Don’t Panic: A Practical Handbook for Navigating Job Transitions in the Age of AI – Substack

I wrote this because I’m really dissatisfied with AI and jobs discourse. 98% of what I read feels designed to drive clicks rather than be helpful. Everywhere you look, someone’s sounding the alarm: “AI is coming for your job!” It’s frankly exhausting. I feel the tiredness in my TikTok comments. I feel it in the unspoken questions swirling around the edges of tech conferences and get togethers. There will be more news about it on Monday when Meta’s layoffs go from rumor to news (company communications were today).

Amidst all this chaos and the ever-accelerating path of model performance, how do we think constructively? I want to suggest that it’s worth examining a bit more whether the old lessons of history might still hold. I think we’re all assuming that AI is a special case, and it’s somehow different from all the other tech waves that have come before—but that’s what every tech wave has looked like in the middle! What if AI is the same? What if the lesson hasn’t changed? And the lesson is this: When technology automates tasks, human ingenuity gets a promotion. We are a tool-using species.

Introduction

This longread takes you on a journey through the past, present, and near-future of automation. We’ll unearth insights from NASA’s “human computers”—the women who performed math by hand for the earliest aerospace missions—and draw parallels to our modern era. You’ll learn why today’s jobs are better understood as bundles of skills, each vulnerable (or not) to automation’s steady march. We’ll also explore the concept of tipping points in technology adoption—how to watch for them, and how to prepare for the moment they arrive. Lastly, we’ll look at a mini-case study in engineering and product management, two fields whose lines are blurring even as automation handles more of their routine tasks.

Think of this read as your compass for navigating an ever-shifting employment landscape. If you’re worried about job security—or simply curious about how AI is reshaping the future of work—this guide will show you the signals to watch, the skills to nurture, and the mindsets to adopt.

I. Rethinking the Automation Narrative

Media Hype vs. Measured Reality

Not long ago, my social feeds were filled with bleak discussions about “the end of work.” You’ve probably seen the same: articles with clickbaity titles like this: “Robots Are Taking Over!” or “In Five Years, Your Job Won’t Exist.” On LinkedIn, I’d see anxious debates under posts from CEOs praising the latest automation tool. Conversations swirled around “We’re doomed!” there and on Reddit.

But beneath the drama lies a more nuanced truth—one that repeatedly shows up in history. Automation certainly does alter the job market, but rarely does it eradicate entire swaths of employment overnight. Instead, it often chips away at specific tasks, requiring us to reconfigure the way we work. Yes, entire roles can fade out, but typically in the disruptive chaos that emerges new opportunities emerge just as quickly. The net effect is change, not necessarily catastrophe.

Why this matters to me

I’ve felt that “change, not catastrophe” dynamic in my own career. My very first foray into tech involved managing an Oracle iStore—an online storefront system that required specialized knowledge. My title was E-Commerce Manager, and that title is all but dead now. When I checked writing this post, there were just 277 jobs left on LinkedIn with that title. Over time, updates and advanced e-commerce platforms did away with most of the need for that role, and the title quietly disappeared from job boards.

My next role was Conversion Optimization Manager, and I watched that one disappear too. There’s only 327 of those jobs on LinkedIn now. Imagine if that was my career path all these years! I thought it was for awhile. But the role has faded away as analytics software and automated testing tools took over large portions of that job. Now there’s a startup that does more than I could ever do with testing with AI: personalizing the site automatically on a per visitor basis.

It’s a good thing I didn’t bet on those roles! The tech wave kept coming, and as a human I had to shift toward tasks that enabled me to deliver a higher level of business value with the tools in hand. I think that’s a fundamental pattern we’ll see across this guide.

To confront the elephant in the room: I am not convinced that pattern breaks even with Artificial General Intelligence (apparently) coming soon. Why? Because intelligence without context is mis-applied and adds little value. We’re seeing already how hard it is to add agency to intelligence. Operator is clunky. Deep Research does one thing very very well. We may see more agents like Deep Research. Agents will get smarter. But even smarter agents need good training data, and they will have difficulty soaking up the unwritten context that drives most of business. And no, I do not think an employee recording a Loom video will be enough.

