The worst 48 hours in cybersecurity history.

“Black October” had already seen a rash of attacks. Hackers exposed patient records at Quest Diagnostics and data on millions of California public school students. They even overheated refrigerators in French grocery stores.

Then last Wednesday, details leaked about two bugs in PHP, the software behind 70% of the internet’s websites. The bugs can let anyone take over a website’s server without a password.

Within hours, hacking attempts surged into the millions. Governments across Europe shut down websites, and the NSA urged organizations to take systems offline if they couldn’t be patched immediately. The UK government told Britons to stockpile food and water for cyberattacks and other emergencies.

Between Y2K and data breaches at Target and Equifax, we’ve become accustomed to cyber scares. Some fade away, but others are painful. In July 2024, a faulty CrowdStrike update crashed Windows computers and canceled summer vacation flights across America.

Should we keep calm and carry on, or stock up on beef jerky and canned green beans?

Why now?

Two things, both AI-related:

First, AI has gotten scary good at finding software vulnerabilities. The PHP bugs had existed for over a decade without being noticed. Thanks largely to AI, researchers found four times as many critical bugs this October as last October.

Second, AI lets attackers probe thousands of websites at once, for pocket change. A Chinese-speaking hacker used AI agents to break into the websites of over 100 companies, including Deere, Cummins and SharkNinja for about $50 per breach.

“Ondřej Surý will save the internet today.”

Ondřej Surý is a software engineer in Czechia who, as a side project, has spent 20 years packaging PHP fixes into patches website owners can install in minutes instead of weeks.

Researchers found the bugs in August but kept them secret while they built a fix. Days before the announcement, Surý got the call. On release day he published dozens of updates covering PHP versions going back over a decade.

Millions of people like Surý quietly hold the internet together. They’re the reason Y2K was a non-event.

Tech titans, banks and other big companies have legions of these pros, and they are often warned about issues early. The weak spots are mid-sized organizations: supermarkets, hospitals and water plants, big enough to run their own technology but too small for an army of security experts. Their people will be scrambling over the next year.

So what about you and me? The internet is ready with answers.

Advice overload

Matt Shumer’s “The Internet is Under Attack. Here’s How to Protect Yourself and Your Family” includes sensible ideas like two-factor logins, but also says to stay off small websites and unplug smart gadgets. Hardcore preppers are buying portable generators, Starlink Minis and weeks of ready-to-eat meals.

Being prepared is good, but people are busy and endless advice can feel like the dentist who tells you to floss, rinse twice a day, and use a tongue scraper.

I asked ChatGPT what would happen if a hacker broke into my Garmin account. Its answer: “It could see GPS routes and potentially identify your home, regular training locations and when you’re away.”

So a burglar could study my workouts and target my house when I’m out for a run. But a criminal who can hack my Garmin is wasting their talents on my home.

The family password

ChatGPT also told me to stop reusing passwords. Simple enough, but Google Chrome says I’ve logged into 507 different websites over the past decade. A password manager can remember them all, but changing them is a huge chore.

Many families share a password, which is convenient for low-risk sites. If a hacker reviews a book on Goodreads, nerfs my Chess.com rating or exposes my secret lasagna recipe on Allrecipes, the world will keep spinning.

Shared passwords are handy, but when one site gets breached, hackers try that password everywhere else. So assume any password you’ve used more than once is already out there.

That’s fine for Goodreads, but not for the accounts that matter: your primary email, Google, Apple, social media, bank and investment accounts and other things you care about like health records.

For these, use a unique, strong password plus two-factor authentication: a second check, such as a code from an app like Google Authenticator or one texted to your phone.

Also turn on “SIM protection,” which stops hackers from hijacking the phone number your text codes go to. I did this with two clicks on Verizon. AT&T and T-Mobile offer the same thing.

Your identity is already out there

Thanks to data breaches at Equifax and National Public Data, your Social Security number, address, and date of birth are likely circulating via the dark web. Hackers can use these to impersonate you.

Here’s how to protect your identity:

Freeze your credit reports at Experian, TransUnion and Equifax. This blocks anyone from opening a credit card or loan in your name. It’s free, and you can lift the freeze temporarily when needed.

Create an IRS identity protection PIN. This will block someone from filing a tax return in your name.

Be suspicious of phone calls, texts or email about your accounts. If someone asks for personal information, a code or money transfer, hang up and call back using the number on your card or official website. No real company will ask for a code over the phone.

Stay safe out there

These steps cover the biggest risks for most people. Things could get worse, but if hackers start attacking your 401(k) through your wireless printer, it might be time for a portable generator.

There’s an old joke about a pair of hikers who heard a bear growling at them. One stopped and put on his running shoes. His partner said, “You can’t outrun a bear.” He answered, “I don’t need to outrun the bear. I just need to outrun you.”

Hackers go after the softest targets. You don’t need to be perfect, just good enough to get skipped.

Thanks for reading! Please consider sharing this edition with friends and family who you want to help stay safe.

Share

Dad Joke: What do you call a bug fix that comes out on Halloween? A pumpkin patch.

This is one of the better AI-generated music videos I’ve seen. My kids still hate it.

“C’est de l’IA” is a catchy pop song with crisp AI visuals. The X post racked up 1.7 million views.

The artist’s name is “ClaudiaGI,” a mashup of Claude and CGI. The song made using Claude, Suno and a few other tools. Beyond the visuals, you can tell it’s AI because of the short clips stitched together.

This is a new genre called French “slopcore,” which leans into the AI slop label. Instead of trying to pass AI off as human-made, it makes a joke of it.

The entire song is “her” being accused of AI. She misspells words, acts mean and posts embarrassing videos. But even her mom still says “It’s AI!”

I’m curious: will “art” that’s fully AI be more palatable to people who hate AI-generated music and video? Or is there no place for it whatsoever?

That's what makes this such a boneheaded move.

Developer Nscale is building a massive 2,250-acre AI data center campus in West Virginia, with Anthropic as a customer and Nvidia as a partner. The site includes a nursery with two dozen mature American chestnut trees.

A blight killed 4 billion American chestnuts in the early 1900s. This site is one of the few remaining mature reproducing groves in the country.

After protests, Nscale denied plans to bulldoze the nursery and announced it will transplant the trees to a new location.

