Musical accompaniment for this week's newsletter.
Scott Galloway says the AI job apocalypse narrative is “mostly BS.” A16z's David George published a piece this week calling it “a complete fantasy.” Both are making the lump-of-labor fallacy argument: there's no fixed amount of work, automation always creates more jobs than it destroys, and history says relax.
They're not wrong. They're also not neutral.
Galloway is a professor who writes for an audience of reasonably comfortable people with strong opinions about tech. A16z is a venture capital firm with billions invested in the AI companies that benefit most from a world where nobody panics about job displacement. When the entity that funds the disruption publishes reassuring essays about how disruption is fine, actually — that's not analysis. That's asset protection with a Substack account.
To be fair to the argument itself: the data is genuinely mixed. Coder job listings are up 11% year-over-year. Radiologist listings are up. The U.S. labor market, at a macro level, is not in freefall. History really does show that technological waves create more work than they kill — eventually.
But “eventually” is doing a lot of heavy lifting in that sentence.
Meanwhile, Freshworks laid off 11% of its workforce last week and told Reuters, out loud, that more than half of its code is now written by AI. The WSJ frames the current moment as a binary CEO decision: cut headcount or extract more output from the people who remain. Both of those things happened in the same week as the “it's all a fantasy” essay. That's not a coincidence. It’s where we are right now.
The volume on both sides of this argument has gotten louder than the signal inside it. In case you haven’t noticed, things are a little amped right now. Someone threw a Molotov cocktail at Sam Altman's house last month. Gen Z's excitement about AI dropped 14 points in a single year. There's now a serious legislative proposal to tax compute to fund worker displacement programs.
Meanwhile, the actual reality of AI in most workplaces is profoundly, almost disappointingly, mundane. It's a slightly faster first draft. It's an expense report that takes four minutes instead of fourteen. It's a recruiter screening tool that works okay except when it doesn't. AI at work is not the second coming of electricity yet. It's a useful, uneven, frequently overhyped tool that most people are still figuring out how to use without embarrassing themselves in front of their boss or accidently deleting an entire database in the CRM.
The Baroque visions of AI disruption — the gleaming robot armies, the permanent underclass, the overnight abundance — are being written by people who spend very little time in actual workplaces, talking to actual workers, or watching actual software get adopted at actual companies. Which, not coincidentally, is the exact audience both sides of this debate are writing for.
For anyone in Work Tech right now, I think you need to read the room. Your buyers are living in the middle of this argument, not above it. The HR leader at a mid-market company just watched colleagues get replaced by Freshworks-style efficiency math. She's also getting pitched “AI-powered agentic workforce blahty blah” by three vendors this week. Everyone is anxious and most of what you are telling them doesn’t make a lick of sense to them.
Get your story straight. Anyone who can talk about what's actually working and achievable will have a very different Q3 than those painting grand pictures of visionary AI maximalism.
The apocalypse might be a fantasy, but the anxiety is real.
What else is going on this week?
QuickBooks Workforce dropped this week — an all-in-one, AI-native HCM platform covering payroll, recruiting, onboarding, benefits, compliance, and performance management. In one launch. For small and mid-market businesses. Powered by agentic AI workflows and a conversational chat interface.
Here's the number that matters: QuickBooks Payroll already serves 18 million U.S. workers. Intuit didn't enter the HCM market this week. They revealed they were already in it — they just hadn't told anyone yet.
ADP, Gusto, Rippling, and every SMB-focused HR platform should be looking at this the way the hotel industry looked at Airbnb in 2011. Not panicking yet, but definitely not sleeping great either.
The GTM implication is blunt: if your positioning depends on owning the SMB HR workflow, you just got a very large, very well-capitalized new competitor who already has your customer's financial data. That's not a feature gap. That's a moat problem.
Applications per recruiter on Greenhouse's platform are up 412% since 2023. Not a typo. Four hundred and twelve percent. And 74% of those candidates used AI to apply. So Greenhouse's response was to acquire Ezra AI Labs — a voice AI interviewer built to give every applicant an actual structured conversation at the top of the funnel, instead of a resume black hole.
Greenhouse is betting the exit ramp is better AI — transparent, structured, voice-based — not less AI. That's a contested thesis, and it's worth watching. Especially paired with its MCP launch, which gives recruiters a governed API layer to connect any AI tool to Greenhouse's data. They're not just buying a product. They're building a platform play around the hiring signal crisis. Smart move. The timing is good.
SAP acquired Dremio (agentic data lakehouse) and Prior Labs (tabular foundation models) in the same week, committing over $1.16 billion to Prior Labs alone over four years. Both deals target the same bottleneck. SAP's CTO said it out loud: “Enterprise AI doesn't stall because the models aren't good enough. It stalls because the data isn't ready for AI agents.”
That's a confession, not a press release.
SAP just told the market that the enterprise AI problem isn't intelligence — it's infrastructure. Fragmented data, proprietary formats, no business context. Dremio fixes the access layer. Prior Labs fixes the reasoning layer for structured tabular data, which is where most enterprise decisions actually live.
The move that should get Workday's attention: SAP is explicitly building for non-SAP data. They're not trying to own your data estate. They're trying to be indispensable across it. That's a fundamentally different — and frankly more honest — platform strategy than most of their competitors are running right now.
In the latest episode of Work Tech Weekly, I sat down with Matt Poepsel, VP and Godfather of Talent Optimization at Predictive Index. He dropped this on me: The average AI deployment puts 80 to 90% of its energy into technology. Maybe 10 to 20% into people.
