How the task-versus-job defence stops working
The old rebuttal was always the same. AI can do tasks, but tasks are not jobs. Jobs are bundles of judgement, relationships, context, and physical-world friction. AI can write a paragraph or fill a spreadsheet, but it cannot run a legal practice, manage a team, close a deal, or sit in a meeting. So jobs, the argument went, were safe.
This was the most respectable version of the continuity story. Autor wrote it. Brynjolfsson endorsed it. Mollick turned it into a business-school cottage industry. It held up for a while because it was true in the same way a gate is true. It was true until enough of it came off the hinges.
Version 3.3 of the Discontinuity Thesis closes that gate.
The thesis does not need AI to replace whole occupations in one step. It never did. It needs only enough economically valuable task-units to cross the cost and quality threshold that firms start rebuilding workflows around machines instead of people. A job is not a sacred object. It is a bundle of tasks. Once the bundle becomes decomposable, the worker becomes decomposable with it.
What the evidence now says
Version 3.2 argued that unit-cost dominance would eventually cross at the task level. That argument is no longer prospective. OpenAI reports GPT-5.2 Thinking beating or tying top industry professionals on 70.9 percent of GDPval comparisons, which is not a trivia benchmark but a test of real professional deliverables across 44 occupations. GPT-5.4 pushed that to 83 percent, cleared 75 percent on OSWorld-Verified where the previous model hit 47, and reached 87 percent on internal junior investment-banking modelling tasks. GPT-5.5 continues the trajectory at 84.9 percent on GDPval and 78.7 percent on OSWorld-Verified, with OpenAI describing it as able to research, analyse, build documents, operate software, move across tools, check its work, and push through messy multi-part tasks.
This does not mean the labour market has collapsed. It means the lower-layer crossover is no longer speculative. The only variable left in the argument is how fast that lower-layer crossover propagates upward.
The new clause: interface collapse
The most important addition in v3.3 is interface collapse.
A lot of white-collar labour was historically protected by software fragmentation. The worker did not only think. The worker moved. They moved between the inbox and the spreadsheet, the CRM and the browser, the calendar and the document editor, the dashboard and the ticketing system, the codebase and the internal wiki. Humans survived because they were the integration layer between systems that did not talk to each other cleanly.
That was a real moat. It is now collapsing.
Once an AI can see screens, click, type, browse, call tools, manipulate files, and keep going across a multi-step workflow, the boundary between producing output and doing work starts to dissolve. This is what the GPT-5.4 and GPT-5.5 benchmarks on OSWorld and BrowseComp actually mean. The model is no longer a text generator on the other side of a keyboard. It is an interface operator.
The old moat was fragmentation. The new capability is integration. The first thing it dissolves is the layer of work that looked like thinking but was actually stitching.
The four layers
The cleanest way to think about this is as a cascade, not a single event.
The first layer is task-level unit-cost dominance. AI plus thin human oversight produces professional task outputs at equal or better quality, faster speed, and lower marginal cost. This has already crossed for a large and growing set of well-specified cognitive deliverables.
The second layer is interface and workflow dominance. AI operates through the same software environments where work happens, and handles the stitching between them. This is rapidly crossing, and the public benchmarks now track it directly.
The third layer is job-level dominance. Whole roles become economically unnecessary as enough human task volume is stripped out. This is partial and uneven, and the thesis does not require it.
The fourth layer is labour-market dominance, which is where the thesis actually lives. Wage labour stops being the mass route to economic agency. This has not fully arrived, but the pathway from the first two layers to the fourth is now specified and visible.
Critics prefer to argue at the third layer because residual human tasks are easy to find inside existing jobs. But that is not where the thesis is defended, and never was.
No scream, just non-absorption
Mass displacement will not first look like mass unemployment. It will look like non-absorption.
Fewer entry-level roles. Fewer junior ladders. Fewer graduate pathways. Fewer promotions. More contractors. More review roles. Productivity gains that do not reach wages. Incumbents held in place while new entrants fail to launch. A headline unemployment rate that looks fine while the system underneath it has stopped reproducing itself.
The tells to monitor are specific. Entry-level hiring decline in AI-exposed fields. Junior-to-senior ratio compression. Weak graduate absorption despite stable aggregate employment. Wage stagnation or wage compression in exposed cognitive work. Productivity gains not passed to labour. Contractorisation and project-based substitution. Delayed retirement of incumbents combined with fewer new entrants. Expansion of review and validation roles that do not scale into careers. Collapse of training ladders because AI now does the junior work from which expertise used to grow. Rising dependence on capital income, transfers, rents, or platform ownership rather than wages.
Stanford’s Digital Economy Lab has already reported the first expected signal. Early-career workers aged 22 to 25 in the most AI-exposed occupations have seen a 16 percent relative employment decline after controlling for firm-level shocks, while less-exposed fields and more experienced workers remained stable or grew. That is not the end state. It is the first tell, and it is arriving on schedule.