II. Historical Context: Lessons from NASA’s Human Computers

A. The Era of Manual Computation

The notion of “human computers” often catches people by surprise. In the mid-20th century, long before the invention of microchips and personal computers, the word “computer” described human beings—usually women—who performed painstaking math calculations by hand. Places like the National Advisory Committee for Aeronautics (NACA, later NASA) employed entire rooms of these mathematicians to plot flight paths, test aerodynamics, and compute rocket trajectories.

Think about the context: America was in the throes of WWII and later the Cold War. Speed was paramount in military and aerospace research. With no digital mainframes to rely on, a single rocket test might require days of arithmetic scribbled on reams of paper. The “human computers” were literally the processing power for these critical computations (Wikipedia).

But of course, we know that job went away. Let’s explore a few stories of former human computers and discover what happened next…

B. The Legacy of NASA’s Human Computers

Katherine Johnson’s Precision

Katherine Johnson’s story is perhaps the most famous. Gifted with an extraordinary aptitude for numbers, Johnson calculated the trajectories for America’s earliest space missions, including John Glenn’s orbit around Earth. Glenn famously insisted Johnson verify the numbers generated by the newly installed IBM systems, saying, “If she says they’re good, then I’m ready to go.” In that single request—machine output verified by human logic—you see the classic interplay of technology augmenting but not fully replacing human skill. Katherine Johnson helped break the color barrier at NASA and ultimately went on an aerospace technologist after human computers were replaced by IBM mainframes.

Dorothy Vaughan’s Adaptability

Dorothy Vaughan, another pioneering figure, also started as a human computer. When NASA introduced digital mainframes and programming languages like FORTRAN, Vaughan realized that mastering these new tools would be crucial. She proactively taught herself FORTRAN and then made sure her entire team of women mathematicians acquired the same skills. In this, she demonstrated a theme we’ll see repeatedly: those who anticipate and adapt to emerging tools not only survive but often lead in the new technological era. She taught Katherine Johnson and eventually joined the Analysis and Computation Division at NASA.

Susan G. Finley’s Continuous Evolution

Susan Finley rounds out this trio of role models: she began working on rocket trajectory calculations by hand, but swiftly transitioned into writing FORTRAN code as computing systems became more reliable. Decades later, she remains a pivotal figure at the Jet Propulsion Laboratory and has worked on NASA’s Mars missions, evidence that fundamental skills—mathematical reasoning, problem-solving, creativity—remain valuable even after the tools and methods drastically evolve.

C. Tipping Points and Transformation at NASA

The “human computer” role effectively vanished once NASA fully trusted its IBM mainframes. But rather than fueling a wave of mass layoffs, this shift propelled many human computers into advanced roles—programming, system design, oversight, and more. The advent of digital technology introduced a new set of tasks that required fresh skills, but also a deep understanding of math and science—expertise these human computers already possessed.

Now I can hear what you’re saying: “that was one company, what about the industry? AI is transforming things at a different scale.” I disagree. First the human computer was an industry wide job classification with industry-wide implications. Second, yes AI is bigger but the arc of the transformation story is very similar: humans do a critical technical task because technology can’t until computers can take over. That’s the story of AI too! (And yes, operating a computer to build economically valuable work is a technical task—even if you’re in marketing!)

So why does this story resonate today? Because it’s a perfect microcosm of the automation process. At first, technology merely supports the human role. Then, as the tech matures, it automates significant chunks of that role. Eventually, the people who once did those tasks either move on or learn the emerging technology, evolving into higher-level positions. Understanding that cycle is crucial for interpreting our modern concerns about AI taking over jobs.

III. Jobs as Skill Bundles: Rethinking What We Do

A. Deconstructing the Job

Now, let’s bring this narrative closer to home. When you think of your job title—maybe “marketer” or “software engineer,” or “product manager”—you typically imagine one cohesive role. It looks like that on the job description! But on closer inspection, that role breaks down into distinct tasks and competencies. A product mansger isn’t just “someone who sits in meetings.” (I kid lol.) The job also involves research, brainstorming concepts, employing empathy to resonate with specific audiences, building requirements, working with engineers, conducting QA, and pushing through the launch process.

