The Decision That Contradicts Everything I’ve Gotten Used to Reading

I’ve lost count of how many posts I’ve written here about companies cutting entry-level jobs because of AI. So when I saw IBM’s announcement, I had to read it twice: the company automated a huge chunk of routine HR work — its internal AskHR assistant already handles about 94% of routine interactions on its own — and instead of cutting entry-level roles, it announced it’s tripling entry-level hiring in the US for 2026.

The explanation came straight from IBM’s CHRO, Nickle LaMoreaux, at a conference in New York: if the company stops investing in entry-level hiring now, what happens in 3 to 5 years? There’s no pipeline. The well simply dries up.

I find that line uncomfortably simple for an entire industry that seems to have forgotten that logic.

The Mistake I Think the Market Is Making at Scale

Hiring someone at the start of their career was never just about executing basic tasks. The mistakes a junior makes aren’t wasted time — they’re the shaping process someone needs to actually understand culture, internal politics, and how a company really works underneath the surface. You can’t buy that ready-made on the open market.

And that’s exactly where Meta chose the opposite path.

Instead of building talent internally, Mark Zuckerberg created Meta Superintelligence Labs and started offering compensation packages reaching into nine figures — hundreds of millions of dollars — to pull senior researchers directly from OpenAI, Google, and Apple. He personally recruited much of the team himself, reportedly WhatsApping hundreds of researchers and hosting some at his own homes.

The Cost of Buying the Top Without Cultivating the Base

The result of this “buy the top, automate the base” strategy wasn’t quiet. In an internal memo reported by Reuters, Zuckerberg himself admitted Meta “made mistakes and will almost certainly make more” during a restructuring that affected roughly 20% of its global workforce — 8,000 layoffs and another 7,000 employees reassigned to AI initiatives.

He promised no more company-wide layoffs this year, and announced measures to try to fix the internal climate: bigger budgets for team events, scaling back the lopsided manager-to-employee ratio in the AI engineering unit, a company-wide hackathon in July. That’s, essentially, a public admission that star researchers with no history at the company collided directly with the culture and bureaucracy already in place there.

Two Bets, Two Time Horizons

Strategic indicatorIBM’s approach (build the pipeline)Meta’s approach (hunt the top)
Main actionTriple entry-level hiringUp to 9-figure packages for renowned researchers
Short-term viewAutomating a large share of operational routines with AITrying to speed up delivery of frontier models
5-year impactA consolidated pipeline of future managers and leadersRisk of a mid-level talent blackout and succession gap
Culture and frictionLow friction — people shaped by the company’s own cultureHigh friction — culture clashes, turnover, public mea culpa

What I Actually Think

I don’t think Meta was wrong to want top-tier researchers — to build a frontier model, sometimes you genuinely need people who already know how to do that specific thing, and you can’t wait for someone to grow into it internally on a competitive timeline. The problem isn’t buying senior talent. It’s doing that instead of cultivating the base, as if the two were substitute strategies.

You can buy someone’s technical experience on the open market. What you can’t buy is the years it takes for a person to actually understand how a specific organization works from the inside — the unwritten shortcuts, the internal politics, who actually decides what. That only gets built from the inside out, by people who grew up there.

And that’s what worries me about companies closing the entry-level door to save money right now: they’re trading a visible cost today (entry-level salary) for an invisible cost tomorrow (having nobody seasoned enough to promote when the leadership seat opens up).

I’m Left With This Question

Automation should be amplifying the learning of people just starting out — not closing the front door of the company. IBM seems to have understood that distinction. Meta is learning it the hard way, in real time, with an internal memo and a reconciliation hackathon.

Is your company already feeling the absence of mid-level and senior people because it cut entry-level hiring in recent years? How do you see the job market for people just starting out?

94% HR automation, and IBM is still tripling entry-level hiring. Meanwhile, a 9-figure salary didn’t stop Meta from having to apologize to its own team in writing.

Read Also