The Great Unraveling, How AI Is Breaking the Career Ladder and Reshaping the Future of Work
There is a moment, just before a cataclysm, when the world holds its breath. In February 2020, as the first whispers of a novel coronavirus began to circulate, there was an eerie calm—a brief pause before a global upheaval that would forever alter how we live, work, and connect. Matt Shumaker, CEO of HyperWrite, has drawn a chilling parallel to that moment. In his viral post, “Something Big Is Happening,” he describes a quiet revolution unfolding not in the headlines, but in the trenches of the tech industry. And like the pandemic, its consequences will be inescapable.
Shumaker’s revelation is deceptively simple. He assigned a complex coding task to an AI agent, GPT-3.5 Codex, and left it to work. When he returned four hours later, the work was not just finished. It was executed with “taste” and “judgment”—qualities we have, for centuries, reserved for humans. This is the qualitative leap that has sent shockwaves through boardrooms and break rooms alike. AI has crossed the line from “helpful tool” to “independent doer.” Coding, legal drafting, financial analysis—execution that once took days now takes minutes. The implications for the traditional white-collar worker are terrifying.
As analyst Jaspreet Bindra explores in his incisive commentary, we are witnessing not just the automation of jobs, but the breaking of the very mechanism by which people get them. The bottom rung of the career ladder is not just wobbling; it is being pulled away. And while optimists and sceptics debate the timeline, the direction of travel is unmistakable: the world of work will never be the same.
The “Taste” and “Judgment” Threshold
For years, the dominant narrative around AI and employment was one of augmentation. Machines would handle the dull, dirty, and dangerous tasks, freeing humans to focus on higher-order thinking, creativity, and emotional intelligence. AI was a tool, like a calculator or a word processor, that amplified human capability. That narrative is now obsolete.
What shook Shumaker—and what has resonated with millions who shared his post—was not that the AI completed the task, but how it completed it. The code was not merely functional; it was elegant. It demonstrated an understanding of context, an appreciation for efficiency, and a sense of design that went beyond the literal instructions. It exhibited “taste” and “judgment.”
These are the very qualities that have long been considered the sanctuary of the human mind. They are what differentiate a skilled craftsman from a novice, a seasoned lawyer from a paralegal, a senior architect from a draftsman. If AI can now exercise judgment in execution, then the boundary between “human work” and “machine work” has been irrevocably blurred. The question is no longer “Can a machine do this task?” but “Why would a human be paid to do this task?”
The Breaking of the Bottom Rung
The most acute impact of this shift will be felt by those at the beginning of their careers. Anthropic founder Dario Amodei has issued a stark prediction: AI could eliminate half of all entry-level positions within five years. Aneesh Raman, a senior leader at LinkedIn, describes the phenomenon even more vividly: “the bottom rung of the job ladder [is] breaking.”
Consider the traditional career trajectory. A young graduate joins a law firm, an accounting practice, or a software company. They spend their first few years doing the “grunt work”—document review, basic coding, data entry, legal research. This work is not glamorous, but it is essential. It is how novices learn the ropes, develop judgment, and prove their worth. It is the apprenticeship model that has underpinned professional development for centuries.
What happens when that grunt work can be done by an AI in minutes, with perfect accuracy and at near-zero marginal cost? The answer is that the entry-level job disappears. The bottom rung is removed from the ladder. A young person entering the workforce is no longer asked to start at the bottom and climb; they are asked to start at a level that requires experience they do not have, competing against AI systems that need no training, no salary, and no sleep.
This is not just a problem for the individuals who cannot find work. It is a systemic crisis for the entire economy. If the pipeline for developing mid-career and senior professionals is cut off at the source, we face a future of severe skills shortages at the top, even as legions of young people are locked out of the workforce. The career ladder becomes a cliff face.
