Ferrari's Designers Still Start With a Pencil

Opening

Dear reader, Pininfarina—the Italian design studio behind Ferraris and Peugeots—has one rule for its junior designers: every new car must start with a pencil. In an age when AI can spit out 200 renderings, why insist on hand sketches?

This week, the Financial Times offered an interesting answer. AI isn’t eliminating entry-level jobs—it’s rewriting how entry-level employees learn the job in the first place. Here’s the bottom line up front: what global companies are doing right now isn’t a resumption of entry-level hiring. It’s a construction project to rebuild the “learning ladder.” And Korea is one of the markets where this project is most urgently needed.

---

The Ladder That Broke, the Hiring That Returned

Let’s rewind to 2023. A Harvard Business School research team ran an experiment on 758 Boston Consulting Group (BCG) consultants: the group using GPT-4 completed 12.2% more of 18 standardized tasks, 25.1% faster. But on complex management tasks that fell outside AI’s competence, the same group was actually 19% less likely to produce a correct answer. The researchers called this uneven boundary the “jagged frontier”—but here’s what matters most: the tasks AI handled best happened to be exactly the tasks typically assigned to junior employees.

The numbers confirm it. According to the “Canaries in the Coal Mine” paper published by Erik Brynjolfsson’s team at the Stanford Digital Economy Lab, employment among 22-to-25-year-olds in AI-exposed occupations fell by roughly 13% relative to other groups since generative AI took off. Meanwhile, employment for experienced workers in the same occupations stayed stable. In Brynjolfsson’s words, “what young workers know and what LLMs can replace” turned out to overlap.

But the trend has been shifting this year. In the National Association of Colleges and Employers (NACE) Spring 2026 hiring outlook, employers said they plan to increase new-college-graduate hiring by 5.6% this year. That’s a full reversal from just six months earlier, when the same figure was -2.4%. That said, the sample is small at 185 responding companies, and the growth is concentrated in large employers with 5,000+ staff, at 8.7%.

What’s more interesting is an analysis by Ramp and Revelio Labs, which linked spending and workforce data across roughly 21,000 U.S. companies. Companies that invested most aggressively in AI grew overall headcount by about 10% over two years after adoption, and entry-level hiring by as much as 12%. In other words, the companies using AI the most are hiring the most junior talent. To be clear, this isn’t proof that AI creates jobs—the more natural reading is that companies successful enough to invest heavily in AI are also companies successful enough to hire broadly. The researchers themselves attached the same caveat.

LinkedIn’s Aneesh Raman sums up the shift this way: “We’re starting to see signals that entry-level work is moving from grunt work to real work.” Hiring hasn’t come back—a different kind of job is being created.

---

Grunt Work Was Actually Tuition

There’s something worth pausing on here. The tasks we called grunt work—cleaning up meeting notes, drafting research, reviewing contracts, revising drawings—were never mere busywork. They were tuition paid to absorb an organization’s tacit knowledge¹ through the body. Learning, by doing grunt work next to a senior colleague, why this organization judges things the way it does—that was the core of the apprenticeship model.

AI has shut down this tuition window. The problem is that companies have started demanding the diploma without offering the class. According to PwC’s 2026 AI Jobs Barometer, which analyzed over a billion job postings across 27 countries, entry-level postings that demand senior-level skills—call them “seniorized” postings—have grown 35% since 2019, while ordinary entry-level postings have fallen 10%. In AI-exposed occupations, 52% of the skills newly appearing in entry-level postings were skills once reserved for experienced hires. In low-AI-exposure occupations, that figure was just 7%.

It’s a paradox: take away the ladder for learning, then demand the skills that sit at the top of it. The same gap shows up in education. In a Pearson/AWS AI readiness report surveying 2,700 people across six countries, 78% of university leaders believe their graduates meet employer expectations—yet 53% of employers say they struggle to find graduates with AI skills. Only 14% of graduates rated themselves as having high-level practical AI proficiency.

