I’m not sure whether your office communications are confined to Slack, Microsoft Teams, or if you’re using good ol’ Outlook. One thing they all have in common is that they trained us like Pavlov’s dogs.

Nowadays, those pings aren’t confined to the sigh of yet another meeting invite that could have been reduced to an email or message. The emotional response to the “ping” has changed. Besides exhaustion or eye rolling, we’ve added angst-inducing values.

“Will the next Teams or Zoom invite be another meeting?” Maybe it’s a “mass meeting”. Will my team and I stare at a screen, finding commonality not in the task, but in the message? Words like reorganization, focusing on strategic growth areas, empowering employees through technology, etc.

Those are all the necessary words to set the stage for the one word that wrecks lives, destroys families, and impacts futures: lay off.

One long-serving employee named Carsolina Walton, whose commitment spanned over seven years at Microsoft, publicly chronicled her experience on LinkedIn. She recounted how her day took an unforeseen turn when a last-minute meeting was added to her calendar by her immediate superior. Initially, she was filled with confusion, trying to grasp the purpose of the urgent gathering, which mentioned topics ranging from the company’s priorities for the upcoming fiscal year to potential reorganization plans. Her realization was stark and sudden when an unfamiliar voice joined the call—her introduction to the reality that she, too, was among those being laid off (ET).

In the Valley, it’s high-fives and flutes of Dom. “92 million jobs displaced, 170 million created,” investors and shareholders cheering each other, like the numbers were just avatars in a PowerPoint, not lives uprooted. The story might look good on CNBC, until you watch this as an affected (former) employee or recognize that they’re cheering your demise.

In recent history, the words used to justify a layoff (reduction in force) have a common denominator: Large Language Models + Computation = what we know to be called “Artificial Intelligence”.

“Not using AI is stagnation,” says Shopify CEO Tobias Lütke from the plush security of his billion-dollar algorithm-proof office. The revolution isn’t coming. It came quietly, in the background of Zoom calls and SCRUM/Project Management (Jira, Kanban, etc.) boards. Most of us were too busy gaslighting ourselves into “adaptability” and “how to best use the tools” to recognize that we’ve invited the “enemy” into the boardroom. Perhaps there’s an argument that “if you don’t use it, others will. And those who will, will bypass you at breakneck speed”. However, I also argue that no one saw how fast this thing would bite.

Life in the Fast Lane

Let’s start with the speed. Generative AI didn’t casually stroll into our lives. It launched. ChatGPT hit a million users in days, making every prior tech adoption curve look quaint. Mary Meeker’s May 2025 report spells it out: adoption is happening at twice the speed of previous tech cycles. It’s an investor’s and tech protagonist’s dream: Not a slow curve of adoption, but a hockey stick race to getting our hands on it. We’re traveling on the road to adoption at 300 mph, and we’re slowly waking up to the fact that we’re heading toward a wall. The crash is inevitable. Whether the airbag of social safety deploys is questionable.

Here’s more from Tobi:

Maybe you are already there and find this memo puzzling. In that case you already use AI as a thought partner, deep researcher, critic, tutor, or pair programmer. I use it all the time, but even I feel I'm only scratching the surface. It’s the most rapid shift to how work is done that I’ve seen in my career, and I’ve been pretty clear about my enthusiasm for it: you've heard me talk about AI in weekly videos, podcasts, town halls, and… Summit! Last summer, I used agents to create my talk, and presented about that. I did this as a call to action and invitation for everyone to tinker with AI, to dispel any scepticism or confusion that this matters at all levels. Many of you took up the call, and all of us who did have been in absolute awe of the new capabilities and tools that AI can deliver to augment our skills, crafts, and fill in our gaps.

Trends Artificial Intelligence

12.1MB ∙ PDF file

Download

Download

“AI is here today,” Meeker et al write. But the language of disruption feels like a threat. It’s more than merely a disruption. AI, it’s use cases, and combination with advancing robotics is nothing short of dispossession.

