a16z合伙人驳斥AI末日论，：别慌，技术变革会做大蛋糕

Original Author: David George

Translated and Compiled by: Felix, PANews

Editor's Note: The current AI "doomsday narrative" seems to be dominating public discourse, with fears of "AI taking jobs" and "unemployment" spreading globally, prompting people from all walks of life to devise solutions for the disruptive changes AI is about to bring. However, a16z general partner David George argues that the "doomsday" perspective is completely unfounded, lacking evidence, imagination, and an understanding of humanity. The following is the full text.

The "permanent underclass" argument put forward by AI alarmists is simply not convincing. This is nothing new; it's just the "lump of labor fallacy" wrapped in a new package.

The "lump of labor fallacy" claims that the total amount of work needed in the world is fixed. It assumes a zero-sum game between existing workers and any person or thing (be it other workers, machines, or now AI) that could potentially do the same job. If the total amount of useful work to be done is fixed, then if AI does more, humans must necessarily do less.

The problem with this premise is that it contradicts everything we know about people, markets, and economics. Human needs and desires are by no means fixed. Nearly a century ago, Keynes predicted automation would lead to a 15-hour work week, but history proved him wrong. He was correct about the "technological unemployment" caused by automation, but instead of resting on our laurels, we found new, different productive activities to fill our time.

Of course, AI will absolutely eliminate some jobs and compress certain roles (and there's evidence this may already be happening). The landscape of the labor market will change, as it inevitably does with every transformative technology. However, the notion that AI will lead to economy-wide, permanent unemployment is bad marketing hype, bad economics, and a profound ignorance of history. On the contrary, productivity increases should boost the demand for labor, as labor becomes more valuable.

Here are our reasons.

"Humanity is Doomed?" Don't Be Ridiculous

We agree with the doomsayers on one point: the cost of cognition is plummeting. AI is becoming increasingly adept at tasks that, until recently, were considered the exclusive domain of the human brain.

The doomsayer argument goes: "If AI can think for us, then humanity's 'moat' disappears, and our ultimate value falls to zero." Humanity is finished. Apparently, we have already done all the thinking that needs or wants to be done, and now AI will take on an increasing cognitive load, gradually leading to human obsolescence.

However, the reality is this: precedent (and intuition) show that when the cost of a powerful input falls, the economy does not stagnate. Costs drop, quality improves, speed increases, new products become viable, and demand expands outward. Jevons paradox strikes again. When fossil fuels first made energy cheap and abundant, we didn't just put whalers and woodcutters out of work; we invented plastics.

Contrary to the doomsayers' views, we have every reason to expect a similar impact from AI. Since AI will bear an increasing cognitive load, humans will be freed to explore new, grander frontiers than ever before.

History shows that technological change inevitably expands the economic pie.

Each "dominant economic sector" was replaced by a larger successor... which in turn further expanded the economy.

Today's technology sector is vastly larger than finance, railroads, or industry, yet its share of the overall economy or market is still relatively small. Productivity enhancement is far from a negative-sum game; it is a powerful positive-sum force. Delegating so much work to machines ultimately results in a larger, more diverse, and more complex economy and labor market.

Doomsayers want you to ignore the history of innovation, focusing solely on the drastic decline in cognitive costs as the whole truth. They see task substitution and stop thinking.

"We'll increase cognitive output tenfold, but we won't engage in more thinking; we'll just pat our bellies, have an early lunch, and everyone else will do the same." This claim not only reflects a severe lack of imagination but also a failure to observe basic facts. Doomsayers call it "realism," but it's simply not going to happen.

The Failure of Luddism

(PANews Note: Luddism refers to a social movement in early 19th century Britain where workers, opposing the Industrial Revolution, destroyed industrial machinery to protest worsening working conditions and unemployment.)

Let's look at what actually happens when a massive leap in productivity sweeps through the economy.

Agriculture

At the beginning of the 20th century, before the widespread adoption of agricultural mechanization, about one-third of the U.S. workforce was employed in agriculture. By 2017, this figure had dropped to about 2%.

If automation caused permanent unemployment, the tractor should have completely devastated the labor market. It didn't. Agricultural output nearly tripled, supporting a massive population increase, and those displaced workers, far from being permanently unemployed, flowed into previously unimaginable industries, factories, shops, offices, hospitals, laboratories, and eventually into the service and software sectors.

So, while you can argue that technology disrupted the career prospects of the average farm worker, it simultaneously released a surplus of global labor (and resources) and gave birth to an entirely new economic system.

Electrification

The story of electricity is similar.

Electrification wasn't just swapping one energy source for another. It replaced drive shafts and belts with individual electric motors, forcing factories to reorganize around entirely new workflows, and creating entirely new categories of consumer and industrial goods.

This is precisely what we expect at different stages of a technological revolution, as documented by Carlota Perez in "Technological Revolutions and Financial Capital": massive upfront investment and financial interests, a sharp decline in the cost of durable goods, and subsequent generational prosperity for the manufacturers of those goods.

The productivity advantages of electricity didn't materialize overnight. In the early 20th century, only 5% of U.S. factories used electric power to drive machinery, and fewer than 10% of homes were electrified.

By 1930, electricity supplied nearly 80% of manufacturing power, and labor productivity doubled over the following decades.

Far from diminishing the demand for labor, productivity gains led to more manufacturing, more salespeople, more credit, and more business activity, not to mention the ripple effects from labor-saving devices like washing machines and cars. These devices allowed more people to engage in high-value work previously inaccessible to them.

As car prices fell, both car production and employment exploded.

This is what genuine General Purpose Technologies (GPTs) do: they reorganize the economy and expand the boundaries of useful work.

