Prelude
The technology industry brims with apocalyptic warnings. Social media prophecies proclaim the demise of programmers, product managers, and creative professionals. We're told that algorithmic systems will automate code, interface design, and product strategy.
After two decades building systems and witnessing countless technological shifts, I recognize something deeper: the fear of replacement stems not from AI capability but from professional fragility.
Those whose entire identity revolves around repetitive tasks—formulaic coding, ticket shuffling—face genuine disruption. Yet builders and systems thinkers encounter an unprecedented opportunity.
The uncomfortable reality: you won't be displaced by technology. You'll be revealed by it.
The Orthodoxy
Current discourse presents a false binary.
Doomsayers view tools like GitHub Copilot as existential threats, equating code generation with professional obsolescence. Defensive engineers retreat into arrogance, dismissing AI limitations while clinging to low-value work.
Both perspectives miss crucial truths.
Engineering transcends typing—it's decision-making. Progress demands releasing our attachment to drudgery.
"74% of IT professionals believe AI will make their skills obsolete," according to industry research. Yet evidence consistently demonstrates that "AI is not a replacement for human expertise but rather a powerful augmentation tool." The data reveals that technological shifts displace certain roles while creating different ones—though the transformation proves uncomfortable.
The difficult truth nobody wants hearing: professional work fundamentally changes. Complexity increases. Pace accelerates. Genuine competence becomes mandatory.
The Cracks
I recently built an agent for refactoring legacy Python services. It generated flawless code—elegant typing, exceptional documentation.
It was completely broken.
The system hallucinated nonexistent dependencies, optimized unused functions, and misunderstood irrational business logic.
A junior developer gauging worth through output would have deployed it. The system would crash.
The mythology of software engineering perpetuates the god complex—the belief that syntax memorization equals genius. AI demolishes this illusion. Knowing kubectl arguments isn't genius; it's trivia. Machines excel at trivia.
The "one-person unicorn" myth similarly fractures under scrutiny. Simultaneously inhabiting product manager, developer, and designer roles creates unsustainable cognitive load. These demand different thinking modes.
Collapsing roles produces mediocrity—systems competent at everything, exceptional at nothing.
Skill obsolescence accelerates continuously. Half of current capabilities may become outdated within two years. Yet this doesn't render jobs obsolete—merely the toolsets. The "human versus AI" framing misses the point entirely: AI functions as a filter.
The Deeper Truth
AI raises the waterline of quality expectations.
Envision competence as a sea. Above water lies value; below sits commodity.
For decades, professionals accumulated wealth performing underwater work—CRUD applications, Webpack configuration. We branded this plumbing as "craft."
It wasn't.
AI floods the engine room, raising that waterline. Everything below automatically drowns.
Developers whose value proposition centered on writing React components from mockups now face submersion—this is commodity work now. Product managers who tallied meeting notes and drafted user stories encounter the same fate.
The panic erupts from those treading just above the previous waterline, suddenly aware of drowning.
Yet masters of genuine craft celebrate rising waters. It eliminates drudgery and exposes excellence.
The Death of Mediocrity
I previously spent 40% of time writing unit tests—the boring ones checking if functions return true when given "true."
Now I prompt an AI. It writes tests. I review. I refine. I continue.
Does this render me obsolete? Quite the opposite—it makes me dangerous.
I now focus on architecture, system resilience, and problem-solving rather than syntax.
The claim that novices achieve expert performance through AI is dangerous mythology. Novices using AI simply produce garbage at accelerated speed.
We're witnessing craft renaissance. "Average" has become free and automated. Therefore, "exceptional" represents the sole remaining currency.
Uniquely human capabilities—critical reasoning, empathy, creativity—constitute the realm's new currency. As commodified work automation expands, "curation and expertise skyrocket" in value.
The Quality Paradox
The Quality Paradox emerges as the cost of generating code and content approaches zero.
Volume explodes. Most AI-generated applications, articles, and products become rubbish.
Consequently, "curation" and "expertise" become invaluable.
Deleting code matters more than writing it. Saying "no" surpasses generating "yes."
This paradoxical reality: more AI creation demands greater human judgment. Genuine expertise remains our sole filter against entropy.
The Indie Agency Era
Another transformation captivates me more than any benchmark.
The demolition of silos.
Traditional relay teams fragmented work—product specifications, design mockups, engineering code, quality assurance, operations deployment. Slow. Inefficient. Boring.
AI collapses this chain entirely.