Take James, a newly minted paralegal. His first year was supposed to be spent in the paper mines, combing through contracts and redlines like every associate before him. Instead, he watches as GPT-law devours 9,000 pages in minutes and flags things he’d need two weeks to spot. “They say I should focus on higher-level work,” he laments over bourbon (hopefully the good stuff). However, what is higher-level work? (preview: it’s solution-seeking work) His firm hasn’t laid him off yet, but he’s still in a probationary period. Chances are, he won’t make it out of probation, into FTE employment. Behind him, the firm had already stopped hiring anyone else like him. Door shut. AI will do it faster, cheaper, and better.

The canary in the coal mine? It’s not chirping. It’s also deeply immersed in GPT, prompting “Act as an Ornithologist, and tell me: How do I get my wings back?”

Those aren’t isolated glitches. Those are systemic patterns. Jobs that deal in repetition, pattern recognition, or data crunching are being scalped by the machine. Tax prep, bookkeeping, claims processing, insurance approvals or reviews, once white-collar moats, now AI chum.

• Administrative assistants? Being replaced by AI agents with better recall and no lunch breaks.

• Customer service? Outsourced to emotionally bland chatbots with infinite patience and no sick days.

• Entry-level coding? Stripped down to prompt engineering, vibe coding, and QA audits.

The numbers back the dread. Companies are already slowing their hiring engines, quietly testing whether AI can fill the gaps without pissing off clients. Spoiler: it can.

But here’s where the marketing spin hits: the World Economic Forum says it’s not all doom. Yes, 92 million roles will be displaced, but 170 million will be created by 2030. Sounds hopeful, until you realize you might not get one of the 170. Their spin is that AI will provide more jobs than it takes.

When I look at the latest advancements in robotics, I have my doubts:

Ask Maria. Fifteen years transcribing for a hospital system, her hands know every medical term and ICD10 code like muscle memory. Then, ambient AI arrives, recording visits, summarizing diagnostics, and writing the notes. For some hospitalists, clinicians, and providers, AI is simplifying their job. For others, it endangers their career. This is the edge: where AI becomes a silent triage mechanism, deciding who stays, who pivots, and who’s reduced in force, because AI is faster, and over time, provides even greater standardized accuracy.

From a personal perspective, I’ve seen the onset of this in 2005. I worked at a medical software company, dispatched to Memphis, TN. On location, I trained their staff, from triage to oncologists, to the owners of the clinic. They had satelite offices as well, breaching into Mississippi. For thosen who are local to TN, you may know which clinic I’m talking about. I had the “privilege” to train the “head honcho”, and as went to through the rules engine of how treatment plans are constructed, from Dx to Rx, I saw his wheels turning, calculating on the fly. His calculation wasn’t geared toward patient relationship improvement. At some point, he drifted into business mode, “this means we work faster. We can schedule more patients.” That was in 2005. We’ve come a long way since. The patient is an electronic number, and while some providers still practice with great empathy, the “suits”, “analysts”, and the “shareholders” don’t care whether you’re confronted with a deadly disease. They care about unit economics.

But not all work sinks. The survivors share some DNA: empathy, creativity, dexterity, and ambiguity. Although not eternally safe, registered nurses, therapists, and early-childhood teachers may have a longer runway before AI and robotics are coming for their paychecks, too. Not because AI can’t learn the scripts, but because it can’t build innate human trust. AI is optimized, but care, when done right, is ambiguous, nuanced, deeply interpersonal, and messy. And messy breaks the algorithm.

Then there’s the skilled laborers: plumbers, roofers, motorcycle mechanics. The hands-on warriors who fix what breaks in the physical world. Try asking a transformer model to fix your septic tank. For now, the individual fixer is safe. But, not for long.

Then, add a drywall robot to the mix. Note: this video is six years old! Six years, in AI and robotics. That’s an evolutionary eternity. Compare and contrast with the other videos in this article, and now imagine the fluidity of movement applied to the six year old video of the drywaller.

And in this twisted irony, those building the AI, from the top level engineers, to the data scientists, ethics officers, model trainers, are watching demand explode. Structurally, though, they airspace will get tighter and tighter, even for the specialists, who themselves will get worked out of a job by the creations they helped forge.