NYU’s Gary Marcus adds one more concern: junior employees who grow up dependent on AI never develop critical thinking, and end up unable to catch AI’s own mistakes.

This is where Pininfarina’s pencil rule reenters the picture. Daniel Lee, chief designer at the studio’s Shanghai location, explains the problem with AI renderings this way: a wrong proportion used to show up immediately in a line, but now vivid colors and backgrounds mask the mistake. That’s how a rendering of a combustion-engine car can end up missing its radiator vents entirely—a basic-of-basics error. Insisting on pencils isn’t nostalgia. It’s a training design meant to build fundamentals in an environment where mistakes are visible.

---

Three Ways to Rebuild the Ladder

The responses from leading companies fall into three broad categories.

First: using AI as a trainer, not a replacement. Germany’s DHL trains AI on its official manuals, then has veteran employees nearing retirement supplement and correct that knowledge. It’s a way of converting tacit knowledge that would otherwise vanish into a shared organizational asset. Consulting firm Cognizant built an AI “harness” that guides new hires; CEO Ravi Kumar compares it to a self-driving system built on the experience of millions of drivers. What used to be learned by walking the halls and watching over someone’s shoulder is now passed down by the system itself.

Second: redefining the role of the middle manager. Kumar believes middle managers, who used to measure and coordinate, must become “player-coaches” who directly develop junior staff. The CEO of legal AI company Luminance offers a more radical forecast: only AI-literate juniors and AI-boosted seniors survive, and the middle layer disappears entirely. Either way, there’s no scenario where the middle manager survives unchanged.

Third: restoring in-person collaboration. At engineering consultancy Arup, 70% of junior growth comes from hands-on work, 20% from peers, and only 10% from formal training. Niels Fischer of Zaha Hadid Architects goes a step further, suggesting that it may have been remote work, more than AI, that damaged junior employees’ soft skills. It’s a paradox of the AI age: the more technology advances, the more companies are herding people back into the same room.

---

In Korea, the Ladder Broke First

Overlay this global trend onto Korea, and an uncomfortable fact emerges: Korea kicked away its ladder before AI ever arrived.

It’s been years since large Korean conglomerates shifted from scheduled mass hiring (gonggae, Korea’s traditional twice-a-year corporate recruitment cycle) to year-round, rolling recruitment, and what filled the gap was the “junggo-sinip”—literally “used entry-level”—job seekers who already have some work experience but apply for entry-level roles anyway. According to an Incruit survey, 28.9% of new college-graduate hires last year were junggo-sinip, up 3.2 percentage points from 25.7% in 2023. HR managers put the ceiling for entry-level applicants at 3.1 years of prior experience, and 73.8% of not-yet-hired job seekers said they’d be willing to apply as junggo-sinip themselves.

Here’s the underlying structure. While global companies rebuild the learning ladder inside the firm, Korean companies have been offloading the cost of that ladder onto individuals—looking for a finished product that trained somewhere else. This model collapses the moment everyone adopts it: if no one is willing to be someone’s first job, there’s no experienced hire to poach three years later. And now that AI is starting to take over entry-level tasks, there are even fewer places left where an individual can pay that tuition on their own.

That’s why I read this FT piece less as a trend report and more as a warning for Korean companies. Experiments like AI trainers, player-coaches, and deliberately designed in-person collaboration aren’t optional extras—they’re the only insurance policy for when the junggo-sinip model finally runs out of road.

---

Oz’s Lens

In my work consulting on GTM strategy and organizational design, I often sit in on hiring-plan discussions. What surprises me every time is how easily a room decides, “We won’t hire entry-level for a while.” What that decision actually means is: we will import our organization’s future seniors from the market. The problem is that if every company makes the same call, there’s no senior left in the market to import—and Korea is already standing at the entrance to that exact problem.

The sentence I found most important in this piece was LinkedIn’s Raman’s: “Entry-level work is a preview of where all work is headed.” What’s happening to junior employees right now—repetitive work shifting to AI, humans shifting to judgment, verification, and coordination—is a trailer for what will happen at every level within a few years. A company that solves its entry-level onboarding problem is, in effect, the first company to solve the organizational design problem of the AI era.