Automation can easily replicate pieces of that puzzle—like composing clear requirements from bullets—while other elements remain human. Want a human example? How about sorting out how to frame a product roadmap for a CRO who’s made a promise to a high value client—that is a human problem. And if you’re a PM you’ve been in that position. I suspect it is these harder-to-automate skills that become even more critical as routine tasks get offloaded to machines.

B. Evidence from the Gig Economy

None of this means that jobs aren’t rapidly evolving in the age of AI! For a frontline view of this trend, look no further than freelance platforms such as Upwork and Fiverr. A study from 2024 shows a 21% drop in postings for “low-skill, repetitive” tasks—like basic copywriting, data-entry, and template-based design. It’s probably accelerated since. At the same time, there’s a notable uptick in listings that ask for “AI-augmented” or “AI-assisted” skills. This shift can look like an employer saying: “We’ll generate a first draft using ChatGPT, but we need a creative writer to refine the voice and ensure factual accuracy.”

This realignment isn’t just for the gig economy. To me it suggests that “human oversight” or “strategic guidance” is becoming a new locus of value in the AI economy. And that feels like an appropriate insight for the agentic economy. Indeed, an Upwork Research Institute paper noted that as certain tasks become automated, they can start to erode away from the open marketplace; the gigs that remain command higher rates precisely because they require the intangible, irreducible traits that AI can’t replicate.

I’ll add that these effects were very moderate at marketplace levels: 1.3% increase in earnings per contract and mixed results across job families, with -4% earnings for Customer Success roles. An Upwork study later in 2024 showed that 77% of workers thought AI had increased their workloads. There seems to be a lot of work to go around.

Through the lens of my own work history, I’ve personally seen how routine portions of my past jobs got swallowed up by automation. It happened with the Oracle iStore manager role, and again with conversion optimization. Each time, I worried about what would replace those tasks. Then a new layer of responsibility—management of more complex workflows, creative problem-solving, or bigger-picture strategy—emerged. In hindsight, it’s clear that many job titles vanish primarily because the discrete skills underneath them either get automated or re-bundled into new roles that prioritize higher-level thinking. And the evidence so far on AI doesn’t an indicate that that pattern is changing.

IV. Tipping Points and the S-Curve: Measuring and Understanding Change

A. The Dynamics of the S-Curve

Let me introduce a handy model for understanding how new technologies displace (or augment) human roles. Welcome to the S-curve of adoption:

So what do these curves look like at work?

1. Initial Phase

Adoption of the tech is slow and experimental. Think NASA’s first IBM machine, running side-by-side with the human computers. Organizations are exploring the technology’s reliability, cost-effectiveness, and practical applications.

2. Rapid Growth (Tipping Point) Phase

Once a technology demonstrates its viability—often when it can automate at least 50–70% of a routine task—adoption spikes. This is the tipping point. Consider how quickly companies are now flocking to AI tools once they see a real competitive edge in speed or cost savings.

3. Maturation Phase

The market stabilizes, technology becomes ubiquitous, and any “routine” tasks that could be automated have been. By then, the purely human-centric tasks (like complex strategic thinking or nuanced decision-making) soar in value.

B. Recognizing the Signs Beyond Metrics

It’s tempting to just look for a big statistic—like “AI replaced 70% of call center tasks!”—to mark the tipping point. But the narrative behind the number is equally instructive. For instance (and this is hypothetical), if Google Trends shows a sudden surge in searches for “ChatGPT coding,” and you concurrently see an 8% drop in entry-level coder job postings, those two pieces of data together paint a more compelling picture than either alone.

For another example, consider a corporate memo stating, “We expect our new AI system to handle 60% of claims processing tasks by Q4.” That signals a confident commitment. It’s even more confident when the company will say something externally. But one company isn’t enough: to really understand how roles are changing you need to have the discipline to look at clusters of statements that correlate around particular roles. If that sounds hard, it is! Someone should probably build an AI driven product to analyze job automation signals. Observing these internal statements, plus external data like job-market shifts and regulatory adjustments (like new guidelines for AI-generated content), offers a holistic way to measure when routine tasks are being phased out.