That number explains almost everything going wrong with AI at work right now. Every business problem is a people problem. So why are most organizations still treating AI like a technology problem? Good stuff. Give it a listen.
Deel acquires Sastrify, an AI-powered SaaS spend management platform. Deel now wants to own not just how you pay global workers, but how you buy the software they use — a quiet but real expansion of its platform ambitions beyond payroll into the broader enterprise spending layer. (LinkedIn)
Anthropic nears a $1.5 billion joint venture with Wall Street firms. The company that makes the AI everyone wants to tax is now co-investing with the firms everyone wants to tax to fund it — a remarkably tidy illustration of whose interests are actually aligned in the AI economy. (Wall Street Journal)
Corgi raises $160 million in Series B funding at a $1.3B valuation. An AI-native insurance carrier for startups that doubled its valuation in four months — the kind of velocity that makes you check the footnotes twice, except the numbers hold up. (FinSMEs)
Personio achieves profitability and acquires Aurio, a Munich-based recruiting AI startup. Europe's most prominent SMB HR platform hitting its first profitable quarter is genuinely meaningful, even if Chad and Cheese are right that burying it in the same press release as the acquisition is "sloppy marketing." (Personio)
Tessera Labs raises $60 million in funding. The AI infrastructure bet continues; Tessera focuses on enterprise AI deployment tooling, adding to a funding stack that increasingly favors picks-and-shovels plays over finished applications. (FinSMEs)
SageOx raises $15 million in seed funding. Early-stage capital is still moving in AI-adjacent enterprise tooling, even as later-stage rounds consolidate around fewer, larger bets. (FinSMEs)
ServiceNow CEO says $30 billion in subscription revenue by 2030 is the bear case. Bill McDermott bought $3 million of company stock to back it up, which is either a genuine conviction or the most expensive PR stunt in enterprise software history. Either way, the $27.7 billion in remaining performance obligations suggests he's not making it up. (Seeking Alpha)
ICIMS names Marc Thompson Chief Executive Officer. A new CEO at one of the larger standalone ATS providers, arriving at a moment when the ATS category is under real pressure from platform consolidation and AI-native alternatives — the timing makes this a strategy story as much as a leadership one. (ICIMS)
isolved names Michael Haske CEO as company enters its next AI-led growth phase. Two CEO changes at mid-market HCM platforms in the same week — isolved and ICIMS both reaching for fresh leadership as AI reshapes what "growth" actually means in this category. (isolved)
Phenom achieves FedRAMP Ready status to advance U.S. government agency hiring. A quiet but strategically important certification — government hiring is a massive, underserved market, and Phenom just cleared a significant barrier to entry that most talent platforms haven't bothered with. (Press Release)
Origin partners with ServiceNow to bring benefits intelligence directly to employees. Benefits navigation embedded into the workflow platform employees actually use daily. The "last mile" problem in benefits administration is real, and this integration is a sensible stab at it. (Press Release)
Workvivo launches Seer, a standalone people intelligence platform. The headline stat from their own research: 62% of employees are comfortable giving feedback, but only 49% see any change as a result — which is either a strong market thesis or a damning indictment of how Workvivo's existing customers have been using it. (Press Release)
Atlassian adds AI enablement to its people strategy. The company that laid off 10% of staff last month citing AI efficiency is now building AI into how it develops and supports its remaining people — a tension the press release does not mention and the market should probably notice. (HR Chief)
Upwork rolls out Spring 2026 updates, deepening Uma AI across the platform. The moves are real — AI shortlisting on the Basic plan, in-meeting contract generation, work diary summaries — and the underlying story is that Upwork is quietly becoming an AI-native talent marketplace while the freelance market itself gets more complicated. (Press Release)
Payscale launches Smart Reporting, letting comp teams query compensation data in natural language. The pitch is "data custodians to strategic advisors" — a reframe that will resonate with every comp analyst who has ever spent a Thursday rebuilding the same spreadsheet for the fifth time this quarter. (Payscale)
Deel launches Akai, an agentic AI layer designed to eliminate manual HR workflows. The promise is bold — "manual work stops here" — and the product targets the same operational overhead that has made payroll and compliance such a pain point for globally distributed teams. (Deel)
Built In launches the first employer intelligence platform for AI search. As candidates increasingly find jobs via AI-powered search rather than job boards, Built In is betting employers will pay to be findable. It’s a reasonable bet, and a quiet acknowledgment that Indeed's distribution model has a new ceiling. (Press Release)
Eightfold AI announces agentic interviewing integration with Oracle Fusion Cloud Recruiting. Two enterprise heavyweights connecting their AI layers directly into the hiring workflow is a sign that agentic interviewing is moving from experiment to infrastructure faster than most TA teams are probably ready for. (Eightfold)
The Workday–A16z debate has five smart takes. Here's what they're all missing. George LaRocque had one of the sharpest pieces of meta-commentary on the enterprise software valuation wars. It’s a useful context for anyone trying to make sense of the week's platform noise. (1WorkTech)
Mom and Dad: The performance review. “Mom, having three glasses of wine (‘because of the pandemic’) doesn’t mean it’s O.K. to loudly sing Charli XCX.” (The New Yorker)
The office of the future sounds like a library full of people talking to themselves. Nobody warned us the post-keyboard future would be this weird and this domestic at the same time. (TechCrunch)