So here’s one practical suggestion: before you debate whether to hire entry-level staff, build a list of every educational function that grunt work used to serve in your organization. That list becomes the spec sheet for the learning system your organization needs to deliberately redesign for the AI era.

---

Closing

Here’s today’s story in three lines. AI eliminated how junior employees learned before it eliminated their jobs. Leading companies are rebuilding that ladder with AI trainers, player-coaches, and in-person collaboration. And Korea, having relied on junggo-sinip for years, is the market where this redesign is most urgently needed.

What was the first thing entry-level employees learned in your organization? If AI has since made that task disappear, I’d love to know what’s filling the gap now. Whether you’re experiencing this as a junior employee or wrestling with it as someone developing junior talent, either perspective is welcome. Leave a comment, and I’ll fold your thoughts into the next issue.

---

💬 Tell us in the comments about the “vanished grunt work” in your organization and what’s filling its place. I’ll fold it into the next issue. 📨 If you have a colleague or manager wrestling with entry-level hiring, please share this piece with them.

---

References & Further Reading

Primary sources

• Brynjolfsson, E., Chandar, B., & Chen, R., “Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence”, Stanford Digital Economy Lab, 2025. (original PDF) ··· This paper provides empirical evidence of employment declines among 22-to-25-year-olds in AI-exposed occupations. It’s the data that started today’s piece, so if you only read one source, make it this one.
• PwC, “Two Futures for Jobs in an AI Era: 2026 Global AI Jobs Barometer”, 2026. (full report PDF) ··· Shows the “seniorization” of entry-level postings across a billion job listings. Especially worth reading if you’re the one designing job postings.
• Ramp Economics Lab & Revelio Labs, “A New Look at AI’s Impact on Jobs”, 2026. (Revelio Labs commentary) ··· A counterintuitive finding that heavy AI investors are hiring more entry-level staff, not less. Read alongside the researchers’ own caveat that this isn’t causation, to avoid misreading it.
• Dell’Acqua, F. et al., “Navigating the Jagged Technological Frontier”, Harvard Business School Working Paper No. 24-013, 2023. ··· The original paper behind the 758-consultant BCG experiment. The source of the “jagged frontier” concept describing the uneven boundary between what AI does well and poorly.
• NACE, “Employers Expect to Hire 5.6% More New College Graduates This Year”, 2026. ··· A survey showing the U.S. new-grad hiring outlook reversing within six months. Keep in mind the sample is only 185 companies.
• Financial Times, “AI isn’t destroying entry-level jobs. It’s changing them”, 2026. ··· The article that inspired today’s piece. Contains detailed accounts of Pininfarina, DHL, and Cognizant. (FT subscription required.)

Background

• Pearson & AWS, “AI Readiness: Building the Bridge from Higher Education to Work”, 2026. ··· Shows the perception gap between universities (78%) and employers (53%) across six countries. If you’re involved in education, the framework section is especially useful.
• Incruit, “More ‘Junggo-sinip’ Job Seekers: 61% of Workers Say They’d Give Up Career Level to Apply as Entry-Level”, 2025. ··· The source for today’s Korea-section figures.

The author, Kwangseob Ahn, is a professor of business administration at Sejong University and lead consultant at OBF (Oswarld Boutique Consulting Firm). He teaches statistics and data analysis — business data management and business analytics — while leading GTM and AI strategy consulting in the field, designing the seam between technology and business. He has published academic research on a memory architecture for AI dialogue systems (HEMA) and runs Daily Arxiv, a daily curation of global AI papers. He holds a master's from Korea University's Graduate School of Technology Management and a KMBA. He is the author of Homo Brainless: The People Who Outsource Their Thinking.

Footnotes

1. Tacit knowledge: knowledge that isn’t written down in any document or manual, and is passed on only through experience. Think of it as embodied know-how—like knowing you should phrase things a certain way with a particular client. ↩