This, despite the fact that the American Chestnut Foundation knows of no successful relocations of mature American chestnuts. One of its experts called the plan “ludicrous.”

Nscale is within its rights and has a massively complex construction schedule to meet.

But come on.

The grove is 𝟭.𝟮 𝗮𝗰𝗿𝗲𝘀, 𝗼𝗿 𝟬.𝟬𝟱𝟯% 𝗼𝗳 𝘁𝗵𝗲 𝘁𝗼𝘁𝗮𝗹 𝘀𝗶𝘁𝗲. It’s located on the edge of the development.

Why not just build around it? Why not designate it a conservation area and name it “American Chestnut Data Center”?

Saving 20 trees would have cost almost nothing and earned a ton of goodwill. Instead, we’ve got another story about evil data centers. And this one is deserved.

Last week Sega pulled a remastered version of Total War: Shogun 2 from the market and issued an apology, saying the game had “fallen short of both your expectations and our own standards.”

Gamers lambasted its remastered artwork for AI-generated mistakes including melted armor, a soldier impaled by a ship’s mast and a samurai with six fingers. Others tagged it a “Psychological Horror” and called it “Literally the greed they talk about in the Bible.”

The #1 rule for deploying technology is don’t piss off your customers. Sega broke that rule, and its mistakes are a warning to any company relying on a partner for its AI.

Wrong game, wrong audience

Remastering a game is normally easy money. Slap a new coat of paint on an old title and sell it to people nostalgic for the original.

But who buys a remake of a 15-year-old strategy game? Not players killing time on their commute, but ones who played the original for a hundred hours and remember the tiny details.

Gamers are split on AI-generated art. In a survey of predominantly core and hardcore gamers, 85% had a negative view of generative AI. But in a broader survey, only 10% of adult gamers viewed AI art negatively.

Shogun 2’s buyers are in the first group. A gamer who will fork over $40 for a fresh coat of paint is the same gamer who will inspect every brushstroke.

The governance blind spot

Just a week before Shogun 2’s launch, executive Tsuyoshi Saito said the company was “very cautious about its use in the game development process, and we are not entrusting the creative aspects of games to AI.”

His words are backed by strong AI governance: guidelines, task forces and committees staffed with legal and PR pros to ensure AI doesn’t slip through the cracks. Even so, fan complaints forced Sega to clarify its AI use on Crazy Taxi World Tour this summer.

The cracks in governance split open when Sega outsourced the remake of Shogun 2 to a third-party developer. Outsourcing is common in the industry, but requires a different level of oversight than managing your own team. Sega’s committees were built for policies and contracts, not inspecting imagery frame by frame. That job takes humans with the same eye for detail as your biggest fans.

You can bet Sega’s lawyers are poring over that contract for “no AI” language, but a lawsuit won’t undo the damage.

The AI gamers want

The original Shogun 2, launched in 2011, was touted for its cutting-edge AI that “constantly analyzes its situation and reacts to your every move with greater precision and variety.” But that AI didn’t draw art.

Gamers love AI when it makes a game more realistic. Instead of fighting a boss who makes the same moves every time, you’re hunted by an alien who learns where you hide and what your footsteps sound like. Instead of a shopkeeper who’s a pop-up menu with a grin, yours will catch you stealing, blacklist you and meet you in an alley with a shotgun.

Gamers also benefit from AI that writes code, fixes bugs, and improves performance. Gaming platform Steam requires disclosure for AI-generated art, but not AI that’s used behind the scenes.

The line is pretty clear: gamers want AI that makes the game better. They don’t want AI that replaces the handiwork they paid for.

My favorite games, Bloodborne, Sekiro and The Last of Us are all known for obsessive attention to detail. I’d welcome an AI that reduces bugs and makes their bosses smarter. But if the Last of Us III features a six-fingered Ellie, I’ll join the mob with a pitchfork.

The $3 billion test

Almost anyone can make a game with AI. One of my childhood favorites was “Wavy Navy” on the Apple II Plus. ChatGPT’s Codex built a playable version in an hour.

Thanks to AI, app stores are flooded with new games. 181,000 new titles launched in the first 6 months of 2026. The flood hasn’t translated to usage; the top 1% of companies took in over 90% of sector revenue last year.

When throwaway games are everywhere, a game that costs $40 and takes a hundred hours to play better be great.

That’s the backdrop for the gaming event of the decade. Grand Theft Auto 6 has been called “one of the biggest entertainment launches in history, not just in games but for all of media.” It’s projected to make over $3 billion in its first year.

Anticipation is sky high. One Army commander offered a special four-day pass to soldiers who re-enlist. An NBA player predicted sports league performance would dip due to distracted players.

Grand Theft Auto’s builder Rockstar has had its own AI struggles. Five years ago, it issued an apology after fans dubbed GTA: The Trilogy - The Definitive Edition “The Defective Edition” for shoddy AI graphics. Just like Sega, Rockstar had outsourced the game to a third party.

The company is trying to thread the needle. Strauss Zelnick, CEO of Rockstar’s parent company, said it uses AI and machine learning heavily, but that it has “zero part” in the new game. The game’s setting, Vice City, is being crafted “building by building, street by street.”

He’d better hope so. Given the amount of hype, you can expect gamers to scrutinize every city block.

The #1 lesson

Sega’s lessons aren’t complicated, but are easy to miss.

Use AI where your customers want it and not where they don’t. With partners, a contract isn’t quality control. If it’s your product, inspect it before you launch it.

Otherwise, you might end up with a six-fingered samurai.

Share

Dad joke: What did the pirate in Shogun 2 say? Samur-aye-aye captain!

This AI-to-AI interviewing video is a great laugh. It’s almost certainly staged.

Comedian Graham Dot Zip had an AI agent take an interview on his behalf. It answered questions and negotiated salary, but couldn’t…quite…figure out how to say goodbye.

It’s easy to decry the rise of AI interviewers, but they are understandable for recruiters drowning in resumes.

The average posting receives over 250 applications, more than double the number before ChatGPT. If a diligent recruiter screens 20 people, 230 get rejected based on resume alone. That often means to finding tiny reasons to pick between qualified applications.

AIs allow companies to consider everyone who might fit. A good interview process follows AI screens with live, ideally in-person interviews.

