The Great Culling Is Here—And It's Not What You Think
Layoffs are skyrocketing, but the real story isn't about hard times. It's about AI making entire job categories vanish overnight.
Cloudflare just fired 1,100 people. Not because business is bad. Revenue hit a record high.
That sentence should haunt every person in tech right now.
CEO Matthew Prince was blunt about it: AI efficiency gains made those roles unnecessary. Support positions, mostly—the kind of work that used to require humans to read tickets, diagnose problems, escalate issues. Now a language model does it in 40 milliseconds. The company doesn’t need those humans anymore.
This isn’t a recession layoff. This is a capability layoff. And it’s going to keep happening.
Photo by Jatman 0007 / Pexels
The Real Signal Hiding in Plain Sight
Look at what’s actually happening across the industry and you’ll see a pattern that most coverage completely misses. Porsche just shut down its e-bike, battery, and software subsidiaries. Over 500 people gone. The CEO called it “painful cuts” necessary for “strategic realignment.” Translation: We built businesses we can’t compete in, and we’re burning cash on them. Time to focus on what we’re actually good at.
Intel’s stock jumped 490% in a year. Wall Street is betting on a turnaround that may or may not actually materialize—the article itself notes the stock rise “may be running well ahead of the company’s actual turnaround.” That’s a bubble forming in real time, driven by hope more than fundamentals.
Uber’s self-driving partner Avride is under federal investigation after a dozen crashes and a minor injury. This is the reality check the autonomous vehicle hype machine desperately needed. Self-driving isn’t solved. It’s still dangerous. And the NHTSA is paying attention now.
These aren’t disconnected stories. They’re telling you something consistent: companies are being forced to make hard choices about what actually works versus what they hoped would work. And the winners and losers are becoming obvious fast.
Cybersecurity’s New Credibility Problem
Meanwhile, the infrastructure holding everything together is getting hammered.
Canvas—the learning platform millions of students depend on—got hit by cyberattack during finals. Daemon Tools, a disk utility that’s been around forever, was backdoored for a month straight without anyone catching it. That’s a supply-chain attack. Someone infected the update mechanism itself, and it sat there silently compromising systems across who knows how many organizations.
Here’s the uncomfortable part: Mozilla just announced that its Mythos security vulnerability detection tool found 271 vulnerabilities with “almost no false positives.” That’s both impressive and terrifying. It means we’re about to see a massive flood of discovered security holes that we’ve been living with, completely blind to, for years.
The security industry is going through what every industry goes through when you dramatically improve detection: you suddenly realize how broken everything actually was.
The Jobs That Are Actually Disappearing
I need to be honest about what’s happening with Cloudflare’s layoffs, because there’s a specific angle everyone’s glossing over.
Support roles are the canary. They’re not gone because the economy is weak. They’re gone because they’re automatable. A support ticket comes in. An AI reads it. It matches patterns from thousands of previous tickets. It either solves it automatically or escalates it with context already loaded. A human used to do that work in 15 minutes. The AI does it in milliseconds.
That’s not going away. It’s accelerating.
The hard question: what else looks like support to an AI? Documentation work. Junior engineering roles focused on routine tasks. QA testing for straightforward cases. Data entry. Basic content moderation. Customer success onboarding calls. A huge portion of early-career paths in tech just became optional.
I think we’re looking at a 12-month period where you’ll see 15-20% headcount reductions across mid-market SaaS companies, all justified the same way Cloudflare just did it. The companies that do this first and clean it up will have lower costs and higher margins. The ones that lag will get hammered by comparison. It becomes competitive.
This is going to look like a wave of “efficiency” when it’s really just standardized automation finally reaching critical mass.
Photo by UMA media / Pexels
What Intel’s Stock Surge Actually Means
Intel up 490% in a year is not normal. Companies don’t move that much on fundamentals. That’s panic-buying based on narrative.
The narrative is: Intel is America’s chip company, we need them to compete with Taiwan, the government is funding it, therefore it’s a good investment. All of that is true. But it doesn’t mean their engineering is fixed. It doesn’t mean their fabs will actually be cost-competitive. It doesn’t mean they’ll ship products people want.
What it means is money is chasing a story, not a reality. And stories are great until they collide with actual quarterly results. Intel’s comeback is absolutely “wilder than it seems”—because a lot of that wildness is fiction.
My prediction: we’ll see a correction of 20-30% inside 18 months when execution misses expectations. Not because Intel is doomed, but because the stock got priced for perfection and Intel is a real company with real problems.
The Weird Bright Spot
Mother Ventures just raised a $10 million debut fund focusing on mothers as consumers and economic drivers. It’s a weird niche move. But it’s smart positioning. Everyone else is chasing AI, robots, and moonshots. Someone’s actually thinking about an underserved market segment with real purchasing power.
That’s the kind of specificity that actually works right now. Not broad strokes. Not “we’re an AI company.” Just: here’s a specific group of people we understand, here’s why they matter economically, here’s how we’re going to serve them better.
Photo by Denys Gromov / Pexels
What I’m Actually Uncertain About
I don’t know if the AI job displacement happens gradually or all at once. I don’t know if support teams stay small forever or if companies learn they need humans for the 5% of cases where AI fails catastrophically.
And I genuinely don’t know whether the self-driving industry bounces back from these crashes or whether we’re entering a long period where autonomous vehicle regulations tighten significantly and deployment slows to a crawl.
What I’m confident about: the easy wins are over. Cloudflare can cut support roles because AI is better at that specific task. But other companies can’t just wholesale automate their way to profitability. The jobs getting cut are the ones where AI actually is better. Everything else? You still need people.
The companies that figure this out correctly will thrive. The ones that just cut indiscriminately because they saw Cloudflare do it will struggle.
What I’m Watching
Cloudflare’s next earnings call (Q1 2025). Watch if the efficiency gains actually translate to margin expansion. If revenue stays flat but costs dropped 5-7%, everyone copies this playbook. If revenue drops because the support cuts annoyed customers, the whole story falls apart.
Intel’s actual product timeline. When do the next-gen CPUs land? What’s the defect rate? Are they competitive on performance and cost? The narrative only holds if reality delivers. I’m watching for any slip in the roadmap.
Avride’s NHTSA investigation conclusion (target: mid-2025). This determines whether self-driving gets funded aggressively or enters a long winter. A major safety finding could freeze investment for 2-3 years.
Daemon Tools’ post-backdoor recovery. Will anyone still trust it? Supply-chain attacks are the new frontier. Watch if this becomes case study #1 for why people move away from third-party utilities entirely.