The Thinking-Doing Dichotomy
Shumaker’s warning points to a fundamental revaluation of human labour. “If your value lies in the ‘doing’ rather than the thinking/creativity/dedication-making,” he argues, “you are increasingly redundant.”
This creates a new, unforgiving hierarchy. At the top are the “thinkers”—the strategists, the visionaries, the entrepreneurs, the artists who conceive of what needs to be done. Their work is about defining problems, imagining possibilities, and making high-stakes decisions. Below them are the “doers”—the coders who translate specifications into software, the lawyers who draft contracts based on precedents, the analysts who crunch numbers to generate reports. It is this vast middle layer of “doing” that is now in the crosshairs.
The “doers” are not low-skilled workers. They are college graduates, professionals with years of training. They are the backbone of the modern white-collar economy. And they are the ones whose jobs are most at risk. An AI can now draft a contract, write code, generate a financial model, and even produce marketing copy. The “doing” is becoming a commodity.
This does not mean that all “doing” jobs will vanish overnight. Many will be transformed. A lawyer may no longer draft documents from scratch but will instead become an editor of AI-generated drafts, a role that requires a different, perhaps higher, level of judgment. A coder may no longer write every line of code but will instead architect systems and debug AI outputs. The nature of the work shifts from production to supervision, from creation to curation.
But this transformation carries its own dangers. It raises the bar for entry. To be an effective editor of AI output, one must already understand what good output looks like. How does a young person gain that understanding if they never did the foundational work themselves? The apprentice becomes obsolete before they can learn the trade.
The Skeptics’ View: Bubble or Revolution?
For every prophet of the AI apocalypse, there is a sceptic. Gary Marcus, a prominent AI researcher and entrepreneur, argues that we are witnessing a massive “AI bubble.” In his view, the current wave of generative AI, while impressive, is overhyped and fundamentally unstable. These systems, he points out, lack true understanding, common sense, and reliability. They can produce convincing output, but they also “hallucinate,” generate nonsense, and fail in unpredictable ways.
Marcus and other sceptics contend that the bubble will eventually burst, or at least deflate, as the limitations of current technology become apparent. They argue that the path to true artificial general intelligence (AGI) is much longer and harder than the optimists believe, and that in the meantime, humans will remain indispensable for a wide range of tasks that require genuine comprehension and adaptability.
This is a comforting view, and it may be partially correct. It is certainly true that current AI systems have significant limitations. They do not “understand” in the human sense. They are pattern-matching engines trained on vast datasets, not conscious beings with intentions and beliefs.
However, the sceptics may be missing the point. The threshold for economic disruption is not AGI; it is “good enough.” An AI that can do a competent job 80% of the time, at a fraction of the cost of a human, is economically irresistible to employers. The fact that it occasionally makes mistakes is not a barrier to adoption; it is a problem to be managed, perhaps by a smaller team of human supervisors. The bar for “good enough” is constantly being raised, but the direction is clear: AI capabilities are improving, while costs are falling.
The Policy Vacuum: Preparing for a World Without Ladders
Governments and educational institutions are lagging far behind the curve. The current model of education, designed for the industrial age, assumes a linear progression: learn a set of skills, then apply them in a career for four decades. That model is collapsing.
If entry-level jobs are disappearing, what is the alternative? One answer is a massive expansion of apprenticeship and mentorship programs that integrate work and learning from day one. Another is a fundamental redesign of university curricula to focus less on transmitting static knowledge (which AI can now access and synthesize instantly) and more on developing the uniquely human skills that AI cannot replicate: critical thinking, ethical reasoning, creativity, emotional intelligence, and adaptability.
There is also a growing conversation about broader social interventions, such as universal basic income or portable benefits, to cushion the dislocation caused by rapid technological change. If the link between “having a job” and “having a livelihood” is severed for a significant portion of the population, the social contract will need to be renegotiated.
None of these solutions are easy, and none are being implemented at the scale required. The AI revolution is moving at the speed of software, while policy and education move at the speed of bureaucracy. The gap between them is where the crisis will unfold.
