Deconstructing Bullshit Jobs: The 'Zero-Effect Default' Null Hypothesis and Why 90% of White-Collar Work Can Be Replaced by AI

For decades, modern society has accepted without question the premise that "working diligently has intrinsic value" and that "preserving employment is the supreme priority of public policy."

We must challenge this premise from the ground up. The mere act of working has no intrinsic value. A job should be evaluated solely by what it actually produces and what tangible outcome (differential impact) it delivers to society.

When we pursue this principle to its logical conclusion, it goes far beyond a superficial complaint that "there are too many pointless jobs." It reveals a structural framework that overturns our institutions: reversing the burden of proof for administrative and white-collar work (the zero-effect null hypothesis), replacing 90% of white-collar tasks with AI, separating labor from reward across four distinct criteria, exposing how corporate overhead inflates consumer prices, questioning GDP worship and ownership-based control, and completely unbundling production from income distribution.

💡 Core Summary in 3 Minutes (5 Key Takeaways)

1. Reversing the Burden of Proof ("Zero Effect" as the Default Null Hypothesis): For indirect white-collar and administrative tasks whose value is not self-evident, zero effect is the default—just as with new drugs and public policies. Staying busy is only proof that a job exists, not that it produces results. Especially today, with white-collar work widely considered saturated, rigorous empirical verification of counterfactual impact (Δ = Y(with work) - Y(without work)) is urgent.

2. Society Would Run Smoothly Even If 90% of White-Collar Work Were Replaced by AI Today: This is not an attack on individual competence, but a question of structural necessity. There is no justification for employing massive numbers of humans for meetings about meetings, internal slide decks, repetitive reporting, and multi-layered approval rituals. With AI, we can discard the assumption of human execution and redesign workflows entirely.

3. Labor Activity Itself Has No Intrinsic Value and Commands Zero Respect: Among "Ability," "Labor Activity (hours/effort)," "Outcome," and "Status," only Outcome should determine reward. Diligently laboring on unverified, pointless tasks is not an object of respect. Someone who automates a process and eliminates their own job creates far greater social value than someone who works 100 hours producing reports nobody reads.

4. Corporations Pass the Cost of Feeding Redundant Staff onto Consumers: Even for commodity products requiring no complex R&D (such as beer or canned drinks), prices remain artificially high to fund prestigious downtown headquarters, bloated management layers, and legacy advertising. Companies exist to create value, not to sustain internal employment.

5. Separate "Eliminating Jobs" from "Impoverishing People": Reducing required human labor through AI is a triumph of technological progress. The real problem lies not in production, but in our broken distribution mechanism that ties income exclusively to employment. Instead of manufacturing bullshit jobs to distribute wages, we must redesign distribution itself.

1. Reversing the Burden of Proof: Why "Zero Effect" Must Be the Default Initial Hypothesis#

In conventional organizations, the mere existence of a position is routinely treated as proof of its necessity.

• Staff are assigned to a municipal department at city hall.
• An annual budget is allocated to the division every fiscal year.
• Employees work busily every day, processing forms and documents.
• Weekly status meetings and reports to superiors are conducted without fail.

Yet all of these are merely evidence that the job is being performed (activity), not evidence that the job is producing positive effects (outcomes).

Whether examining municipal bureaucrats or corporate white-collar workers, the exact same analytical standard applies. Even if an employee works tirelessly for a full year on an administrative task, we must ask a separate question: Compared to a counterfactual world where that task was completely eliminated, has the well-being of residents or the state of society actually improved?

We must strictly distinguish between two fundamentally different questions:

Because "activity volume" is easy to count, government agencies and corporations use it to claim accomplishment. Yet the only metric that justifies consuming societal resources is the counterfactual difference.

Causal Inference and the Null Hypothesis of Zero Effect#

This perspective aligns directly with causal inference and statistical hypothesis testing in data science.