AI matters enough to get this right. We need to be willing to be holistic. We need to look at both the quantitative and the qualitative. A 20% drop in postings for a particular task is notable, but why is it dropping? Is it a momentary dip, or is the skill fundamentally being automated into obsolescence? Is there an uptick in roles that require a more strategic level of thinking in that same field? When you see those puzzle pieces fit together, you’re witnessing the reconfiguration of job skills in real-time.

V. Engineering and Product Management: A Mini-Case Study in Role Evolution

This has all been very broad, and I don’t want to leave you with platitudes. Let’s move from broad strokes to something specific: how engineering and product management roles are morphing before our eyes, largely thanks to AI-driven tools.

A. Engineering: From Routine Coding to Strategic Integration

Look at any modern engineering role, and you’ll notice a growing expectation to use AI-assisted coding tools—like GitHub Copilot, ChatGPT, or specialized machine-learning frameworks. Early studies (and confident statements) suggest these AI tools can confidently handle 20–30% of routine coding tasks (writing boilerplate code, generating test scripts, debugging typical errors). Projections show that number inching up to 50–60% within the next few years (here and here).

What does this mean for engineers?

It means that the “mundane” tasks that used to take up hours of coding time are being offloaded. As a result, the remaining tasks—designing system architecture, orchestrating how different software components interact, and making creative decisions about the best approach—become the engineer’s primary focus. There’s already data suggesting that companies are paying a premium for “AI-augmented” engineers who can effectively prompt and supervise AI tools.

I’ll add there’s evidence of here of more work created by AI. One of the hottest topics in tech right now is how engineers handle AI-generated code. How is it integrated? What components are acceptable to build with AI? How does AI code align with the existing codebase conventions? These are new problems AI is generating for engineers.

Indicators of a Tipping Point

One key signal is job listings themselves: when 25–30% of engineering roles explicitly mention AI proficiency or “AI-augmented development,” it’s likely we’re at a tipping point. If you’re wondering what the number is today, I did some work on that and my best estimate is somewhere around 5-10% of total engineering roles, up from 5% in 2024. Another sign might be your own project workflows. If your team is collectively saving 50% of coding time using Copilot, that indicates the technology has moved from novelty to necessity.

B. Product Management: The Rise of the Hybrid Role

Product managers, traditionally, have wrestled with tasks like data aggregation, basic analytics, user research, and writing feature specs. AI can now automate large parts of that workflow—for instance, collecting user data from multiple sources, processing it, and even generating initial summary reports. Once upon a time, an entry-level product manager might spend hours sifting through usage metrics; these days, specialized AI dashboards can do it in seconds, and Kraftful does the same for anecdotes (which used to take me days).

How do PMs stay relevant?

PM is a bridge role, and I won’t pretend to know exactly where my job family is going. It’s fundamentally hard becuase it’s a sponge role: bridging technical capabilities with market needs, interpreting AI outputs to form product roadmaps, and ensuring that user empathy shapes final decisions. Indeed, recent data from various tech job boards show a decline in entry-level PM positions (albeit perhaps not below historic norms) coupled with an increase in senior or hybrid roles that combine analytical, leadership, and creative responsibilities. There’s also evidence that entry level definitions are changing and leveling up.

Indicators of a Tipping Point

Look for more product management job postings emphasizing “AI collaboration” or “machine-generated insights.” If you’re wondering, ~16% of PM roles in the US mention AI now, but less than 2% do globally. Especially pay attention to pricing power for these roles. I did a quick and dirty estimate to get to the PM roles estimate above, and the same Perplexity thread yielded an estimate of an eye-popping 65% pay bump for AI PM roles vs. traditional PM roles. I conclude product management may be closer to a tipping point than engineering.

C. Blurred Boundaries: Convergence of Engineering and Product Management

What’s arguably most exciting is how these fields are merging. As engineers incorporate more product thinking—considering user experiences and market implications—and product managers grow more technically proficient, the line between the two disciplines fades. In some organizations, it’s not unusual for senior engineers to handle aspects of product strategy, while PMs learn enough coding to tinker with prototypes or run AI experiments. Some organizations are just doing away with PM’s altogether (which I have mixed feelings about).