Sam Altman, OpenAI’s patron saint of techno-optimism, argues AI will become our “copilot.” Our “research assistant.” Our “thought partner.” But that sounds a lot like intellectual outsourcing. The future isn’t about replacing you. It’s about replacing the parts of you the machine finds inefficient. And considering our productivity scale, humans cannot be productive 100% of their alloted work time. Context switching alone, is a major brain drain, or lunch, or “zoning out” while thinking about how to reply to an email or customer. Not to mention vacations, sick time, maternity leave, or having “an off day”. AI and Robotics? No such needs. Efficient, productive, and while at first more expensive - long term, more affordable, and less of a headache for HR or management.

Share

A friend of mine, a mid-career graphic designer, explained it best: “I used to spend 70% of my time in InDesign, Photoshop, and Illustrator. Now I spend 90% pitching concepts. The AI does the rest. I am producing faster. But it’s not the same job.”

That “not the same job” line? That will likely be the punchline of the decade.

And the capital flows? They’re betting hard on this new normal. Big Tech’s AI infrastructure spend is up 63% year-over-year. Nvidia’s data center revenue just exploded 78%, like a silicon gold rush. Palantir’s commercial AI bookings are up 65%. We’re beyond the basics of Research and Development, but knee deep in the battle of AI & Robotics supremacy. The AI arms race is now driving exponential training data growth (260% annually), compute usage (360%), and server demand. The language models are getting smarter by the week, while getting greedier for electricity, bandwidth, and human redundancy.

On a long enough timeline, the driving question is this: who’s steering the ship, who’s still needed aboard, and who’s dead weight, being tossed overboard?

Celine McNicholas of EPI frames it bluntly: It would serve the interests of exploitative employers to lose focus on the longstanding systemic reforms that are needed to restore worker power and give workers a voice in the use of AI.” Otherwise, this tide won’t just reshape work. It’ll reshape who gets to be human in the eyes of the market.

The terrain is global. Internet penetration has hit 68%, up from 16% just 19 years ago. The machine has a map of most of humanity now. Every keystroke, every resume, every purchase. Then you must consider that AI isn’t local. It’s planetary. And the decisions being made in San Francisco and Shenzhen are shaping what survival looks like in Nairobi, Naples, and Nebraska.

Takeaways

The AI tide isn’t theoretical. It’s not future tense. It’s here, and it’s hungry.

This is not just another industrial revolution. It’s evolution on high-performance drugs. Capability doubling every 3–4 months. Compute power up 360% annually. The old playbooks are ashes.

The raw math is this: if your job is built on patterns, repetition, or information retrieval and structural interpretation, then you’re a mark. Not today, maybe. But soon. If you can’t explain why your work matters outside of execution, you're already replaceable.

The traits that protect you aren’t “soft.” They’re the new hard skills:

• Empathy.

• Creative ambiguity.

• Physical dexterity in unpredictable environments.

• Strategic decision-making (for now).

• Building human connection that feels real.

In addition, until we’re being replaced, academic skills, formerly looked down upon, will be the bedrock of effective use of AI: English, Communications (systems dialogue), and Philosophy (Logic), because high-quality prompts/instructions depend upon effective use of ECP.

So, if you’re an English, Comms, or Philosophy major, or exceedingly interested and capable of using ECP, you’ll be in demand….for now.

Global Impact Sidebar: AI’s Ripple Effect Across Borders

The real shockwave of AI displacement is hitting far beyond Wall Street and US college campuses. For emerging markets and the Global South, regions long positioned as the outsourced labor engine of the West, the AI tide is a tectonic shift that threatens to sever the global value chain.

Take India and the Philippines, where millions are employed in BPO (Business Process Outsourcing) industries, from call centers, data entry, tech support, to back-office finance. These jobs were once immune to automation by virtue of being “cheap enough not to replace.” That logic no longer holds. The same is true for production jobs.

According to Oxford Economics, up to 20 million manufacturing jobs could be displaced globally by AI and automation by 2030, with outsourced knowledge work at risk even sooner. The World Bank warns that AI and robotics could eliminate two-thirds of jobs in the developing world unless rapid adaptation occurs.