Consider pharmaceutical regulation: when evaluating a candidate medicine, regulators never assume "this drug must work." The default baseline (null hypothesis) is that it has zero effect compared to a placebo (Δ = 0), and its efficacy is recognized only after rigorous randomized controlled trials prove otherwise. Public policy evaluation follows the exact same logic.

Indirect administrative and white-collar routines that produce no clear tangible outcome must be subjected to this exact scientific rigor. Especially today, as white-collar labor markets are widely recognized as saturated, the necessity of verifying whether a task produces any positive outcome compared to its total absence has never been higher.

If we define the true effect of a job or policy as Δ (Delta), what we need to verify is not whether the task was executed, but the difference between the outcome with the task, Y(with work), and the counterfactual outcome without it, Y(without work):

Δ = Y(with work) - Y(without work)

If the real-world outcome remains unchanged whether the task is performed or not, the marginal effect Δ of that job is zero. No matter how many thousands of documents are stamped or how many hours are logged, one cannot infer a positive effect (Δ > 0) from the mere fact that someone is working.

Of course, some public services and organizational functions have effects that are difficult to quantify perfectly. However, the fact that an effect is difficult to measure is never a valid reason to automatically assume the effect is positive.

Just as with new drugs and public policies, society must treat zero effect as the default initial hypothesis for all institutional work. Those who wish to justify the continuation of a position, regulation, or budget bear the burden of demonstrating reasonable evidence of positive net impact.

The Consequence for Bullshit Jobs#

Adopting "zero effect as the default" completely inverts how we frame the problem of bullshit jobs.

The question is no longer: *"What percentage of existing jobs are wasteful?"* Instead, the question becomes: *"Out of all existing jobs in government and business, for what percentage can we actually verify a positive net effect?"*

Especially in public institutions like city halls, where the sole purpose of administration is citizen welfare, evaluation must be based on net benefit to residents—not on how busy clerks look or how many procedural hurdles they administer. Reversing the burden of proof stops the intellectual laziness of treating unverified bureaucracy as "public service."

2. Society Would Run Smoothly Even If 90% of White-Collar Work Were Replaced by AI Today#

Applying this zero-effect baseline to private-sector office work leads to our strongest proposition:

"Even if over 90% of bullshit-job-ridden white-collar work were replaced by AI right now, the world would continue to function without the slightest problem."

This is not a personal critique of whether individual employees are smart or lazy. It is a structural question: Is there any actual need for humans to perform these tasks in these numbers?

Inside modern corporations and agencies, vast amounts of human time are consumed by self-referential activities:

• Meetings to prepare for meetings, and documents created solely to generate more documents
• **Repetitive information processing: internal memos (*Ringisho*), daily reports, and progress tracking**
• Multi-layered message relaying and approval stamp rallies across middle management
• Internal politics and coordination generated to preserve departmental turf rather than produce goods or customer value

While these tasks may feel mandatory within a department's local rules, from the perspective of society as a whole, it is entirely unclear what tangible value would be lost if they vanished tomorrow.

Discarding the Premise of Human Execution#

Now that AI can instantaneously handle information synthesis, drafting, analysis, and coordination, the goal is not merely to make existing white-collar routines 10% or 20% faster. Using AI to write pointless internal approval memos more quickly only accelerates bullshit work.

Instead, AI allows us to discard the premise that humans must staff these administrative layers at all, redesigning workflows around essential functions only.

The "90%" figure serves as a structural provocation against the status quo: there is zero empirical basis for treating current white-collar headcount as the number of people actually required to run society. When we compare the social benefit lost by deleting a bureaucratic workflow against the massive human and economic resources saved, the vast majority of administrative overhead proves redundant.

3. Labor Activity Has No Intrinsic Value: The Four Criteria of Reward#

Why do ineffective jobs persist and continue to command salaries? Because existing institutions conflate four distinct criteria for evaluating and rewarding human contribution:

While Outcome (3) should be the primary determinant of reward, real-world organizations routinely substitute Labor Activity (2) and Status (4) in its place.