A single engineer equipped appropriately accomplishes small-team workloads. Designers prototype functional code. Product managers query databases using natural language.
We're transitioning toward "indie agency" models—small multidisciplinary expert teams leveraging AI for disproportionate impact.
The Multidisciplinary Necessity
In this environment, "not my responsibility" constitutes resignation.
Backend-only specialization becomes insufficient. Expertise must span product understanding, user empathy, business acumen.
Why? AI handles specialization. You integrate the pieces.
Modern system complexity demands interdisciplinary perspectives. The human-in-loop must understand the entire loop, not fragments.
Implications
These shifts reshape every technology position fundamentally. Static titles must transform into evolving platforms for orchestration.
Thread 1: The Product Manager
The Product Manager role persists; "project managers" pretending at product strategy vanish.
Ticket shuffling, standup summaries, release notes—AI handles these commodities.
The emerging "Builder" PM is technical, understanding their materials. They prototype solutions rather than merely requesting features. They leverage AI exploring solution spaces before engineering handoff.
They demand "AI fluency"—knowing how to prompt effectively, recognizing model limitations.
Critically, they become "Why" guardians.
When building code approaches zero cost, constructing the wrong code becomes the principal risk. PM responsibility centers on risk mitigation.
Thread 2: The Software Engineer
Software engineering endures; "code monkeys" disappear.
Memorizing API function arguments wastes effort. AI masters APIs superior to any human.
The emerging "System Architect" engineer owns whole system integrity. They treat AI-generated code as raw material requiring refinement, security hardening, integration.
They transition from "writer" to "editor."
This demands editors exceeding the writers they edit.
Engineers must grasp business context—why build this? Which solution proves most maintainable? What scales optimally?
Thread 3: The Interaction
The PM-engineer dynamic generates the real magic.
Previously, friction dominated—PMs drafted documents engineers ignored, arguments ensued.
Now, artifacts become shared. PMs deliver working AI-generated prototypes. Engineers deconstruct them. Co-creation occurs collaboratively.
Specification becomes code; code becomes specification.
This requires reciprocal respect.
PMs honor engineers' perception of technical hazards. Engineers value PM insights regarding market necessity.
Teams designed for collaboration leverage "collaborative intelligence," where PM and engineer function as unified creative dyads.
Thread 4: Kill Your Ego
Untether your value from tasks executed. Value derives from achieved outcomes.
When AI queries SQL faster than you, permit it. Don't resist proving points. Redirect saved time toward schema optimization.
The technology industry overflows with individuals considering themselves special for computer comprehension. Machines understand themselves now. You're not the translator—you're the guide.
The Mental Model
Consider a different framework for AI-augmented teams.
Abandon "assembly lines" and "handoffs."
Imagine a "jam session."
Jazz musicians master their instruments individually. The drummer doesn't play piano; the pianist doesn't play trumpet.
Yet they listen reciprocally. They react. They build collectively.
AI represents a new instrument—a synthesizer emulating anything.
The PM plays melody, establishing direction and feel. The engineer provides rhythm, ensuring structural integrity and stability.
Sometimes PMs take solos—prototyping. Engineers support them.
Sometimes engineers solo—optimizing infrastructure. PMs support them.
They perform the same composition.
Should the PM declare "I no longer need the drummer since this synthesizer has drum loops," music deteriorates. It becomes mechanical. It loses soul.
Users (your audience) distinguish loops from drummers instantly.
Respect the drummer. Respect the pianist.
Respect the craft.
Conclusion
AI doesn't hunt your job. It targets your boredom.
It comes for documentation updates, boilerplate authorship hours, and the ego-driven mythology that centering a div makes you a wizard.
The AI era functions as a mirror. If you perceive replacement, you've made yourself replaceable through mediocrity settlement. If you perceive leverage—an instrument amplifying speed, quality, scale—you've nothing to fear.
The "end of work" narrative exhausts me with its laziness and fear. The "I'm godlike" narrative frustrates me with arrogance and foolishness.
Reality: we're all apprentices once more, learning to harness this transformative force.
Product managers—cease attempting to eliminate engineers. You require them preventing disasters and ensuring visions don't crash overnight.
Engineers—stop dismissing product managers. You require them preventing cathedral construction in deserts and maintaining human-centered focus.
The future belongs to teams integrating AI while preserving humanity.
It belongs to teams respecting craft.
The waterline rises. Standing alone guarantees drowning. Standing together ensures floating.
Now I'm returning to building.