This cross-pollination fosters teams where everyone has a stake in both the technical and human-centric sides of product development—a synergy that becomes increasingly advantageous in AI-driven environments. In short, we’re witnessing a new hybrid role starting to emerge we don’t have a good name for yet. Maybe we’ll call them “strategic technologists” or “technical product visionaries,” but regardless of what we name them we’re going to see more blurring here between traditional PM and Eng roles, largely driven by AI.

VI. Embracing Transformation with Confidence

I hope this has been a helpful read. I hope it’s been more interesting than mos tof hte panic reads we get on these topics. I want to leave you with some practical paths forward.

The Human Edge

First, believe in the human edge. From NASA’s hand-calculators to the unstoppable rise of AI in coding and beyond, one truth endures: automation does not end work—it reshapes it. Tasks become unbundled, roles get reconfigured, and the people who adapt find themselves in more creative, strategic positions. Routine, repetitive tasks vanish, but new avenues—requiring holistic thinking and emotional intelligence—are born.

Practical Next Steps

1. Audit Your Skills

Break down your job into its core tasks. Which are repetitive and rules-based? Which require empathy, nuanced judgment, or creative thinking? Focus your professional development on the latter.

2. Monitor the Data

Keep an eye on job postings, Google Trends, and corporate memos in your sector. A consistent decline in one area and a rising demand in another means you’re at the cusp of an S-curve acceleration. And yes, you can ask ChatGPT or Perplexity to help you with this!

3. Embrace the Hybrid Mindset

If you’re in engineering, develop your communication and product insight. If you’re in product management, strengthen your technical literacy (I have a Maven course for that). Hybrid skill sets are quickly becoming the bedrock of many future-facing roles.

4. Stay Agile and Curious

Rather than fear new tools, explore them. Dorothy Vaughan’s story at NASA taught us the value of self-driven upskilling; you’ll not only remain relevant but can also help guide your organization in using these tools effectively.

The Road Ahead

Yes, automation will absorb a large portion of rote work. AGI will absolutely reshape work. But over the long arc of history, I think we are seeing a new chapter of the same story we saw when mechanical looms replaced hand-weavers, and again when word processors made typewriter pools obsolete. Over time, people move up the value chain. The key is to sense these shifts as early as possible—by recognizing tipping points, reading market signals, and investing in the strategic and empathic layers of our professions. I think learning how to keep an eye on these shifts for our jobs and planning ahead is a much more useful response than the panic the newspapers generally advocate.

In many ways, the NASA human computers laid down a blueprint. They were experts in math and logic, but they also had the foresight (or the mentors) to see that new technology enhances the fundamentals rather than destroying them. As you navigate today’s AI wave, remember that the heart of your value is the creative, strategic, and emotional intelligence you bring to the table. Machines can take over the rote tasks, but they can’t replicate the distinctly human parts of problem-solving and collaboration.

So let this guide serve as your call to thoughtful action. Pay attention to the small, often subtle signals that foretell automation’s advance; invest in the skills that are hardest to automate; and above all, maintain a posture of curiosity and adaptability. With each shift in technology, a window opens for innovative, deeply human work—and you want to be first in line to seize that opportunity.

A Final Word

Maybe we should think of AI-driven automation as less a grim reaper and more a big mirror, reflecting back to us what is truly essential in our work. It reveals the tasks we do that a machine can’t approximate—and those tasks frequently revolve around creativity, empathy, ethical judgment, and visionary leadership. As NASA’s human computers taught us, the path from manual processes to automation doesn’t have to be a straight line to obsolescence. Instead, it can be a ladder to roles requiring deeper human insight.

No matter where you stand today—engineer, product manager, marketer, or something else entirely—take a step toward the future now. Embrace the shift, refine the skills that only a person can provide, and trust that in the uncharted waters of AI-powered work, your distinctly human capabilities will be not just relevant, but indispensable. And if that sounds too hopeful at least embrace Douglas Adams’ advice and Don’t Panic!