The Career Ladder Is Breaking. Here Is What That Means for How You Hire.
By Creatives Takeover Editorial Team · August 10, 2026
Traditional career ladders are giving way to skills-first hiring.
Entry-level job postings in the United States have fallen 35 percent since early 2023. In specific categories, the drop is far steeper: junior software development and data analysis postings are down as much as 67 percent in some sector-level tracking. At the biggest technology companies, new graduates made up just 7 percent of new hires in 2024, a 25 percent decline from the year before, and more than 50 percent below pre-pandemic hiring levels. At startups, the picture is even starker: graduate hiring fell from 30 percent of new hires in 2019 to under 6 percent by 2024.
Stanford research tracking workers aged 22 to 25 in roles most exposed to AI, including software development, customer service, and accounting, found a 16 percent relative drop in employment for that group in under three years. Anthropic CEO Dario Amodei has publicly warned that AI could eliminate half of all entry-level white-collar jobs within one to five years, potentially pushing unemployment as high as 10 to 20 percent. Brookings Institution research estimates AI could automate more than 50 percent of tasks in entry-level positions, roughly five times the automation risk facing more senior roles.
Those numbers, taken together, describe something considerably more structural than a rough hiring season. They describe the bottom rung of the traditional career ladder narrowing in real time.
The Debate Worth Taking Seriously Before Drawing Conclusions
Before accepting a single, tidy explanation for what is happening, it is worth sitting with a genuine and unresolved disagreement among people who study this closely.
One camp, anchored by figures like Amodei and reinforced by SignalFire's hiring data and Brookings' automation modeling, argues this is fundamentally an AI story: routine coding, junior financial analysis, first-draft writing, and basic data work are exactly the tasks current AI tools already do competently, and entry-level roles have historically been built almost entirely around exactly that kind of routine, learnable-by-repetition work.
A second, more skeptical camp pushes back with real evidence of its own. Analysis from Technical.ly, drawing on Federal Reserve data, points out that underemployment among recent college graduates reached 42 percent this spring, the highest since the pandemic year of 2020, but was actually higher following the recessions of 1990, 2001, and 2008 to 2009, well before generative AI existed in any commercial form. That analysis also notes that only about 18 percent of jobs are currently considered at high near-term automation risk, a meaningful number, but not one that alone explains a 35 percent collapse in entry-level postings. Their argument is that a frozen broader hiring market, a demographic bulge of job seekers, and general economic uncertainty are doing more of the actual work than AI is, with AI serving as a convenient, easily understood explanation for a more complicated and less satisfying set of causes.
Both camps are working from real data. The honest position is that this is very likely not a single-cause story: AI is a genuine and measurable factor in specific, automation-exposed roles, layered on top of a broader cyclical hiring slowdown that would be occurring, to some degree, even without it. Trying to force the entire phenomenon into one clean narrative, in either direction, means missing at least half of what is actually happening.
Why "Wait and See" Is Not a Neutral Choice
Regardless of which explanation carries more weight, the practical consequence for how organizations hire is the same, and it is already visible in the data. Nearly half of hiring managers surveyed by ResumeTemplates.com, 48 percent, said they would rather invest in AI tools than hire and train a recent college graduate. Forty percent of chief executives surveyed by Boston Consulting Group plan to reduce junior roles specifically over the next one to two years. Nearly 4.6 million students who wanted an internship this past year could not secure one, cutting off the traditional pipeline that used to feed entry-level hiring in the first place.
The risk in that collective behavior is not really about fairness to new graduates, though that is a real and separate concern. It is a structural risk to the organizations making the choice. SignalFire's research, tracking hiring patterns across more than 650 million professionals, frames the deeper concern precisely: removing the first rung of the ladder raises real questions about how institutional knowledge actually transfers inside an organization, and how anyone rises into senior roles a decade from now if almost nobody was hired and developed at the junior level today. A company that quietly stops hiring and training junior talent for several consecutive years is not simply saving money in the short term. It is deferring a very specific, very predictable talent gap into its own future, one that will arrive at exactly the moment that cohort would otherwise have been ready to step into mid-level and senior positions.
The Company Making the Opposite Bet
Not every organization is responding to this pressure the same way, and the contrast is instructive.
IBM announced it will triple entry-level hiring in the United States in 2026, even as most of its industry peers pull back. The company's own explanation is specific and worth taking seriously rather than dismissing as a public relations gesture: junior software developers at IBM are reportedly spending measurably less time on routine coding tasks, now handled largely by AI tools, and measurably more time working directly with customers, a shift in role composition rather than a wholesale elimination of the role itself. IBM's framing is that younger workers, entering the workforce already fluent in AI-assisted tools as a baseline expectation rather than a novel skill, may actually be a better long-term investment during a period of genuine technological transition than a strategy built entirely around retaining only already-experienced staff.
