Jack Clark thinks coding is the new literacy. Not in the vague, aspirational way that phrase has been tossed around Silicon Valley for the past decade. The Anthropic co-founder — whose net worth now sits comfortably in billionaire territory thanks to the AI company’s meteoric valuation — is making a specific, urgent argument: that the United States faces a generational employment crisis unless it fundamentally rethinks how it educates young people for a world being reshaped by artificial intelligence.
His timing is deliberate. Gen Z unemployment has become a quiet crisis, one that doesn’t generate the same headlines as inflation or interest rates but carries implications just as serious for long-term economic growth. And Clark, who helped build one of the most powerful AI systems on the planet, believes the fix starts in classrooms, not boardrooms.
In a recent interview with Fortune, Clark laid out a vision that’s equal parts alarming and pragmatic. He argues that programming education should be treated with the same seriousness as reading and mathematics — woven into curricula starting in elementary school, not bolted on as an elective in high school. “We’re preparing kids for an economy that doesn’t exist anymore,” Clark told Fortune. The economy that does exist, and the one forming rapidly ahead, demands computational thinking as a baseline skill, not a specialization.
This isn’t just philosophical musing from a tech executive with time on his hands. Clark’s argument rests on data that’s hard to dismiss. Youth unemployment in the United States has been running persistently above the national average, and the jobs that are available increasingly require some level of technical fluency. The Bureau of Labor Statistics projects that computer and information technology occupations will grow by roughly 15% through 2032, far outpacing the average for all occupations. Meanwhile, the jobs disappearing fastest — routine administrative roles, basic data entry, certain categories of customer service — are precisely the entry-level positions that have historically served as on-ramps for young workers without specialized training.
The irony is thick. AI companies like Anthropic are building the very systems accelerating this displacement, and now one of Anthropic’s founders is sounding the alarm about the consequences. Clark doesn’t shy away from this tension. He acknowledges that AI will eliminate certain categories of work entirely. But he frames coding education not as a way to turn every child into a software engineer — rather, as a way to ensure they can work alongside AI systems, direct them, audit them, and build on top of them.
Think of it this way. A century ago, the ability to read wasn’t considered a universal necessity. It was a skill for clerks, scholars, and the professional class. Industrialization changed that calculus entirely, and within a few decades, literacy became non-negotiable for participation in modern economic life. Clark sees an identical inflection point arriving now, compressed into a much shorter timeline.
Wall Street has started paying attention to the workforce implications of AI, though mostly through the lens of corporate productivity gains and margin expansion. The investment thesis is straightforward: companies deploy AI, headcount falls or flattens, earnings per share climbs. Goldman Sachs estimated last year that generative AI could eventually automate the equivalent of 300 million full-time jobs globally. Morgan Stanley’s research division has published extensively on AI-driven productivity, generally framing it as a net positive for corporate profitability.
But there’s a demand-side problem that gets less airtime on earnings calls. If a generation of workers can’t find meaningful employment — or can only access low-wage service jobs with limited upward mobility — consumer spending suffers. Housing formation slows. Household debt rises. The macro picture gets uglier in ways that don’t show up in near-term EPS projections.
Clark is essentially arguing that coding education is infrastructure. Not the bridges-and-broadband kind that politicians love to fund, but human capital infrastructure — the kind that determines whether the AI productivity boom generates broadly shared prosperity or a deeper stratification between those who can work with these systems and those who can’t.
He’s not alone in this thinking, though he may be the most prominent voice making the case right now. Hadi Partovi, the CEO of Code.org, has spent years pushing for computer science to be treated as a core academic subject rather than a vocational afterthought. As of 2024, only about 57% of U.S. high schools offer a foundational computer science course, according to Code.org’s annual state-by-state analysis. That number has improved dramatically from a decade ago, when it was in the low single digits. But it remains wildly insufficient given the pace of technological change.
And the disparities are stark. Rural schools, under-resourced urban districts, and schools serving predominantly Black and Latino students are far less likely to offer computer science instruction. This isn’t just an equity issue in the abstract. It’s a labor market pipeline problem that will compound over time as AI fluency becomes a prerequisite for an expanding share of jobs.