Separating these four criteria yields a direct conclusion:

The mere fact that someone is working hard or logging long hours does not mean their work has any social value whatsoever. Furthermore, unverified, pointless labor activity is entirely unworthy of respect.

If an employee works until midnight every day producing hundreds of pages of internal slides that improve zero decisions, are read by no one, and have no impact on the final product, rewarding them for their "hard work" is irrational. Sanctifying useless busywork purely out of emotional sympathy for someone's sweat and late hours only protects institutional parasites.

Conversely, someone who solves a critical problem in ten minutes, or an engineer who writes a script or AI workflow to completely automate and eliminate their own job, has not produced less value because their labor time was short. Eliminating a job so that no human ever has to do it again is infinitely more valuable than spending 40 years diligently performing unnecessary work.

Shattering the Greatest Illusion: "If They Get Paid a Salary, Their Work Must Have Value"#

Even after following this logic, many readers still struggle to shake one intuitive feeling: *"Still, if those workers are getting paid a proper salary every month, doesn't that prove someone needs their work?"*

To break this spell, we must recognize that salaries are funded not only by (1) Productive Value (Δ > 0), but also by (2) Tollbooth Rents (erecting legal, licensing, or bureaucratic barriers across a public path and splitting the compulsory tolls extracted from captive users) and (3) Disguised Welfare (paying conditional basic income in exchange for eight hours of daily bureaucratic theater to prevent mass white-collar unemployment and political revolt).

At the macro level, commercial bank remittance and settlement can be completely replaced by Bitcoin and P2P protocols (as proven when Ireland's banks shut down for six months in 1970 without collapsing the economy, and by Thomas Philippon's 130-year study of financial intermediation rents), while 99% of municipal clerical work can be automated away as proven by Estonia's X-Road. Bankers, bureaucrats, and middlemen receive salaries not because their work is indispensable, but because regulations and disconnected systems are deliberately preserved to justify their paychecks.

👉 Read the Full Empirical Analysis: *Why Do Valueless Jobs Command Salaries? Shattering the "Salary = Proof of Value" Illusion and the Empirical Evidence That Society Runs Without Banks, Bureaucrats, and Middlemen*

4. How Corporations Manufacture Redundant Work and Pass the Cost to Consumers#

One might assume that private corporations, driven by profit, naturally eliminate bullshit jobs unlike government bureaucracies. Yet examining large corporations and oligopolistic markets reveals the same institutional pathology.

Consider everyday commodity products such as beer or canned beverages, where production technology is mature and requires no frontier scientific R&D. In a rational structure, a manufacturer should concentrate resources strictly on production, logistics, and distribution, driving consumer prices down.

Why, then, do major manufacturers maintain high, nearly identical prices? Because they maintain massive headquarters in prime metropolitan real estate, employ thousands of white-collar staff for internal coordination and management layers, and spend fortunes on legacy mass-media advertising. The cost of feeding this bloated organizational apparatus is embedded directly into the price of every can sold.

Again, the issue is not individual laziness. When an organization makes "preserving headcount, departmental budgets, and executive posts" an end in itself, it rationally invents internal tasks to keep those employees occupied.

We must ask a fundamental question of order: Does a company exist to maintain employee headcount, or does an organization exist to produce value needed by society? Consumers have no obligation to subsidize corporate bullshit jobs through inflated product prices.

5. Does GDP Growth Mean Social Progress? Deconstructing Metrics and Capital Control#

Following this logic further challenges two macroeconomic sacred cows: GDP worship and control through capital ownership.

The Illusion of GDP as Progress#

We must never confuse changes in a measurement metric (GDP) with changes in real productive capacity or human flourishing.

Suppose AI and automation allow a service previously requiring 100 workers to be delivered at the exact same quality and volume by 50 workers:

• Because output is maintained with half the labor, real labor productivity doubles.
• If the freed-up time is returned to individuals as leisure, learning, or creative pursuits, human freedom and well-being genuinely increase.
• However, if reduced payrolls lower aggregate wage spending under legacy rules, nominal GDP might stagnate or contract.