Conclusion: The Calm Before the Storm
Matt Shumaker’s comparison to February 2020 is apt. There is a palpable sense of an impending upheaval, a feeling that the ground is shifting beneath our feet. The difference is that this time, there is no vaccine on the horizon. There is no “return to normal.” The change is structural, permanent, and accelerating.
The bottom rung of the job ladder is breaking. The “doing” layer of the workforce is being hollowed out. The qualities of “taste” and “judgment” are no longer exclusively human. We are entering a world where the relationship between human beings and work must be fundamentally reimagined.
The eerie calm will not last. The upheaval is coming. The only question is whether we will be ready.
Q&A: Unpacking the AI Employment Crisis
Q1: The article mentions AI exhibiting “taste” and “judgment.” What does that mean in practical terms?
A: In practical terms, it means the AI’s output is not just technically correct, but contextually appropriate, aesthetically pleasing, and efficient. For example, in coding, a human programmer might write code that works but is messy and difficult for others to maintain. An AI with “taste” writes code that is clean, well-commented, and follows best practices. In legal drafting, it means a contract that not only includes the necessary clauses but structures them in a logical, clear way that anticipates potential disputes. It is the difference between a functional but ugly piece of furniture and one that is both functional and beautifully crafted. This leap from mere correctness to quality is what makes the current wave of AI so disruptive.
Q2: Why is the impact on entry-level jobs considered so critical?
A: Entry-level jobs are the apprenticeship system of the modern economy. They are where young workers learn professional norms, develop practical skills, and build the judgment that comes from experience. If these jobs are automated away, there is no pipeline for developing the next generation of senior professionals. A lawyer cannot become a skilled litigator without first spending years doing research and drafting. A programmer cannot become a software architect without first writing thousands of lines of code. Removing the bottom rung of the ladder doesn’t just create unemployment for the young; it creates a future skills shortage for everyone. The entire system of professional development relies on the existence of these foundational roles.
Q3: The article distinguishes between “thinkers” and “doers.” Is this a new distinction?
A: The distinction itself is not new, but its consequences are becoming far more severe. In the past, the line between thinking and doing was blurred. A senior professional both conceived the plan and executed parts of it. The “doing” was where they refined their thinking. What AI does is to separate these functions more sharply. The “thinking” becomes about high-level strategy, problem definition, and oversight. The “doing” becomes a task that can be delegated to a machine. This creates a new class divide in the workforce between those who direct the machines and those who are replaced by them. The danger is that the “doing” layer, which has historically been a source of stable, middle-class employment, is systematically eliminated.
Q4: What do sceptics like Gary Marcus get wrong about this disruption?
A: Sceptics are correct that current AI systems have profound limitations—they lack true understanding, they make mistakes, and they can be brittle. However, their error may be in underestimating the economic logic of “good enough.” A system that is 80% as good as a human but costs 1% as much will be adopted, even if it requires some human oversight. The fact that AI is imperfect is not a barrier to adoption; it is a problem to be solved by reducing the number of humans needed to supervise it. The sceptics are fighting the last war, focusing on whether AI can match human intelligence, when the real question is whether it can replace human labour at a scale that reshapes the economy.
Q5: What can an individual worker do to prepare for this future?
A: The most important adaptation is to shift one’s value proposition from “doing” to “thinking” and “relating.” This means focusing on skills that AI cannot easily replicate: critical thinking, complex problem-solving, ethical judgment, emotional intelligence, leadership, and creativity. It also means becoming proficient in using AI as a tool. The workers who will thrive are not those who compete with AI, but those who can leverage AI to amplify their own capabilities. A lawyer who can use AI to draft documents in minutes and then spend their time on strategy and client relationships will be far more valuable than one who insists on doing the drafting manually. The future belongs to the “centaurs”—humans and AI working together, with the human providing the direction and the AI providing the execution.
