The Career Ladder Is Losing Its First Rung

Every career you admire started with a job someone gave to an unproven person. That handoff, the bet a company makes on someone with potential and no track record, is the mechanism quietly breaking right now. The work that entry-level people used to learn on, the research memos, the first-draft everything, the data cleanup, the ticket triage, is increasingly the exact work AI does cheapest.

The numbers have stopped being anecdotal. Venture firm SignalFire, which tracks hiring across the tech industry, found that entry-level hiring at the fifteen biggest tech companies fell about 25% from 2023 to 2024, while senior hiring held up. A ResumeTemplates survey of hiring managers found nearly half saying they’d rather invest in training AI than hire from the 2026 graduating class. And when Fortune reported on Yale research into agentic AI and early-career work, the framing was the same one workers are reaching for themselves: AI may not take your job, but it’s taking the path to your first one.

Predictions go further. Anthropic’s CEO has suggested AI could eliminate as much as half of entry-level white-collar roles within five years. That is a projection from someone with an obvious stake in AI’s importance, not a measurement, and it deserves the skepticism any five-year forecast deserves. The measured 25% decline deserves none. It already happened.

Why companies stopped hiring at the bottom

No executive announced a policy of not hiring juniors. What happened is quieter: when a team can route routine work through AI, the business case for the next junior hire stops penciling out. Nobody gets fired. The role just doesn’t get backfilled, the req doesn’t get opened, and IEEE Spectrum’s reporting on early-career engineers describes the result, with responsibilities that were the traditional on-ramp now the most automatable part of the org chart.

The uncomfortable economics: a junior employee is a multi-year investment that pays off as they grow. AI’s output at junior-task level is available today, scales instantly, and never accepts a better offer after you’ve trained it. Companies that think in quarters will keep choosing the second option, which is how a structural problem gets built one reasonable-seeming decision at a time.

What firms are actually buying is a delayed liability. Skip junior hiring for five years and you haven’t eliminated the need for senior people in 2031; you’ve eliminated the supply. A few companies will run apprenticeship-style programs precisely because the talent pipeline became scarce, and “we develop people” is becoming a recruiting weapon. But the gap between pipeline-minded firms and quarter-minded firms is where individual careers fall.

If you’re trying to get in

The honest version first: it is harder than it was three years ago, through no fault of yours, and anyone selling a guaranteed path through it is lying.

What still works is attacking the actual reason juniors aren’t hired. Companies skip junior hires because training someone to be useful is expensive. So arrive partially trained. Not with certificates, with evidence: a real project in the domain, built end to end, where you can explain every decision (the longer case for projects over credentials is in Stop Collecting Certificates). A portfolio that proves you can already do the year-one work converts you from investment to bargain.

Aim at the parts of junior work AI handles worst. Anything requiring physical presence, messy coordination between people, accountability for a result, or earning trust with clients resists automation far better than document production. The junior roles that survive will be heavier on those elements, and your application should be too.

And widen the target. The 25% decline was measured at the biggest tech firms, the most AI-saturated employers on earth. Small companies, regulated industries, and unglamorous sectors automate slower and still need hands. The first rung hasn’t vanished evenly; it’s vanished where everyone is looking.

If you’re mid-career, this is about you too

It’s tempting to read the first-rung story with relief. You’re already in. But juniors weren’t just trainees, they were the leverage under every experienced person’s job, and their absence lands on you in specific ways.

The first-draft work comes back to your desk, except now it routes through AI and you’re the quality control. Reviewing machine output is real work, often less pleasant than the drafting it replaced, and it’s accruing to mid-career people without anyone renaming their jobs. The skill of verifying, correcting, and taking responsibility for AI-produced work is rapidly becoming the actual content of experienced roles.

Management paths change shape too. Career ladders assumed you’d eventually run a growing team; teams that stop adding juniors stop growing. Advancement tilts toward scope, owning bigger problems, more systems, more cross-team surface, rather than headcount. If your promotion case is built on “I need more people,” it’s worth rebuilding it around “I can own more” instead.

There’s also a longer game worth playing deliberately. Organizations that quit developing people will eventually pay for it, and the mid-career professionals who stay good at developing others, who can take the one junior the team does hire and make them productive fast, hold a skill that gets scarcer every year it goes unpracticed. Scarce skills get paid.

The reframe

The ladder metaphor was always a little off. What’s disappearing is the standardized first rung, the mass-produced junior role that converted graduates into professionals at scale. Paths into careers still exist; they’re just becoming irregular, narrower, and more dependent on demonstrated work than on credentials and patience. That’s worse for almost everyone starting out, genuinely. It is also the environment now, and it rewards the same thing at every career stage: being able to show, concretely, what you can do that the tools can’t do alone.