Aggregate calm is not a refutation of the thesis if these indicators are moving in the predicted direction. That is the whole point of calling it the No-Scream Principle.
The augmentation argument is a stage error
The standard reply to all of this is that humans will work with AI. That sentence is doing more work than it can carry.
There are three kinds of complementarity, and they are not the same thing.
Genuine complementarity means AI raises the marginal value of human labour. The worker becomes more productive, more valuable, and captures some of the gain through wages, bargaining power, or advancement. This is real, and it exists. It is not guaranteed to persist.
Transitional complementarity means humans supervise, correct, validate, integrate, and absorb responsibility while the AI improves. This is the phase most commonly mistaken for the future of work. It is unstable by design. The human role gets thinner as the system gets better.
Theatrical complementarity means humans remain for trust, liability, regulation, customer comfort, ritual legitimacy, or institutional optics. The human is still in the room, but no longer economically central. This is not augmentation. It is managed displacement wearing augmentation’s clothes.
The augmentation narrative points at the second stage and calls it the destination. The thesis argues it is the corridor. Once that corridor is named, “humans will work with AI” stops being a rebuttal and becomes a question about which stage you mean and why that stage is stable under competitive pressure. Nobody making the argument ever answers that question, because the honest answer is that it isn’t.
Regulation defines categories; economics decides which ones survive
Some argue regulation can preserve human-only work zones. This misunderstands the problem.
Regulation can define categories. It can create compliance duties, liability regimes, procurement rules, audit trails, disclosure requirements, and sectoral restrictions. The EU AI Act does this competently. Those boundaries can exist in statute. The question has never been whether law can define categories. The question is whether those categories remain labour-preserving once the economic gradient points toward automation.
AI erodes task boundaries continuously. Spell-check becomes drafting. Drafting becomes composition. Composition becomes analysis. Analysis becomes decision-making. Decision-making becomes execution. Each permitted augmentation pathway becomes a staging ground for the next substitution step. There is no stable line between assisted and automated when the same system can assist, draft, decide, execute, and verify.
Coordination therefore fails on two axes at once. Actors who keep expensive human labour lose to actors who automate. And the task boundaries that regulators try to defend blur faster than regulators can redefine them. The prisoner’s dilemma runs at quarterly earnings pace. Treaty negotiation runs at OECD summit pace. Deployment outruns coordination. This is not a claim that redistribution is impossible. It is a claim about timescale.
Why redistribution does not save the system
Capital redistribution may be necessary. It may be humane. It may be the only politically viable way to preserve mass consumption after wage labour weakens. But it does not preserve postwar capitalism.
A National AI Dividend Fund, a sovereign compute fund, a UBI, or a public equity scheme all preserve consumption while conceding production. Recipients may eat, vote, and live better than the unemployed did in the old failure mode. But they no longer hold the structural leverage that came from being economically necessary. They have citizenship without productive agency.
This is the distinction closed in v3.2 and worth restating. Consumption continuity is not system continuity. Dividends are a successor system, not a rescue. The dispute about whether that successor system is good or bad is a different dispute. It is not the dispute about whether postwar capitalism survives. That one is already over.
The right threshold
Version 3.2 contained a rhetorical line that needs to come out. It described the time of death as the moment the last human worker becomes more expensive than AI. That threshold is unnecessary and invites the residual-work dodge. There will always be some human work. Kings had servants. Aristocracies had artisans. Feudal societies had labour. The existence of residual human work does not preserve a system built on mass productive participation.
The correct threshold is not the last worker. It is the majority.
Postwar capitalism dies when wage labour no longer provides mass economic agency, because the majority of working-age adults cannot sell labour at socially sustaining wages without subsidy, protection, artificial scarcity, makework, or political intervention. Not total labour extinction. Mass agency collapse. Not consumption. Productive participation.
What critics now have to argue
The thesis has narrowed the defensible counterposition to almost nothing. The claim that AI cannot do economically valuable cognitive work is over. The claim that tasks are not jobs no longer saves continuity because capitalism unbundles jobs the moment task-level arbitrage is profitable. The claim that humans will work with AI has become an unanswered question about which complementarity stage is meant and why it lasts. The claim that regulation will hold the line has become a question about economic durability rather than legal definition. The claim that aggregate labour markets look fine has to confront the No-Scream indicators rather than the unemployment rate.
What is left for critics is a specific, committed claim about propagation speed. They need to explain why task-level dominance will not propagate into workflow recomposition. Why workflow recomposition will not suppress hiring. Why hiring suppression will not break training ladders. Why broken training ladders will not compress wages and bargaining power. Why all of that will not eventually reach the threshold where the majority cannot sell labour at socially sustaining wages. “Not yet” is not an answer to any of those questions. It is the interval between mechanism and manifestation.
The discontinuity is not caused by AI replacing every job. It is caused by AI replacing the economic necessity of mass human labour. That is the thing postwar capitalism cannot survive, and the last moat protecting it was the interface.
The corpse does not need to be cold for the wound to be fatal.