That is a meaningfully different read on the same underlying disruption than the one driving most of the industry's pullback. Where many organizations appear to be treating AI capability as a direct substitute for junior headcount, IBM appears to be treating it as a reason to redesign what a junior role actually does, keeping the hire, changing the job description underneath it.
What a Redesigned Entry-Level Role Actually Looks Like
The organizations navigating this most effectively are not simply choosing between "hire junior people the old way" and "stop hiring junior people entirely." They are doing something more specific: redesigning junior roles around outcomes and applied judgment rather than around the routine, repetitive tasks that AI tools now handle competently on their own.
In practice, that means being explicit about what a junior hire actually owns, rather than treating the role as a general pool of low-cost labor for whatever administrative or repetitive work needs doing. It means building an AI-enabled training plan deliberately, treating AI proficiency itself as a core skill to be developed and evaluated, alongside judgment, quality control, and the ability to know when an AI-generated output is actually good enough to ship. It means being honest that roles built almost entirely around administrative or purely repetitive tasks are genuinely disappearing and are not worth preserving in their old form, while roles that combine some routine work with real customer exposure, real problem-solving, and real ownership of an outcome are considerably more durable, because those are precisely the parts of the job current AI tools do not yet reliably replace.
Rathod, a career strategist quoted in recent Forbes reporting on this shift, put the practical implication for candidates directly: specificity matters more than ever when there are fewer entry points available, because it signals a results-driven person capable of demonstrating actual impact, not simply a credential. That same logic applies directly to how a founder should be designing the role itself, not just how a candidate should be applying for it. A junior role with a vague, catch-all job description built around administrative support is exactly the role most likely to be quietly eliminated. A junior role built around a specific, ownable outcome, with AI treated as a tool the person uses rather than a replacement for the person, is the version far more likely to survive the next several years of this transition intact.
What This Means for a Founder Deciding Whether to Hire Junior Talent Right Now
For an early-stage founder specifically, this entire debate collapses into a much more practical, immediate question: is a junior hire, right now, actually the right decision for this specific task, or is defaulting to "just use AI for this" quietly setting up a talent gap the company will feel acutely in three or four years, once the tasks that junior person would have grown into no longer have anyone ready to fill them.
The useful diagnostic is not "can AI currently do this task," because for an increasing number of narrow, well-defined tasks, the honest answer is yes. The more useful diagnostic is whether the role, designed correctly, combines some of that AI-assisted routine work with genuine customer exposure, real decision-making under ambiguity, and a defined path toward ownership of something larger. A role that clears that bar is one worth hiring and training a person into, even in a tight market, because the judgment, context, and institutional knowledge that person accumulates over eighteen months is not something a language model accumulates on the company's behalf in the same way. A role that does not clear that bar, one built entirely around routine, repeatable, low-ambiguity tasks, is very likely correctly identified as a role AI genuinely should absorb, and trying to preserve it as a headcount line purely out of habit or sentiment is not a sound long-term strategy either.
Five Things Worth Taking From This
The AI-versus-cyclical-downturn debate is not actually resolved, and treating it as settled in either direction is a mistake. Real data supports both explanations simultaneously. Plan for a combination of both forces rather than betting the company's hiring strategy on a single, tidy narrative.
Cutting junior hiring saves money today and creates a specific, predictable gap later. SignalFire's core warning, that removing the first rung of the ladder raises real questions about institutional knowledge transfer and future senior pipeline, deserves to be treated as a genuine business risk, not just an abstract concern about fairness to new graduates.
Redesigning the role is a better response than simply eliminating it. IBM's bet, that a junior role reshaped around customer interaction and applied judgment rather than routine tasks is worth tripling investment in, is a specific, concrete alternative to the more common industry response of quietly shrinking junior headcount altogether.
The right diagnostic is not "can AI do this task" but "does this role combine AI-assisted work with genuine ownership." A junior role built entirely around routine, repeatable tasks is very likely correctly identified for AI to absorb. A junior role that pairs AI-assisted efficiency with real customer exposure and decision-making is exactly the kind of role worth continuing to hire and train people into.
Specificity is now doing more work than credentials, for both the candidate and the role itself. Just as career strategists are telling graduates that a vague resume no longer competes in this market, founders should hold their own junior job descriptions to the same standard: a role built around a specific, ownable outcome survives this transition. A vague, catch-all junior role does not.
The career ladder is not disappearing entirely, and the debate over exactly how much of its narrowing is AI's doing will likely stay unresolved for some time. What is already clear, and already actionable, is that the old, undifferentiated version of the entry-level role, a general pool of low-cost labor for routine tasks, is genuinely going away. The founders who redesign what that role actually does, rather than simply deciding whether to keep or cut the headcount line, are the ones most likely to have a functioning talent pipeline in place when they need one most.