Conversely, if society multiplies parasitic middlemen, complex regulatory compliance rituals, and corporate bullshit jobs—generating billions in consulting fees and administrative transactions—nominal GDP swells without adding a single unit of real utility to human life. Treating GDP growth as synonymous with social progress incentivizes the state to manufacture pointless work just to keep money circulating.

Shareholders, Employees, and Ownership as Control#

Rejecting the myth that "labor activity has intrinsic value" also requires scrutinizing the opposite side of the coin: capital and ownership.

Voluntary work and voluntary exchange of money are fine. The structural problem arises when holding vast financial assets or organizational status grants indirect command over other people's time and behavior—or when inherited wealth placed passively into index funds compounds automatically without the holder contributing any new ingenuity or value to society.

While capital returns partly reflect risk-bearing and liquidity provision, we must recognize a parallel truth: just as logging labor hours is not synonymous with creating value, merely holding title to capital is not synonymous with creating new social value.

6. The Core Problem Is Distribution, Not Production: Three Propositions#

We can synthesize this entire analysis into a single structural progression:

When AI replaces human labor, society experiences two distinct shifts:

• 1. The Production Shift (Technological Triumph): Fewer humans and fewer hours are needed to produce the goods and services required for a high standard of living.
• 2. The Distribution Shift (Institutional Bug): Because income is currently distributed primarily through wages and employment, people whose jobs are automated lose their claim on the wealth produced by AI and capital.

Losing a job and losing the ability to live are two entirely different problems.

A reduction in total required human labor is a victory for civilization, not a tragedy. The tragedy lies entirely in our legacy distribution bug, which forces society to bundle "production" and "income distribution" into a single mechanism called "employment."

When we cling to employment as the sole pipeline for income in an age of AI abundance, technological progress becomes a threat rather than a liberation. To keep citizens from starving, the state and corporations are forced to invent and subsidize zero-effect bullshit jobs, chaining human beings to desks for eight hours a day just to justify handing them a paycheck.

The Three Core Propositions#

1. The Descriptive Proposition (Zero-Effect Default): All institutional work should be evaluated against a default null hypothesis of zero effect (Δ = 0). A vast share of administrative and white-collar work exists only to sustain the organization itself and can be replaced or eliminated via AI without loss of social utility.
2. The Evaluative Proposition (Outcome Over Effort): Reward and evaluation should reflect tangible outcomes and social benefit, not hours worked or hierarchical status. Automating and abolishing a task is superior to diligently performing unnecessary labor.
3. The Institutional Proposition (Unbundling Production from Distribution): In a society where AI and automation drastically reduce the labor required for production, we must stop preserving pointless employment as a disguised welfare program and instead redesign how wealth and income are distributed.

Conclusion: Destroying Bullshit Jobs and Building a Selectable Distribution Protocol#

Our target for destruction is never human beings; it is the institutional machinery that forces valueless work to persist and chains human lives to it.

The fact that a bureaucracy or corporate division has existed for fifty years, or employs ten thousand people, proves nothing about its effectiveness.

• Abolish documents created for the sake of documents, audits that exist to sustain auditors, projects designed solely to capture subsidies, and meetings held to prove an organization exists.
• Enforce a reversed burden of proof (Δ = Y(with work) - Y(without work)) on all public and organizational functions.
• Completely decouple the elimination of unnecessary labor from human economic survival.

This is precisely why UNIVGOV (Universe Government) explicitly champions the *"Destruction of Bullshit Jobs"* in our Hundred-Year Thesis, unbundles governance into verifiable modules, and experiments with an unconditional initial grant of 1,000,000 UNI to every registered citizen alongside AI-driven policy audits.

Preserving vested interests and administrative busywork is neither "employment policy" nor "economic growth." By treating zero effect as the default, handing redundant information processing over to AI, and separating income distribution from forced organizational servitude, we reclaim human time for genuine creation, inquiry, and freedom.