The Career Ladder Is Being Taken Down From Both Ends

The two ends of the corporate ladder are moving at the same time, and it is the same force doing both. At the bottom, the share of CEOs planning to cut junior roles jumped from 17% to 43% in a single year ^[1]. At the top, senior professionals are voluntarily leaving management to become individual contributors again, because one person armed with AI agents can now do work that used to require a whole team ^[2]. These get reported as two separate stories. They are one structure being dismantled from both ends. Start at the bottom. Entry-level postings in the US fell 35% between January 2023 and June 2025 ^[3]. Workers aged 22 to 25 in AI-exposed occupations saw a 13% relative employment decline while experienced workers were largely untouched ^[4]. Anthropic's Dario Amodei has warned that half of entry-level white-collar jobs could disappear within five years ^[5]. The grunt work that used to be the price of admission, the copying and drafting and first-pass analysis, is exactly what the models now do. Then the top. Manager engagement fell from 27% to 22% in a year, and the people leaving management are not being pushed out; they are choosing to go ^[2]. The term is "conscious unbossing." Robinhood's Vlad Tenev predicts a "job singularity" of single-person companies, and Alibaba's president describes one-person unicorns where competitive advantage shifts from headcount to judgment ^[6][7]. The extreme case is a dot: one person, a staff of agents, no one to manage. The middle is where the policy becomes explicit. Amazon's Andy Jassy mandated a 15% increase in the ratio of individual contributors to managers ^[8]. Block cut 4,000 jobs and published a thesis titled "From Hierarchy to Intelligence" proposing that AI coordination replace middle management ^[9]. Deloitte scrapped its analyst-to-consultant-to-manager progression entirely, backed by a $3 billion AI commitment ^[10]. McKinsey's "25 squared" plan runs 25,000 AI agents alongside 40,000 humans, growing client-facing roles while shrinking everything else ^[11]. Analysts already have a name for what is left. Not a flattened pyramid but an obelisk: fewer people at every level, and in the extreme a dot structure where one entrepreneur runs a massive firm ^[12]. The obelisk has a problem in its foundation. It depends on seasoned people whose judgment the models cannot replicate, the very people companies are now competing to hire. But judgment is not downloaded; it is grown. Helen Leis of Oliver Wyman put it plainly.

To have the mid-level people that can manage an agentic workforce, they need to learn the company and the job. — Helen Leis

The companies closest to the problem see the same stakes. IBM is tripling its entry-level hiring ^[1]. AWS's own chief executive, Matt Garman, has pushed back in the bluntest terms.

you have no talent pipeline. — Matt Garman

The European Central Bank found that AI-intensive firms tend to hire rather than fire ^[13]. And TeamLease reported that firms which cut half their staff after adopting AI came back months later needing people to run the technology ^[14]. None of this points in a different direction. It measures the cost of the one we are already traveling. The bet being placed is that companies can import the judgment they have stopped growing. The loudest objections to that bet come from inside the companies making it.