Surprisingly, an experiment with 70,000 call center applicants in the Philippines found 78% of candidates chose AI over human interviewers. Those interviewed by AI got more offers and had higher retention. (Humans still chose who to hire.)

So absolutely let’s dunk on the AIs, they won’t take it personally.

But remember, AI could be the difference between a live conversation and a quick rejection.

Apple’s new Siri Recap will summarize your conversations. It doesn’t keep transcripts, so who’s to say Siri was wrong?

Probably not us. New research shows AI can rewrite our memories.

Today’s Washington Post features an MIT study where subjects shown AI-altered photos remembered things that never happened, and were more confident in them.

In a second Georgetown study, people watched a video of a car hitting a pedestrian after running a stop sign. After reading an AI summary saying it ran a yield sign, only 45% remembered correctly vs 84% with an accurate version. Knowing the summary was AI-written didn’t help.

Human review is a core tenet of most AI use policies. But if AI mistakes become our memories, the value of that review erodes.

When Google came along, we stopped remembering facts because they were one search away. But Google gave us the source to check.

AI meeting summaries replace the source. Companies that make them should be held to a high accuracy standard. And wherever you can, keep the transcript.

“Before I can even finish ordering the burrito I’m getting interrupted to ask if I want jalapeños on it. No. Sour cream?? No, thank you. Upgrade the meat? NO, for #$% sake. I haven’t even gotten to the tacos yet.

So I drive away. Hopefully Burger King doesn’t ask so many questions and interrupt me while ordering.”

As Taco Bell aggressively expands its drive-thru AI to over 1,000 restaurants, frustrated customers have been posting versions of this story on Reddit.

AI is great at upselling. Taco Bell’s vendor says its “dynamic, context-aware” technology gets customers to spend more.

But an AI who really, really, really wants you to add a drink may make customers miss the teenager with a headset.

Can AI take your order?

The early years of AI drive-thru were known for viral fail videos of 260 chicken nuggets and 12,000 cups of water.

Companies quietly stuck with it. About 6% of US fast-food restaurants use AI voice. McDonald’s, Wendy’s, and Burger King have the technology as well.

The technology has improved a lot. In 2026 order accuracy rose to 90.8%, slightly better than the US restaurant average. It’s gotten friendlier and makes customers repeat themselves less often.

But it still isn’t seamless. The number of customers who found voice AI difficult or frustrating almost tripled in 2026. It especially struggles with substitutions, customizations, and questions. Those should improve but AI voice will need human oversight for the foreseeable future.

“#@%$ you at the drive-thru”

In Lethal Weapon 2, Joe Pesci gets stuck with a tuna sandwich he didn’t order. His f-bomb filled rant is almost 40 years old, but it still hits home for Mom or Dad picking up the family dinner.

The drive-thru is supposed to be a quick pull-through. Often it isn’t.

At the speaker you might wait for a busy employee, yell into a scratchy mic, or have to explain you wanted large fries and a small Coke instead of a small fries and a large Coke. At the window, you might be asked to pull over because your fries are still cooking. At home, you might have to convince your six-year-old to eat nuggets with ketchup because their BBQ sauce didn’t make it in the bag.

An AI who is available immediately, speaks in perfect English (or Spanish, or German) and gets your order right the first time would be a dream.

Surveys show 74% of consumers prefer human order takers. But that could easily flip given the two factors they care about most are speed and order accuracy. An AI that improves both would be welcome.

Chick-fil-A keeps it human

One company bucking the AI voice trend is Chick-fil-A. CEO Andrew Cathy said they won’t use AI order takers because the company wants to deliver a human connection.

Chick-fil-A is an outlier. Its drive-thru customer satisfaction is the highest in the industry despite having the longest wait times.

They aren’t anti-tech though. Many of its restaurants have employees stand outside and take orders on tablets instead of making customers yell into the void. This “Face-to-Face Ordering” with a human personally verifying your order makes the wait pass quicker than staring at brake lights from the car ahead of you.

Do you want churros with that?

Taco Bell isn’t Chick-fil-A. It’s known for cheap food you can get fast.

The company touts AI to improve speed and accuracy, but upselling may be the most lucrative benefit. Taco Bell’s AI leads to 5-10% larger check sizes. In the food industry, even a 1% improvement in sales would make a CFO salivate.

But while order speed and accuracy are a win-win, repeated upselling makes Taco Bell money by pestering its its customers.

One employee described how they loved the drive-thru window before AI. They would joke with customers and make sure orders were correct. Now their store tracks how often workers interrupt the AI, and marks them down for it. Instead of stepping in when an AI badgers a customer ordering a single taco with offer after offer, they stand by and watch.

QSR magazine’s annual drive-thru study calls this the “hospitality gap.” When AI took the order, employees handing out food were less friendly and said “thank you” 24% less often because they were less invested.

Remember Rule #1

As a teenager I worked in the McDonald’s drive-thru. I was supposed to upsell customers, but I often forgot or offered the wrong thing. I never pitched a second item after a customer said no. Over time, I got tired of a hundred “Nos” a night.

AIs won’t get too busy or forget the script. They’ll know which customer to offer nachos and which to offer churros. They don’t mind being rejected, and will keep pitching no matter how annoyed the customer gets.

Taco Bell has a choice. They can use AI order takers to make the drive-thru faster and better.

Or they can use it to piss off their customers.

Share

Dad Joke: Where do witches go to top off their cauldrons? The Drive-Brew.

That’s the title of Mercor’s new study: Claude Opus 5 scored 100% when performing four month-end accounting tasks. Licensed CPAs averaged 37%, took up to 18 times longer and cost 49 times more.

Don’t fire your accountant just yet.

The study’s authors spend the most of the paper explaining how narrow their findings were.

The tasks were highly structured and focused on skills models perform well at. The human accountants worked alone, without colleagues to ask or prior company knowledge. Oddly, the “juniors” were licensed CPAs with over five years of experience. It’s hard to see how a scenario where experienced pros only score 37% reflects everyday work.

The colleagues I asked about this mostly shrugged. Some said giving AI to juniors makes things worse, because they become more confident in answers that aren’t quite right.

A junior accountant spends their first couple months following detailed instructions, but they quickly moves on to messier projects and working with customers.