We see this time and again. Did VisiCalc and Excel end the careers of bookkeepers? Absolutely not. Instead, vastly more efficient computing technology led to a surge in the number of bookkeepers and spawned the entire Financial Planning & Analysis (FP&A) industry.

We lost about 1 million "bookkeepers," but gained about 1.5 million "financial analysts."

New Service Sector Jobs

Of course, job displacement doesn't always drive employment growth in related economic sectors. Sometimes, productivity gains translate into new jobs in entirely unrelated industries.

But what if AI means some people become incredibly wealthy while others are left behind?

At the very least, those super-rich will have to spend their money somewhere, just as they have in the past, creating entirely new service industries from scratch:

Massive productivity gains and the resulting wealth creation gave rise to entirely new fields of work that might never have emerged without increased income and labor supply (even though these fields were technically feasible long before the 1990s). Whatever one thinks of service industries catering to the wealthy, the end result benefits everyone because increased demand leads to a significant rise in median wages (thus creating more "wealthy" individuals).

Ernie Tedeschi, an in-house economist at Stripe, provides a comprehensive case study of how technology disrupted, transformed, and reshaped the travel agent profession.

Did technology reduce demand for travel agents? Yes.

Today, the number of travel agent employees is about half of what it was around 2000, almost certainly due to technological advancements.

So, does this mean technology killed jobs? No, because travel agents didn't become permanently unemployed. They found work in other sectors of the economy, and the overall employment-to-population ratio for the economy today is roughly the same as it was in 2000 (adjusted for an aging population).

At the same time, for those who remained in the now tech-enabled travel agent industry, higher productivity means higher wages than ever before:

"In their peak year 2000, travel agents earned on average 87% of the overall average weekly wage. By 2025, that figure stands at 99%, meaning travel agents' wages have grown faster than the rest of the private sector over this period."

Therefore, even though technology significantly impacted travel agent employment, the overall employment rate for the working-age population remained stable, and the remaining travel agents are better off than ever before.

Augmentation > Replacement (And Jobs Yet to Exist)

This last point is crucial and shows, once again, that doomsayers are only telling a small part of the story.

For some jobs, AI poses an existential threat. True enough. But for other jobs, AI acts as a multiplier: making those jobs more valuable. For every job at risk of replacement by AI, other jobs are poised to benefit:

Goldman Sachs' estimated "AI replacement" effect is far outweighed by the "AI augmentation" effect.

It's also worth noting that management teams seem to focus more on augmentation than replacement:

To date, mentions of "AI as an enabler/feature" on earnings calls outnumber mentions of "AI as a replacement" by roughly 8 to 1.

Although Goldman Sachs doesn't even list software engineers on its "augmented" list, they are perhaps the best example of AI-augmented talent.

AI is a multiplier for coding. Not only have git pushes surged (along with the creation of new applications and businesses), but the demand for software engineers also appears to be rising:

Since the beginning of 2025, software development job openings (both in number and as a percentage of the overall job market) have been steadily increasing.

Is this related to AI? Frankly, it might be too early to tell, but AI undoubtedly enhances productivity in software engineering, not to mention that AI has become a focal point for executives at every company.

Given everyone's efforts to figure out how to integrate AI into their businesses, it's not surprising that companies are hiring aggressively, which undoubtedly enhances the value of some employees rather than diminishing it.

The adoption of AI seems to be driving above-average wage growth (especially in systems design).

These gains might be limited for now, but it's still early days. As expertise expands, opportunities will grow. Regardless, this isn't the data doomsayers want you to see.

Meanwhile, according to Lenny Rachitsky (founder of Lenny's Newsletter, a platform for the tech community), the number of open project manager positions continues to climb (after dipping significantly due to interest rate fluctuations), now higher than at any point since 2022:

The growth in hiring for both software engineers and product managers strongly validates the fallacy of the "lump of labor" argument. If AI completely replaced human thinking, you might expect "engineers need fewer product managers," or vice versa, but that's not happening. We're seeing a sustained rebound in demand for both types of talent because the key is that people are becoming more productive.

This is precisely why the doomsayers' claims are fundamentally lacking in imagination. They focus only on the jobs that will be automated away, ignoring the demand areas that will create entirely new jobs we haven't even conceived of yet:

The majority of jobs added since 1940 didn't even exist in 1940. By 2000, it was easy to imagine travel agents losing their jobs, but much harder to envision a mid-market tech services industry built around "cloud migration," as cloud adoption was at least a decade away.

What Does the Current Situation Look Like?

So far, we've focused on theory and precedent, both of which support the optimistic view:

That's right. Every productivity increase has led to demand growth or the reallocation of surplus resources to other parts of the economy. This means more jobs, many of which are significantly more valuable, and some of which are unprecedented. If this time is different, the doomsayers need to provide much stronger evidence than just empty rhetoric.

The idea that "job displacement" is an end to civilization (when in fact it's the opposite) makes a lot of sense. Human nature is to be restless. We finish one task and look for another.

But, setting aside theory and precedent, what do the actual data say about AI and jobs? While it's still early days (for better or worse), existing data doesn't support the doomsayer's view. If anything, it points to "no significant change," but there is emerging data pointing in the opposite direction: AI creating more jobs than it displaces.

First, let's look at some academic research. This isn't an exhaustive literature review, just a few examples of recent papers:

• "AI, Productivity, and the Labor Force: Evidence from Business Leaders" (NBER Working Paper 34984): "In sum, these results suggest that while AI adoption has not yet led to significant changes in total employment, it has begun to reshape the allocation of tasks and occupations within firms. Specifically, routine clerical and administrative activities appear more susceptible to replacement, while analytical, technical, and managerial tasks are more often described as being complemented by AI."