The corporate world’s response has been uneven. Some of the largest technology companies — Google, Microsoft, Amazon — fund coding education initiatives, though critics argue these programs are often more about talent pipeline development and public relations than systemic reform. Apple’s Everyone Can Code curriculum has been adopted by schools and community colleges, but adoption remains patchy. The federal government, meanwhile, has been slow to act with the kind of urgency Clark is calling for. The CHIPS and Science Act included provisions for STEM education, but coding-specific mandates remain largely absent from federal policy.
State-level action has been more promising. In 2023, Nevada became one of several states to require all high schools to offer computer science. Arkansas, under then-Governor Asa Hutchinson, was an early mover, mandating computer science education in every public high school starting in 2015. But “offering” a course and ensuring students actually take it — and take it seriously — are very different things. Without graduation requirements tied to computer science, enrollment often remains optional and skews heavily toward students already inclined toward technical subjects.
Clark’s framing as reported by Fortune pushes further than most education advocates have been willing to go. He’s not talking about offering more AP Computer Science sections. He’s talking about integrating computational thinking into how children learn from the earliest grades — the same way number sense and phonics are treated as foundational cognitive skills, not electives.
There’s a reasonable counterargument, and it deserves airing. Some educators and economists push back on the “everyone must code” narrative, arguing that AI itself will increasingly handle the coding. If large language models can already generate functional software from natural language prompts, why would basic programming be a universal requirement? Won’t AI make coding obsolete before the next generation of students even enters the workforce?
Clark has addressed this directly. His response: coding isn’t really about writing syntax. It’s about understanding logic, systems thinking, abstraction, and how to decompose complex problems into manageable components. Even if AI writes most of the code in the future, someone has to understand what the code is supposed to do, verify that it works correctly, and identify when it fails. That someone needs computational literacy. The analogy to reading holds. Spell-check didn’t eliminate the need for literacy. Calculators didn’t eliminate the need for numeracy. AI won’t eliminate the need for people who understand how software works.
The labor market data supports his urgency, at least directionally. A March 2025 report from the National Association of Colleges and Employers found that employers across sectors — not just technology — increasingly list technical skills as preferred or required for entry-level positions. Financial services firms want analysts who can write Python scripts. Healthcare organizations want administrators who understand data structures. Marketing departments want people who can work with analytics platforms at a level beyond clicking buttons. The floor for technical competency is rising across the economy, and it’s rising fast.
For investors, the implications extend beyond education stocks and edtech plays. The speed at which the U.S. workforce adapts to AI will be a critical variable in determining whether AI-driven productivity gains translate into sustained economic growth or produce the kind of structural unemployment that drags on GDP. Japan’s lost decades offer a cautionary tale about what happens when demographic and structural workforce challenges go unaddressed. The U.S. isn’t Japan, but the parallels are worth considering.
Clark’s Anthropic, meanwhile, continues to push the frontier of what AI systems can do. The company’s Claude model competes directly with OpenAI’s GPT series and Google’s Gemini, and Anthropic has raised billions in funding at valuations that reflect enormous expectations for AI’s economic impact. There’s an inherent tension in a company building tools that automate cognitive work while its co-founder simultaneously warns that society isn’t preparing people for the disruption those tools will cause. But perhaps that tension is precisely what lends Clark’s argument its credibility. He’s not speculating about what AI might do to the job market. He’s watching it happen from the inside.
The question now is whether anyone with the power to act — governors, state legislators, school board members, federal policymakers — will treat Clark’s warning with the urgency it warrants. Education systems are famously slow to change. Curriculum reform is politically fraught, resource-intensive, and subject to the inertia of institutions that measure progress in academic years, not quarters. But the AI industry isn’t waiting. Every month brings new capabilities, new automations, new categories of work that machines can perform faster and cheaper than humans.
So the gap widens. Between what the economy demands and what schools produce. Between the skills employers need and the training young workers receive. Between the speed of technological change and the pace of institutional response.
Jack Clark is betting that coding education can close that gap — or at least prevent it from becoming unbridgeable. Whether he’s right or merely optimistic, the underlying problem he’s identified is real, it’s growing, and it isn’t going away. The billionaire building the AI future is telling us the workforce isn’t ready for it.
That should make everyone — educators, policymakers, and yes, investors — deeply uncomfortable.
The Billionaire Who Says Your Kids Should Learn to Code Like They Learn to Read — And Why Wall Street Should Listen first appeared on Web and IT News.