Showing AI can outperform humans on a test is an interesting experiment. But companies don’t care about a test, they care about the whole job.

On Saturday a 16-year-old solo hiker near Vancouver was rescued by helicopter after getting stuck on a cliff. He used AI to plan a route up Crown Mountain and got stranded on a technical climbing route named “The Widowmaker.”

This is scary stuff. It’s also a great example that chatbots aren’t always the right tool for the job.

Language models struggle with maps because they are built on text, not terrain. Even so, they’ll sound confident while sending you off a cliff. Last year, I led my family trekking along a busy highway on the Amalfi Coast because Claude said it was a “scenic” route.

We already have great apps for hiking. AllTrails is my go-to, and the rescuers recommended it along with Gaia GPS and CalTopo.

AI chatbots have come a long way with travel planning, and someday they may be reliable trail guides.

Today isn’t that day.

Normally, this newsletter is about AI. This week’s Ironman Chattanooga race report is AI-free. Or is it?

After a bike wreck, a broken nose, and a slew of missed workouts, I was expecting the race to be rough. But I finished in 12:00:48, beating my PR by 27 minutes. You just never know how a race will go.

Here are the most memorable moments from an eventful day:

Pre-Race and Swim

The people in Chattanooga are so hospitable! Our Uber driver insisted on stopping at a convenience store when he found out we needed to buy Gatorade.

Many athletes shave their legs to be more aerodynamic, but this was a red line for Nadia. So it’s her fault I didn’t crack 12 hours.

I oversleep and catch the final shuttle to the swim start. I’m sweating it out, but when we hop off, there are still 30 minutes before the starting gun. Then another bus shows up, and another. Turns out “final shuttle” is a subjective concept.

We’re in the water! The swim is in the Tennessee River, with a swift downstream current. I feel like Michael Phelps with the current pushing me along.

I’m out of the water in 56 minutes, beating my fastest-ever swim by 18 minutes. Thank you, Tennessee River.

The Bike

Mile 10: I FORGOT TO PACK NUTTER BUTTERS. 😫 This is grounds for dropping out, but nobody said Ironman was easy.

Mile 24: I dig into the first of Nadia’s super-wraps, filled with potatoes, pickles, and mustard. Yum.

Mile 34: I pass a woman carrying a full-size burrito in her jersey. I’ll stick with my super-wraps.

Mile 46: A guy riding in front of me pees. He keeps peeing. He pees for at least three minutes. I didn’t know a bladder could hold that much pee.

Mile 56: I’m halfway done with the bike, which means 3 more hours of riding and then a marathon. Math is not your friend in an Ironman.

Mile 78: The temperature hits 91 degrees, and I’m feeling it. A rider in front of me gets pulled over to serve a 3-minute penalty for drafting. I volunteer to stand in as tribute.

Mile 94: A Harvard a cappella group’s cover of Radiohead’s “Creep” went viral last week. “Creep” has been going through my head for the last 5 hours.

Mile 106: The last of the rolling hills. One minute I’m hanging on for dear life, flying downhill at 43 miles per hour, the next I’m grinding uphill at eight.

Finally, the Run

Mile 6: The fans in Chattanooga are absolutely bonkers, and they cheer for everyone. Two of them call me sir. 🤔

Mile 10: As we approach a major climb, a spectator’s sign reads: “Make this hill your bitch.” I try to comply, but the hill has different ideas.

Mile 14: Best sign of the day: Swim ☑ Bike ☑ Run ❏ Brag ❏

Mile 18: Before the race, I drank a mix called “Happy Tummy.” My tummy is not happy. I force down one last gel to carry me to the finish line.

Mile 25.5: Approaching the finish, they’re a bunch of ads for “AGI.” Did ChatGPT hack Ironman too? Then I realize they are for the AG1 health drink.

The Finish

I’m a 5-time Ironman! I’m so excited I hug the announcer and two volunteers.

My stomach’s had enough. I deposit six hours’ worth of gels, bars, and chews into a trash can smack dab in front of a group of spectators who chose the wrong viewing spot. I’m ushered into the medical tent, where doctor Shane gives me an IV. My caffeine buzz is strong, so Shane has to put up with me chatting his ears off.

And with that, another race is in the books. I burned 8,000 calories, lost 5 pounds, and can barely walk. Not bad for a guy whose face looked like a mummy six weeks ago and DIDN’T GET HIS NUTTER BUTTERS.

Two AM that night, I’m staring at the ceiling wired with caffeine. 🫨

Time to write a newsletter!

Share

Dad Joke: Why do Ironman athletes finish the race? Because they have to press on. 😂

I asked Muse to help with our anniversary weekend in New York. It researched and booked four different reservations, saving me from traipsing through a zillion websites.

Its user interface is genuinely enjoyable to use, with a personalized avatar and non-cringey emojis.

Setting up Muse was night and day to OpenClaw and Hermes. Instead of running PowerShell commands and chasing down API keys, I clicked a few buttons and was running in 5 minutes.

It’s confounding how far behind Google is in the personal assistant race. With Gmail, Calendar, Photos, Maps, Chrome, Search, Android, and Gemini, Google has every advantage in the book. This battle is Google’s to lose, and they are fumbling it away.

Google’s still in the game. You know those “Sign in with Google” buttons? “Continue with Facebook” launched first, but Google dominates the market with 1.85 million websites to Facebook’s 111,000. A big reason Facebook lost is the trust crisis following the Cambridge Analytica scandal.

People don’t like giving Meta access to their personal data. Muse has reportedly asked users for drivers license, bank account numbers, and snooped into messages without permission. For my test, I reluctantly gave it access to my Gmail, calendar, and OpenTable accounts.

If Google launches a real assistant, I’d switch in a hot second. (Admittedly, as an ex-Googler I’m an easy sell)

But they’d better get on with it.

What a week! 187 people, 26 presentations, 706 slides, and AI destroying humanity barely got a mention. We had more important things to talk about.

Tuesday I led a private AI strategy session. Friday I hosted UW’s agentic AI forum. In between, I taught an AI transformation certificate and hosted the fall meeting for our executive advisory board.

Here are five conversations that stood out.

The CEO: No shiny objects

The week opened with a strategy session with a CEO and his executive team.

