What organizational inertia will slow down the disruption speed of the AI era

Tech optimists often predict that AI will destroy entire industries and employment structures overnight. But reality is much more complex. Anyone who has witnessed multiple economic collapse predictions should understand a profound truth: institutional inertia—the tendency of social, regulatory, and organizational systems to maintain the status quo—is far stronger than we imagine. This invisible force is slowing down AI’s disruptive pace, giving us valuable time to adapt.

Why Institutional Inertia Is More Resilient Than Expected

This is not a new phenomenon. In 2007, peak oil theories predicted the end of U.S. geopolitics; during the 2008 financial crisis, people believed the dollar system would collapse; in 2014, analysts declared AMD and NVIDIA exhausted. Each time, deeply rooted existing institutions proved their resilience far beyond observers’ expectations.

The story of real estate agents best illustrates this. For 20 years, people have been shouting “the death of real estate agents”—something that doesn’t even require superintelligence to realize. Platforms like Zillow, Redfin, Opendoor have existed for years. Yet, institutional inertia and regulatory barriers keep this industry’s survival far stronger than anticipated.

I recently bought a property. The entire transaction required hiring an agent, a reason that’s well-founded but also frustrating. My buyer’s agent earned about $50,000 on this deal, working no more than 10 hours—mainly filling out forms and coordinating parties, tasks I could have done myself. Still, market inertia and regulatory frameworks keep this role surprisingly entrenched.

This isn’t a critique of anyone. I’ve started and sold a company whose core business was helping insurance agents shift from a “manual operation model” to a “software-driven model.” Years of experience have deepened my understanding of this constant truth: human society in the real world is extremely complex, and changing anything takes longer than you think—even when you’ve adjusted your expectations for this truth.

The Software Industry’s Near-Infinite Labor Demand

Recently, the software industry has underperformed, with investors worried that backend system companies like Monday, Salesforce, and Asana lack competitive advantages and that their products are easily copied. Many predict AI programming will end SaaS companies: products become identical and unprofitable, jobs disappear.

But they overlook a decisive fact: most existing software products are generally poor.

I have firsthand experience—I’ve spent hundreds of thousands of dollars on Salesforce and Monday. AI indeed makes it easier for competitors to copy these products, but more importantly, AI enables competitors to create better products. Stock prices dropping isn’t surprising: an industry long dependent on lock-in, lacking competition, and filled with poor legacy companies is finally facing real competition.

From a broader perspective, almost all current software is garbage—an undeniable fact. Every tool I’ve purchased is riddled with bugs. Some programs are so bad I wouldn’t pay for them (over the past three years, I couldn’t even make an overseas transfer via Citibank’s online banking). Most web apps fail to properly adapt to mobile and desktop devices. Not a single product offers all the features I want.

Stripe and Linear are so popular simply because they are not as clunky as their competitors.

If you ask a senior engineer, “Show me a truly perfect software,” you’ll only get long silence and surprised looks.

Here lies a profound truth: even in the “software singularity” era, the demand for human software labor is nearly infinite. We all know that the last few percent of perfection often requires the greatest effort. By this standard, almost every software product has at least 100 times the complexity and potential for feature growth until demand saturates.

Critics claiming the software industry is about to die lack intuition about software development. The industry has existed for 50 years; despite enormous progress, it has always been in a “shortage” state. As a 2020s programmer, my productivity is equivalent to hundreds of people in the 1970s—an incredible leverage—but still leaves vast room for optimization.

Most underestimate the “Jevons Paradox”: efficiency improvements often lead to explosive growth in total demand. This doesn’t mean software engineering offers lifelong security, but the industry’s capacity to absorb labor and its institutional inertia are far beyond what people imagine. The process of demand saturation will be very slow, giving us ample time to calmly respond.

Reality and Reindustrialization Opportunities in the Physical World

Of course, labor reallocation will inevitably happen. As many prophets have pointed out, many white-collar jobs will experience shocks. For roles like real estate agents—whose value is increasingly lost and maintained only by institutional inertia—AI may be the last straw.

But we still have one last hope: America’s nearly unlimited potential and demand for reindustrialization.

You may have heard of “manufacturing returning,” but its scope goes far beyond surface meaning. We have almost completely lost the ability to produce the basic components of modern life: batteries, electric motors, microchips—the entire electrical supply chain relies almost entirely on imports. What if a military conflict occurs? Even worse—do you know that China produces 90% of the world’s synthetic ammonia? If supply is cut off, we won’t even be able to produce fertilizer, risking famine.

If you focus on the physical world, you’ll see many employment opportunities related to infrastructure projects that benefit the country, create jobs, and garner support from multiple political parties. We are already seeing economic and political shifts in this direction—reindustrialization, deep tech, and “American energy” are now mainstream discussions.

My prediction is: when AI impacts white-collar employment, the least politically resistant path will be to finance reindustrialization through “large-scale employment projects” to absorb displaced labor. Fortunately, the physical world has no “singularity”—it is subject to friction. We will rebuild bridges and roads. People will realize that deriving satisfaction from tangible labor outcomes far surpasses wandering in the digital abstract world.

A senior Salesforce product manager earning $180,000 a year might find a new job at a “California desalination plant” to cope with a 25-year drought. These infrastructure projects not only need construction but also must meet the highest standards and provide long-term maintenance. If we are willing, the “Jevons Paradox” applies equally to the physical world.

From Crisis Management to Societal Prosperity

The ultimate goal of large-scale industrial projects is shared prosperity. The U.S. will regain self-sufficiency and undertake large-scale, low-cost manufacturing. Moving beyond material shortages is key: in the long run, if we truly lose most office jobs due to AI, we must be capable of providing high-quality living standards for the population.

Since AI will drive profit margins toward zero, consumer goods will become extremely cheap, and this goal will be automatically achieved.

My view is that different sectors of the economy will “take off” at different speeds, and almost all industry transformations will be slower than prophets predict. This is not a denial of AI’s power—I am very optimistic about AI and expect my own work to eventually become obsolete. But it takes time, and time gives us the opportunity to develop good strategies.

At this stage, preventing a market collapse like Citrini7’s imagined scenario is actually not difficult. The U.S. government’s response during the pandemic demonstrated its proactive and decisive crisis management. If needed, large-scale stimulus measures can be quickly deployed. Though I admit they are inefficient, that’s not the point.

The key is to ensure the material well-being of the population—universal happiness, legitimacy for the state, and social contract—rather than obsessing over past accounting figures or economic doctrines.

Institutional inertia is not a bad thing—it’s a source of social stability. As long as we stay alert and adaptable during this slow but certain technological change, we will ultimately get through it safely. The crucial understanding is: change will come, but at a pace we can handle.