But this isn't just about job loss. It’s about sovereignty erosion. When a Western firm can use AI to analyze contracts, run customer support, and generate marketing content in-house, the economic dependency loop breaks. Labor-rich nations will need to scramble to retool their education systems, digital infrastructure, and fiscal policy.

Countries like Kenya, Vietnam, and Bangladesh now face a grim dilemma: either become training hubs for AI systems they don’t own, or risk being left behind entirely in the new digital caste system. Meanwhile, multinational corporations hedge their bets: automating functions in one region while testing "AI plus human" hybrid models in another.

Leave a comment

US Economy: Reconfiguration, Not Recovery

Domestically, AI is as much about job loss as it is about labor reconfiguration, translated to a restructuring that will gut the working and middle classes under the illusion of innovation.

The Bureau of Labor Statistics reports steady overall employment, but underneath the hood, automation is silently amputating entire job categories. Goldman Sachs projects that 300 million jobs globally could be affected by generative AI, including a quarter of U.S. work.

The Congressional Budget Office warns that AI-driven displacement could suppress wage growth and exacerbate income inequality, especially for non-college-educated workers. And while tech stocks soar. Nvidia up 78% YOY in data center revenue, the rest of the economy shows signs of bifurcation: high-profit tech monopolies on one end, stagnating service economies on the other.

The Antidote

You want to survive this? You’ll need more than a resume refresh. Let’s try a tactical blueprint:

The only way to get clarity is by doing the work. Take a few hours, focus, and dig deep:

1. Audit Your Vulnerability:

• List every task you do.

• Circle the ones that are routine, data-driven, or pattern-based.

• Those are the first to go.

• Now, circle the ones requiring human discretion, ethical nuance, creative leaps, ambiguity, messiness, or client trust. That’s your moat…for now.

2. Build the Skills AI Can’t Touch (Yet):

• Solve messy, undefined problems.

• Learn to embrace ambiguity, which flies straight in the face of our desire for uncertainty avoidance.

• Cultivate emotional intelligence like it’s a second language.

• Connect across silos, industries, and contexts.

• Handle physical environments that don't follow scripts.

3. You won’t be able to avoid it; thus, master the Interface:

• Become your occupation’s AI whisperer.

• Learn prompt design. Not just how to prompt, but what to prompt.

  • Read this: OpenAI Prompt Engineering Best Practices

• Know the limits. Know the hallucinations. Understand bias!

• Direct the tool, go beyond just using it.

4. Shift Your Career Psychology:

• Measure value by outcome, not method.

• Let go of task identity. Embrace strategic thinking.

• Trade “what you do” for “what you solve.”

5. Diversify Your Career Portfolio:

• One job? One income stream? That’s playing Russian roulette with five bullets in the chamber.

• Develop secondary skills. Side income. Optionality.

  • This sucks, because you’re time starved and burned out already. However, it beats the alternative of being obsolete.

6. Learn Publicly.

• Build in the open. Share your experiments. Create with the intent to build community.

  • It doesn’t have to be Substack or YouTube. It can be at home. Share your work.

  • Yes, in part, that’s what I’m doing here, through Substack. But I’m not the final measure of success - far from it.

• In the AI age, your reputation is your resume, because your method can potentially be replicated. Thus, it turns to your skill of communicating the outcome of the problems you’ve solved.

Ask yourself every 6 months and re-audit:
"If AI took over 50% of my job tomorrow, what would still make me worth hiring or to keep me on staff?"

The AI tide will swallow the unprepared. But if you become the bridge, the one who understands both human need and machine potential, you’re not in immediate danger. On the contrary, you might be indispensable until AI comes for you, too.

The analogy applies: “if a bear chases you, you just need to be faster than the slowest guy.”

Question: Would you find value in a “weekly prompt”?

I will bring the prompt. Dissect it. Put it to use. Show you the outcome. Then, we’ll discuss the function, application, and how to modify it for your use. This would be a paywalled section, accessible only to you and the small cohort of like-minded.

###

~Z.