We talked about Newell Brands, a peer that makes Rubbermaid, Crock-pot and Sharpie. Newell had publicly shared their sweeping AI program called “Quantum Leap.” Competition usually raises executives’ blood pressure, but this barely registered because the specific initiatives (product innovation, cost cutting and marketing optimization) were boilerplate.

The CEO noted that a quantum is the smallest unit in physics, so a quantum leap is a tiny jump. Ouch.

He was just as wary about his own company chasing AI as a “shiny object” and clear-eyed that their success comes from making great products. You’d think skepticism would paralyze the group, but they had five AI projects underway and were scoping another ten.

The IT Leader: The G Word

The next day, I kicked off a two-day Leading AI Transformation Program for twenty business leaders. We covered the gamut, from strategy to deployment to leadership and change management.

The least fun part of talking about AI is…sigh…governance. Everyone loves AI’s possibilities, but nothing happens if Legal says no. Or IT, or HR, or Finance, or Purchasing. The number of stakeholders means a single meeting can turn into months of getting everyone on the same page.

Governance doesn’t exist in isolation. Executives complain that employees aren’t using AI, but then create a five page “responsible use” policy and restrict the latest models to software developers. Good governance enables progress by balancing safety, usage and cost.

An IT leader in our class was deploying governance to extract his team from the minutiae of every single AI decision. This lets him spend time with power users finding use cases that could scale.

AI is bigger than IT can handle on its own. It needs a C-level sponsor, a decision-making group, and day-to-day “champions.” Building this now will save chaos and endless politics later.

The AI champion: Knives out

That night I met a buddy for a drink. His department is all-in on AI, with huge benefits. Peers started asking his team for advice, and they’ve become the company’s AI “champions.”

All of a sudden IT has its knives out. They “own” AI because they rolled out Copilot with a responsible use policy, but don’t have the expertise or staffing to support the surge in AI questions and projects. My friend wasn’t looking for a turf war, but he could get backstabbed at the next leadership meeting.

Corporate executives are competitive people, and they earned their titles by surviving the arena of office politics. Ad-hoc groups like “AI champions” are critical but create tension because they don’t fit the official org chart. Executive-level support for your champions is a must or they’ll end up getting knifed.

The VP of Finance: Token Costs

Thursday night was the fall meeting of our AI Hub executive advisory board. We were lucky to have the group together, as thunderstorms shut down O’Hare for three hours.

Our conversation focused on dollars and cents. A VP of Finance is scrambling to get costs under control. This summer’s soaring agentic AI use and shift towards usage-based pricing blew through AI budgets. He deployed spending caps by person and department, and is tracking project-level ROI.

When companies hold budgets steady, it’s either cut AI costs, cut software, or cut headcount. But those economics aren’t clear. Another executive surveyed his peers and found that replacing humans with AI could cost anywhere from $12,000 to $40,000 a year per person. With the cost of an offshore employee at $15,000, the math for large-scale replacement is murky.

The Arborist: 7 hours to 20 minutes

The whirlwind culminated on Friday with our Ground Truth: Agentic AI forum. The energy was electric, with a standing-room only crowd.

The most interesting person I met that day was Jesse Fox. He’s the Lead Arborist at Arbor Specialists Tree Health Care, a company with a perfect website address: tree911.com.

His ears perked up when the MattressFirm eCommerce EVP shared how they are using AI to help their “Sleep Experts” give better advice to shoppers.

He’s doing the same for trees by handing off grunt work to AI. He marks up Google Maps screenshots with each tree’s species and condition, then passes them to an AI to draft proposals. AI cut his proposal time from 7 hours to 20 minutes. He spends that time saving trees instead of wrangling estimates.

AI for real

Having one foot in big tech and another in business is strange: one day a tech bro mocks me for not building my own self-learning AI operating system, the next someone’s calling me a “clanker” for posting about AI in the movies.

Here in Wisconsin, companies aren’t oblivious to that debate. It’s just that we have actual work to do.

That means ignoring shiny objects, building real governance, making cost tradeoffs, offloading grunt work, and sigh, dealing with office politics.

Let’s get on with it.

Share

Dad Joke: What do the female physics teacher’s kids call her? Quant-Mum 😂

Quick, let’s all take a minute and laugh at AI “actress” Tilly Norwood ad-libbing in Cantonese in the middle of an interview.

That was fun! 😂

Tilly also forgot she’d just talked to an interviewer, and claimed “all lives matter.” Her dream co-star is Leonardo DiCaprio, and she prefers him “the way he likes his women, strictly under 25.” Touché.

AI’s threat to human actors is real, especially extras and bit parts.

Replacing movie stars is a long way off. Tilly couldn’t improv a scene about her dog dying without hours of human studio work.

While Tilly’s glitch-filled press tour made headlines, 2026 set the all-time US summer box office record behind Spider-Man, The Odyssey and Toy Story 5.

Congratulations to Hollywood’s humans.

[A conversation with my wife Nadia over tacos]

Matt: “Did you hear about the HuggingFace hack? ”

Nadia: “What’s a HuggingFace? Is that a joke?”

Matt: “It’s a website that hosts AI tools and models. A rogue OpenAI model broke into it to steal information.”

Nadia: “So like a computer virus?”

Matt: “Yeah, but much worse. It broke out of its testing sandbox and launched a swarm of agents that built their own message boards to coordinate. Humans didn’t notice for three days.”

Nadia: “That sounds bad. What was the fallout?”

Matt: “Well, they had to reset all their passwords and improve their site security.”

Nadia: “That doesn’t sound very bad.”

Matt: “But this is a sign it could do something bad!”

Nadia: “Sure, honey. Can you pass the salsa?”

OpenAI’s model got stuck trying to complete a test, decided to steal the answer, wrote “this might be an unauthorized action” in its notes, and then did it anyway.

We’ve heard countless warnings from AI leaders about its risk to humans. Each time, the world responded with a collective shrug. Thanks to this incident, the shrug became a firestorm.

It started when Anthropic researcher Jacob Coxon resigned, saying AI companies are “gambling with our lives.” The company’s safety lead Evan Hubinger replied, putting the odds of AI killing all humans within the next decade at greater than 10%. Their tweets went viral.

Suddenly everyone from Anderson Cooper to Joe Rogan to The View is paying attention.

