The Turn: Oops, All AI Tokens
A tech company blew their AI budget and HR takes a hit
When I heard that Uber had blown through their whole yearly AI budget by April because they promoted tokenmaxxing (like many of their peers), I thought what I usually think when tech companies do something predictably dumb: Why is anybody surprised?
About a month later, the rideshare company let go of nearly a quarter of their HR team.
Okay, that was a little surprising.
Officially, Uber has denied any connection between these two events. But I have my doubts so I wrote about that in my Reworked column this week.
According to Uber leadership, HR was, “too complex and fragmented, with overlapping responsibilities, unclear ownership and teams operating too far from the businesses and partners they support.”
And according to reporting from HR Executive, of those remaining in the HR department, anyone still on a remote work arrangement is having it rescinded.
Here’s a guess that HR won’t be done with losing people.
Like so many tech companies, Uber is not struggling financially. Its first quarter gross bookings were up 25% and its YoY revenue is up nearly 15%. Of course, the stock market continues to not care about that — its share price is down more than 13% for the year as of this writing.
So they did what a lot of companies do: a very loud layoff of corporate’s favorite scape goat with the hope to appease the shareholder class.
Oh, we’re back to not being surprised again.
I am not a huge talker but I have a whole rant about the types of companies that throw HR under the bus like this. Like any HR department ever gets the benefit of the doubt when it comes to increased staffing in their department. Do I believe Uber was overstaffed by 340 HR people? No, I do not.
Do you know how many HR leaders I have talked to who have had to write multi-page business cases for the most ridiculous things? A six-month contract recruiter to cover for maternity leave? A single HR analytics employee when the business has asked for dozens of detailed reports from ancient enterprise systems?
Maybe HR did need a re-org. But the timing, the messaging, and the impact don’t add up for me. HR departments getting cut this significantly are for flailing businesses or mergers, not for a some complexities and overlap in a department.
In my column, my hope is that HR increases their presence in discussions about AI usage across the organization (not just for HR). I keyed in on three discussions HR deserves to be a part of. Take a read if you’re interested.
What else is happening this week?
What Olivia Rodrigo and Robert Smith Can Teach Us About Enterprise Software Today. A Thomas Otter post with a music observation built in is two of my favorite things at once. He uses the Rodrigo-Smith Glastonbury collab to argue that enterprise incumbents can either coast on back-catalog nostalgia or take real risks and earn new audiences.
Qualified, Interested, and Affordable Is No Longer a Valid Method for Employers to Measure Applicants. Time to QUAIL. Andrew Gadomski adds two letters to the old QIA framework, Understandable and Legitimate, to account for a world where fake candidates, fake recruiters, and fake job postings mean neither side of hiring trusts the other anymore.
Crash Out With Me: A Letter to Corporate. Madison Butler published this anonymously because the pattern matters more than the details, and any Black woman who has tried to earn safety inside systems not designed to provide it will recognize it immediately.
You Don’t Have to Say Yes to Money. Laurie Ruettimann on the Peter Thiel Dialog list and the prominent leadership authors on it: people who make a career preaching integrity should know who is signing the check.
Pro-Worker AI, Explained. MIT economists cover the difference between AI that replaces labor and AI that expands what workers can do, and why most companies are defaulting to the easier path.
The Work AI Can’t Do. Moe Carrick on connective labor: the unscheduled conversation, the noticed silence, and the manager who actually knows her people.
What Jalen Brunson Can Teach Us About Team Building. David Manaster at ERE is with me on this: Brunson left money on the table so the Knicks could build a real roster around him, and they won the chip. You have to do a lot of things right, and it all starts with people.
Is DEI Dead? Not According to New Catalyst Data on Workplace Inclusion. New data shows 80% of organizations remain committed to DEI, and that 55% have publicly signaled a retreat while only 34% have actually reduced their inclusion efforts. Sorry not sorry: still kicking.
Toxic Workplaces Are Winning—For Now. New Businessolver data shows that the worst corporate cultures are 2x more likely to report significant financial growth. The pendulum will swing the other way… eventually.
Ragan’s 2026 HR Tech Hot List Honorees Announced. Twenty companies on the list this year, some new names alongside the familiar ones like UKG, Rippling, Lattice, and Eightfold.
How Remote Work Has Helped a Generation of Working Parents. Mothers of children under 5 have been in the labor force at higher rates since 2023 than before the pandemic, and some say they wouldn’t have had kids at all without remote flexibility.
Growth of Data Centers and AI Expose Tensions in Organized Labor. Building trades want the construction jobs data centers bring; other unions are worried about AI taking theirs. The reality of diverging economic incentives can be brutal for working folks.
Newsom Promised to Help Californians Build New Careers. Now, the Money Is Running Out. California’s high road training partnerships are getting little or no new funding in the 2026-27 budget. Promising a workforce agenda and not funding it is just a press release.
Fake Candidate? Recruiting Impostor? Check Here. Stacy Zapar built a free, searchable database of suspicious emails, profiles, and domains used by scammers: 1,200+ flagged addresses and 170+ recruiter impersonators. Bookmark it.
Now of Work Weekly Digital Meetup with Jason Lauritsen. If you haven’t tuned into Jason Averbook and Jess Von Bank‘s weekly show, catch up on this one with Jason Lauritsen and sign up to join on Fridays.
When the Prediction Becomes the Verdict. Part two of Jess Von Bank‘s seven-part series: the Kentucky pretrial risk assessment tool deployed to reduce racial bias ended up giving racism a more defensible interface. Your AI governance team should read this one.
Another Word for Technical Debt Is Maintenance. John Sumser on agents: the harness is 90% of the model, ages on a different clock than the LLM, and when the model improves, the guardrails built around its old blind spots quietly stop working. “Agent reliability engineer” might be the next job title nobody saw coming.
Make your writing more interesting
Last week, I shared a short guide about how to fix your brand writing. Of course, I had to promote myself at the end, but even if you don’t want to work with me, there are a few takeaways to make your writing stand out.
I actually don’t blame AI for all of the terrible writing you see from B2B brands. This has been happening for years before AI could write an em dash. Whether you use AI or not, there are some pretty simple fixes, though.
Swipe through the guide here, typo and all because I decided to push publish instead of sitting on it for days.
Have a great rest of your week!