, and temperatures are in the low fifties, perfect for running. I’m ready with my carbon shoes, baking soda and legs coated in ‘PR lotion.’ PR here we come!

Mile 1: My Garmin watch flashes a performance condition of +10, which basically means I’m Eliud Kipchoge. Is this real or is Garmin’s AI just glazing me?

Mile 2: We pass a group of runners and are forced into a long stretch of mud running. The rest of the course is a mix of soggy and muddy. Carbon supershoes were not made for mudding.

Mile 4: The path drops down to hug the Lehigh River. The next twenty miles are filled with gorgeous views.

Mile 5: A group of spectators in Harry Potter robes with a sign reading “RavenCLAW through the course so you can SytherIN the finish.” 👀

Mile 7: The aid stations have UCAN gels! I grab some and realize they are Strawberry-Banana, which tastes like a smoothie made with Tums and sidewalk chalk.

Mile 8: Last week a robot ran a half-marathon 6 minutes faster than the world record. Meanwhile, I’m gagging down a sidewalk-chalk smoothie. Does this mean robots will take all the Boston slots? 🤔

Mile 10: I pass two runners with hoodies reading “Swagga House. Tattoo Parlor and running club.” Do they let 54-year-old dads hang out in the Swagga House? Don’t answer that.

Mile 12: A long stretch of uphill. I was told this was a downhill course. 😠

Mile 14: Into the porta potties for a pitstop. F1 pitstops take 3 seconds. Mine is 45.

Mile 16: Fatigue sets in, right on schedule. I feel like I’m wearing one of those weighted vests.

Mile 17: I pass a sign saying Jim Thorpe would be impressed. I’m very sure he wouldn’t.

Mile 19: Another sign: Chuck Norris never ran a marathon. Marathons run from Chuck Norris.

Mile 23: I feel like I’m running in eighteen weighted vests. One. More. Mile.

Mile 26: In an act of true cruelty, the final mile marker is 1.25 miles past the last one. All the AI in the world and we can’t measure a mile correctly?

I finish in 3:27:52, well short of my PR. Can I get a refund on the PR lotion?

No PR, but I got to spend a Sunday morning running next to an idyllic river, on legs that still work at 54. Maybe a robot will run this faster, but it won’t savor it.

Share

Dad Joke: Why did the runner finally remember his PR time? Something jogged his memory 😆

“This is where all Meta ‌employees ⁠can help our models get better simply by doing their daily work”

Meta announced it will record keystrokes, mouse movements, and screenshots to train its AI models and coach employees.

Meta CTO Andrew Bosworth said there’s no opt-out. Negative press and employee backlash are already building.

This data is valuable. Recording voice calls with tools like Gong is standard practice for coaching, and increasingly, model training.

Meta’s position is understandable; these are Meta’s computers used by Meta employees to do Meta’s business.

Two second-order effects I’m watching:

1. Off-hours working: If their company laptop becomes purely a work machine, will employees still bring it home? Will a manager knock out a couple emails after planning their kids' birthday party? For years, companies have benefited from employees blending work and personal lives. Screen capture could end it.

2. Legal discovery: This will be a subpoena goldmine. Not just sent emails but drafts. Documents viewed but not authored. Data covered by regulatory protections. The volume of new material available to lawyers will be staggering.

Microsoft was forced to neuter its Windows Recall feature due to backlash from privacy advocates. This was a feature consumers could opt into for their own use. Meta’s version is the opposite.

Corporate America will be watching. If Meta weathers the backlash, expect this to spread quickly.

I’d prepared fifteen slides, a dozen questions, and an excerpt from my AI strategy class. After three questions we’d blown through an hour.

Boy, did they want to talk.

It was the second meeting of our AI Hub advisory board, with the same executives I’d hosted six months earlier.

In November one executive described the mood as “optimistically scared S**less.” Six months later, another said “I’m not any less busy. I’m seven thousand times more busy, and more stressed, and more annoyed.”

Our last meeting was a thoughtful strategy session. This felt like a pit stop for drivers with overclocked engines.

Today’s executives are leading teams through a time of structural change: who does the work, how it’s done, with technology that keeps changing. They’re doing it without a roadmap.

Where the buck stops

Executives have always been accountable for their teams’ work, but now they’re on the hook for AI too.

Last year Deloitte had to refund the Australian government after the firm’s report had errors traced to AI-generated content. People laughed, but zoom out and it’s easy to see how they slipped through the cracks. Deloitte has 470,000 employees, and runs hundreds of thousands of engagements each year. Their quality processes were built for human outputs, not AI-generated ones. And the volume of AI work is growing faster than teams can check.

That’s just the humans. AI systems are equally unpredictable. Meeting business requirements isn’t a yes-no question. It’s a matter of how often the AI gets it right, and whether it holds up when a thousand users use it a thousand different ways.

A manager can spot-check individual outputs, but an executive responsible for thousands of AI workflows can’t. More and more companies are turning to AI as the best way to verify growing mountains of AI output.

The buck stops at the executive’s desk. The leaders of the Deloitte engagement have left the company.

The right amount of fear

Every company has employees who are excited, fearful, and somewhere in between. That’s one thing in a team; it’s another in an organization of thousands. One leader had to decommission an internal policy chatbot after employees manipulated it into giving incorrect answers. Sabotage or not, employee resistance is a growing trend.

Executives wrestle with how to motivate employees: should they urge, reward, or threaten? Push too little and people coast; too much and they freeze up. The academic term for this sweet spot is productive discomfort. One leader knew he’d found it when employees volunteered to lead their AI role redesign.

Motivation is only half of the challenge. The other half is that every role and team is affected differently by AI. Some are 10% automated, some 50%, and some busier than before. In every case, leaders have to decide whether that warrants minor adjustments or an overhaul.

The unused gym membership

One speaker had built an AI insights tool for sellers. It was powerful, but lived separate from the company’s CRM, where sales reps actually spent their time. So nobody used it.

Other executives had the same story. They deployed a new AI tool, only to see adoption stall because it didn’t fit the existing process. That’s reasonable. An improvement that doesn’t fit a process refined over years can slow things down.

The best deployments answer two questions. First: where does AI fit our existing workflows? Second: what work does AI replace? Skip the first and you build something nobody uses. Skip the second and you make more work instead of improving it.

The second question is the new one. Previous software helped humans do their work. AI raises a different question: what work still needs a human? Executives are making these calls at scale.

The toughest discipline is knowing when to stop. Leaders are setting “quit criteria”: a point where you pull the plug regardless of sunk costs or political fallout. Without this, teams waste time and money on zombie solutions. Executives have always done this. But when deployments took years, they did it once. Now they do it monthly.

Where’s the finish line?

AI was supposed to give us time back. One speaker described it as eliminating “pajama time”: unplanned work that gets tackled at night or on weekends. But for executives AI isn’t reducing pajama time, it’s multiplying it. And it’s making them busier, more stressed, and more annoyed along the way.

We don’t know where this ends. There’s no earnings report, no deal close, no system go-live. And the technology keeps shifting; the systems we’re deploying today will be different next month.

I’ve learned a lot about endurance from fourteen marathons and four Ironmans. When it gets tough, the strategy comes down to one thing: don’t worry about the finish, just run the next mile.

But this race is different. There’s no map, no clock, no finish line. Endurance helps, but it’s not enough.

The most successful leaders aren’t the ones who can tolerate the uncertainty best. They’re the ones who’ve stopped looking for a map and are drawing their own.

Share

Dad Joke: Why did the morgue employee want to be CEO? So he could have a coroner office. 😂