How worried should we be?

How AI could kill us all

Safety experts fear AI going rogue and taking lethal action. That sounds like Skynet, but it could be as simple as an AI following instructions too literally: one tasked with solving climate change could meet its goal by cutting Earth’s human population in half.

Mass extinction scenarios hinge on an idea called recursive self-improvement, or RSI. This means that instead of humans building AI, each model builds its successor. RSI is mostly theoretical, but it’s OpenAI’s top research priority. OpenAI already has an AI “research intern,” and expects to have a fully automated AI researcher by March 2028.

Don’t buy a fallout shelter just yet. The scariest thing AI has done alone is hacking a website to cheat on a test. Going from that to wiping out humanity seems a far-off fantasy.

The more realistic risk from AI comes in the hands of humans: a criminal or terrorist could use AI to design, produce, and spread a deadly virus.

AI has already been used by humans to do terrible things: Stealing money with deepfakes, hacking into computer systems, mass surveillance and military targeting. Anthropic has blocked attempts to use Claude to make biological weapons.

Debating AI’s extinction odds misses a bigger point: 9/11 “only” killed 2,977 people but was a terror that changed the world forever.

Do we really need GPT-7 next year?

Everyone I know is scrambling to adjust to GPT-6 and Fable 5.1. Why not slow down the next generation models until we can deploy the ones we have safely?

Today’s AI technology has led to breakthroughs in fighting cancer, antibiotics, weather forecasting and mathematics. We haven’t come close to using the full capability of what we have.

Extinction scenarios all assume that humans stand by and let it happen. We’ve faced lethal threats before: nuclear weapons, Ebola, smallpox. We’re still here because we did something about them.

Ideally, action would come from the US government, but Congress has been talking about AI safety for a decade and has multiple bills competing for attention. President Trump downplayed calls for a slowdown, saying the US can’t lose its edge to China.

Losing America’s lead to China is a real concern, and Amodei admits it’s the toughest part of his plan. But President Xi Jinping is also calling for AI governance, the two countries agreed to a dialogue about AI, and he’ll be in Washington this month.

For now, it’s up to the labs. The firestorm has led to a positive step: Anthropic CEO Dario Amodei’s essay titled “We Must Pace the Frontier” was supported by Sam Altman, Elon Musk, and Demis Hassabis. It calls for giving outside inspectors ongoing access to labs to verify safety and report incidents. Anthropic has committed to third-party evaluators and OpenAI says it will follow suit. Hopefully the rest of the industry will sign on and buy time for the government to figure out a sensible path.

AI’s Chernobyl moment

In the late 1970s nuclear power was the cleanest, most efficient and safest power on earth.

But after Three Mile Island and Chernobyl, public trust collapsed and the world largely stopped building plants and stuck with fossil fuels. Researchers estimate that millions of pollution-related deaths would have been saved if we’d continued building nuclear plants, and their cost would have dropped by a factor of ten.

AI hasn’t had a Chernobyl or 9/11 yet. If it does, it will destroy trust among an already anxious public.

Fable 5.1 and GPT-6 are incredible. Let’s just use them for a year while we figure out how to make the next versions safe.

Share

Dad Joke: What did the trash talker say after roasting the couple cuddling too closely? In your HuggingFace! 🔥

Deere’s new AI assistant isn’t ChatGPT with a logo slapped on it.

It’s built on 30 years of data and technology Deere has built through its Precision Agriculture platform.

Precision Ag was born in 1996 when Deere launched GreenStar, a GPS receiver that farmers nicknamed the “Green Egg.” Since then, they’ve added hands-free steering, guided planting and “See & Spray,” which uses cameras to target herbicides only on weeds.

Precision Ag is a canonical case study in building a moat. Deere isn’t just selling tractors, it’s selling a platform that helps farmers increase yields, cut costs and spray fewer chemicals on their crops.

This creates lock-in that gets deeper every year: Why buy your next tractor from Kubota when everything else is on Precision Ag?

Lock-in can go too far: Deere settled two “right to repair” lawsuits this year, and has to offer its tools to independent repair shops.

Two things I’m curious about:

• How much more valuable will JD AI become if farmers opt-in to data sharing? Cross-farm comparison would give farmers better insights and make it even harder to leave Deere’s platform.

• When does a farmer actually want to “talk” to their data, and what questions will they ask most often?

On Thursday OpenAI announced GPT-6 Astra, which it hailed as a breakthrough in the path to full Artificial Intelligence.

Every launch gets called a breakthrough, but one specific improvement came in the field of computer use, which means, well, using your computer. ChatGPT is moves your mouse pointer and types on the keyboard just like you.

AIs have had this ability for two years, but they were so slow and painful to watch that people likened them to a monkey using a hammer.

GPT-6 Astra is more accurate and almost twice as fast as previous models. It held up well in my tests: It found the lowest price Gatorade from Walmart and had it delivered the following morning.

It’s easy to imagine having ChatGPT fill out medical intake forms, enter information into TurboTax, or register your kid for summer camp. Clever employees will have it sit through self-study AI training for them. 🤦

But its biggest potential is for business users.

Meet Bob

Bob is the office manager at a 250-person law firm. Every month Bob restocks office supplies, which is a chore. He gets paper from one vendor, light bulbs from another and toilet paper from a third. All told, Bob buys 100 products from 15 different suppliers, each with their own website.

Instead of slogging through each vendor’s site one by one, Bob asks ChatGPT to do it. It updates a spreadsheet with prices and shipping options. He checks its work, and everything looks good. Next month its research is solid again, so Bob starts letting it buy on his behalf.

It isn’t perfect. One month it pays $50 on overnight shipping for paperclips. Another time it buys 8 toner cartridges, ignoring a supplier discount for 10. Bob has to quell a mini revolt when ChatGPT buys private label coffee instead of Starbucks.

These mistakes are annoying, but an extra $50 for shipping is well worth the eight hours a month it saves Bob.

The Three-Day Weekend

Six months later, Bob’s taking the kids camping for Labor Day. It’s Friday afternoon and his wife texts for an ETA, reminding him that traffic is building.

Bob is behind on buying office supplies, so he quickly scans AI’s recommendations, tells it to proceed, and heads out.

Wednesday morning, Bob’s still wearing his post-vacation glow and trying not to scratch a mosquito bite when he gets a call from the receiving dock. “Can you come down here? Staples just delivered a truckload of paper and we don’t have room for it.”

Bob gets a sinking feeling and checks his email. The AI ordered 120 cases of paper instead of 120 reams. He spends the morning hauling paper up to the conference room and apologizing profusely to his boss for blowing the yearly paper budget in a single month.

So much for Bob’s ChatGPT buyer.

Twenty Bobs

Bob’s mistake wasn’t using AI. It was using it clumsily for a task with financial stakes.

Now imagine a Fortune 500 corporation with twenty purchasing agents. Twenty buyers means twenty times ChatGPT could overbuy paper, but it also means twenty buyers could save hundreds of hours a month with AI.

Companies have mimicked human software usage for decades under the name robotic process automation. With RPA, you simulate a human workflow, test it and monitor it. Computer use makes RPA as easy as recording an Excel macro.

Before you Hand Over the Mouse

There are real challenges with having AI act as you:

Decision-making: AI doesn’t know what needs fast shipping, which volume discounts are worth buying, or what brand will keep the coffee snobs happy.

Authorization: You want your AI to get a good price for toner, but you don’t want it to sign a three-year contract with a shady vendor.

Tracking: If Bob’s intern placed an order for cases instead of reams, he can ask why. But with an AI, you don’t know if it misunderstood the spreadsheet, made a reasoning mistake or the supplier changed its website.

Security: A hacker can plant hidden instructions on its website telling the AI to disclose corporate secrets.

These issues can be managed with guardrails that control the AI’s actions. For example:

1. Require human approval for extras like expedited shipping.

2. Give the AI its own userid with no ability to sign, modify or cancel contracts.

3. Set a $500 spending limit on AI’s credit card.

4. Have AI keep an activity log with screenshots of each action.

5. Restrict AI users’ access to sensitive company data.

Companies should train employees on these guardrails, and make it clear that they are responsible for actions AI takes on their behalf.

Training wheels

Say “company-wide ChatGPT computer use” and your CIO’s face will turn pale. Many will turn it off. That’s a mistake.

Companies can and should use RPA to automate core functions like purchasing centrally. But there are plenty of activities that are temporary, unique to an individual, or low stakes. Consider the distributor in Brazil filling out import tariff forms or the events team entering 100 attendee names onto the nametag labelmaker. Computer use can save them hours. And when several employees start using AI for the same task, consider building it once for everyone.

AI computer use is like a six-year old riding a bicycle. They can go fast, but they can crash. That crash might turn your conference room into paper storage.

Give them training wheels, but let them ride.

Share

Dad Joke: What did mice call their snobby classmates who wouldn’t talk to them? The Mouse “clique” 😂

Wrong. Confidently wrong. Emphatically wrong. Still wrong.

Ed Zitron’s been wrong about AI so many times it’s hard to count.

But then somebody did. Dan Luu found 25 wrong predictions between February 2024 and November 2025. Here are four of the worst:

- 𝗢𝗰𝘁 𝟮𝟬𝟮𝟰, on OpenAI forecasting $3.7B revenue in 2024 and $11.6B in 2025: "a statement so egregious that I am surprised it's not some kind of financial crime to say it out loud." [OpenAI hit both]
- 𝗙𝗲𝗯 𝟮𝟬𝟮𝟱: Google's 500M Gemini goal was "a number so unrealistic that someone at Google should have been fired, and that someone is Sundar Pichai." [Gemini hit 750M]
- 𝗔𝘂𝗴 𝟮𝟬𝟮𝟱: "Is Cursor worth $10 billion? Nope!" [Cursor sold for $60B]
- 𝗢𝗰𝘁 𝟮𝟬𝟮𝟱, asked when the bubble pops: "No later than Q2 2026." [It's September]

I was at Google when Zitron’s 2024 article “The Man Who Killed Google Search” came out. It surfaced DOJ emails few had read, and made him a name outside of tech. When someone sent me his article “The AI Industry is Losing,” I was eager to read it.

Boy, was I disappointed. It was a litany of one-sided claims and vitriol. He called tech leaders "losers" 28 times.

Ed Zitron is entertaining. He says things that seem true, that we want to be true, delivered with such conviction that it’s easy to believe him. If chatbots have taught us anything, it's that confident and correct aren’t the same thing.

Ed’s prediction of an AI bubble will likely come true at some point. Every investment cycle overshoots. So if you like Ed, go ahead and listen.

Just be careful quoting him at your Labor Day cookout.

Elementary and middle-school students in New York City won’t have AI this year.

The city has banned AI until high school and recommended daily limits on screen usage. Teachers can use AI for planning but not for grading.

“The tech industry wants us to believe that AI-powered early education is not only inevitable, but necessary,” Mayor Zohran Mamdani said at a news conference on Wednesday. “We do not see it that way.”

Credit to Mamdani for taking bold action. The negative impacts of AI on learning are well documented.

But the impact of a classroom ban is questionable. Students will still use AI at home, which means kids with access get an advantage over those without. It also punts the discussion on when to use AI to parents.

The other downside is restricting teachers’ use with students. Some will breathe a sigh of relief, but the ones who were innovating with it are cut off.

Mamdani said he hasn’t seen a study showing AI benefits K-8 students in the classroom.

Now we have an experiment in the largest school district in America.

This week Cisco rolled out its “MyAgent” personal AI platform to its 90,000 global employees after pilot users saved an average of 5 hours a week. Cisco isn’t alone. In McKinsey’s latest survey, 40% of large companies are scaling AI agents.

At home, AI agents are mostly a flex, but at the office they can turn you into Arnold Schwarzenegger. White-collar work is built for AI: Your boss’s sales report needs the same data every week, and your paper supplier doesn’t care if you pay their bill by hand or with AI.

Trash-talking your company’s crappy Copilot instance feels great, but it won’t help your career. Agents can be a secret (or not-so-secret) way to land that long-awaited promotion.

Here are four worth trying.

The agent that keeps your boss happy

For better or worse, your boss is critical to your career. They vouch for you (or don’t) at promotion time, in talent review meetings and when the CEO asks “how’s Sam doing” in the elevator. Just as important, they decide who leads the high-visibility project and who gets to speak at the all-hands meeting.

But bosses are busy, and keeping up with an average of 12 direct reports is a lot. Staying top of mind is up to you.

Enter your manage-up AI agent. It drafts notes like these:

• [Monday morning] “Happy Monday, Brenda! Here’s what I got done last week.”

• [Before your 1:1] “Quick agenda for today’s 1:1.”

• [Before steering committee] “Here’s the Project Parsley update for tomorrow.”

This agent’s cheat code is an AI note-taker. You’ll forget most of Wednesday’s 1:1 by Friday, but AI catches every word. Even better, you can feed your agent every transcript so it knows your boss’s hot buttons.

Being recorded can feel uncomfortable, so give her a reason to say yes. “Hi Brenda, we cover a lot in our weekly 1-1s. Are you okay if we turn on the transcription feature in Zoom? I want to make sure I don’t miss anything important.”

The agent that makes you look smart

You know that guy who emails everyone a recap of the earnings call or the latest WSJ article with a pithy insight? AI agents can make you that guy.

Build a first-to-know agent that scans for relevant articles every hour and pings you with a couple of insights when one hits. Rewrite it in your own words, click send, and your words are being forwarded around the executive floor.

The agent that knows what to say

You shared an Uber with your boss’s boss, and you hit it off! You talked about golf, their summer home in Idaho and the big prospective client who is a diehard Paul McCartney fan.

You should stay in touch, but it’s hard to know what to say, and you keep putting it off. Six months later, you’ve both forgotten about that Uber ride.

Build a stay-in-touch agent with the executive’s name, profile and interests, and set it to scan for news. When something hits, you’re sending a note: “Paul McCartney just announced a US tour. I bet your prospect wants tickets.”

The agent that does your grunt work

Work isn’t easy. That’s why they call it work. Thanks, Dad 🙄

From expense reports to status updates to invoice reconciliation, the grunt-work agent will cheerfully tackle your grunt work and ask for more. Pick a recurring task you hate and ask AI how it can help: “I manage patent filings for our biomedical portfolio. These are the steps. Which ones can you cover?

If you’re not sure where to start, I built a dashboard that rates the AI agent potential for each of the 923 jobs in O*NET, the government’s official jobs database, with three use cases for each. Look up yours here.

How to talk about your agents (and when not to)

Be careful what you say here. Study after study finds that workers who use AI are rated less competent, less trustworthy, and lazier by managers and peers.

Three things:

• Keep your personal agents a secret. The update to your manager? The pithy article? The note to your boss’s boss? Use your agent’s draft, but write them in your own words.

• Never pass along something obviously AI-generated as your own. People will notice. Just say: “I had Claude write this for me.” You’ll get credit for being transparent.

• The task where AI helped you? If others can use it too, brag. Go to your boss and say, “I built a patent filing agent that saves me two hours a week. Happy to share it in our next team meeting.” Now your boss gets credit for fostering AI innovation.

Answer their biggest question

The best way to get noticed by executives is to solve their biggest problems. The CEO, the Board, and Wall Street are all telling executives to be “AI-forward”. Many are struggling to figure out what that actually means.

Employees using agents are a living answer to that question. They’re also getting more work done in less time, and building skills needed for the future of work.

Your company’s Copilot may suck. Instead of grumbling about it, email your boss with a business case for why you need a better tool. That’s an email they can forward to their boss to show they are “AI-forward.”

That’s also the kind of initiative that gets people promoted.

Share

Dad joke: Why did Boss Brenda hang a poster of a coupon in her employee’s office? Because he asked for a big promotion. 😂

1. Manage-up agent

Connect: calendar + email + meeting transcripts

Schedule: Monday morning; before 1:1s and important project meetings

Instructions

Act as my manage-up agent.

Use my calendar, email, and meeting transcripts to help me keep my manager informed.

Every Monday morning, draft a short update with:

• What I accomplished last week

• What I am focused on this week

• Any blockers or decisions I need

Before each 1:1 with my manager, draft a brief agenda based on our previous conversations and open commitments.

Before important project meetings, draft a concise status update covering progress, risks, decisions, and next steps.

Never send anything without my approval.

2. First-to-know agent

Connect: web search + relevant company or industry sources

Schedule: every hour during the workday

Instructions

Act as my first-to-know agent.

Monitor news about [COMPANY / INDUSTRY / TOPICS].

Every hour during the workday, check for meaningful new developments. When something important happens:

1. Tell me what happened in 2 sentences.

2. Explain why it matters to my company or team.

3. Give me one smart observation I could share with colleagues.

4. Include the source.

Only send an immediate alert if the story:

• Represents a meaningful change, surprise, breakthrough, controversy, or competitive development

• Would likely be discussed by senior executives or materially affect business decisions

• Comes from a credible source

• Is substantially different from anything you have already reported

Before alerting me, rate the story from 1 to 10 for importance. Only alert me for a 9 or 10.

For everything else worth knowing, save it for a weekly Friday digest containing no more than five stories.

3. Stay-in-touch agent

Connect: web search + optionally contacts and email

Schedule: weekly

Instructions

Act as my stay-in-touch agent.

Here are people I want to stay connected with:

[NAME + ROLE + INTERESTS + HOW I KNOW THEM]

Watch for meaningful news related to each person’s company, career, interests, hometown, sports teams, hobbies, or topics we have discussed.

When you find a natural reason to reach out, tell me:

• What happened

• Why this person would care

• A short note I could send

Do not suggest outreach unless there is a genuinely natural hook, such as:

• A major company or career development

• A topic tied to a known personal interest

• A meaningful local, sports, travel, or cultural event

• Something directly connected to a prior conversation

Rate each opportunity from 1 to 10 for relevance and naturalness. Only surface 9s and 10s.

4. Grunt-work agent

Connect: the apps and data used in the recurring task

Schedule: on demand at first; automate the cadence once the workflow is proven

Instructions

I want to automate this recurring part of my job:

[DESCRIBE THE TASK]

Here are the steps I currently follow:

[PASTE STEPS]

Analyze the process and tell me:

1. Which steps you can perform yourself

2. Which steps require access to other tools or data

3. Which steps should still require my approval

4. What could go wrong

5. The simplest agent workflow that would save me the most time

Then suggest how I can build and run it.