As AI gets better at technical work, the real differentiator may be skills we usually associate with the humanities: critical thinking, communication, and understanding people.
Loading…
Playbooks on AI and Web3 product marketing, GTM and distribution. No spam, unsubscribe anytime.
Written by Kevan Shah, Web3 and AI product marketing strategist and founder of Hangryfeed.
Head of BD at Zealy and growth operator with 8+ years experience helping AI and Web3 products raise $26M+, scale to $2.2B monthly DeFi volume, and reach 1.2M+ users.
Picture someone who studied literature or philosophy, now working in marketing or operations at a growing startup. For years, they’ve heard the same story: you should have done computer science. Then large language models arrive, and overnight it looks like even their writing and research work might be automated.
That’s the tension a lot of people are living with right now: if AI can code, write, analyze data, and spit out strategy decks, what’s left that humans are uniquely good at—and is there any room in that picture for humanities skills?
One of the more surprising voices weighing in on this is Daniela Amodei, cofounder and president of Anthropic, a leading AI company. She studied literature, not computer science, and says she doesn’t regret it for a second.[3][4][6] Her argument is simple but uncomfortable: as AI gets very good at STEM-style tasks, humanities majors may actually become more important, not less.[3][4][6]
That flips the usual story on its head—and it’s worth unpacking what it really means for people building companies and careers around AI.
Amodei’s starting point is straightforward: modern AI systems are already very good at STEM.[6] They can:
If you’re a founder, you’ve probably already seen this: a single engineer, paired with an AI assistant, can ship more features than a whole team from five years ago. A marketer can spin up ten variants of a campaign in an afternoon. A product manager can generate specs, mock user flows, and pull a quick competitive review before lunch.
The immediate instinct is to assume this kills off non-technical work first—writing, editing, research, basic customer communication. But Amodei draws a different line. In her words, “In a world where AI is very smart and capable of doing so many things, the things that make us human will become much more important instead of much less important.”[3][6][5][8][9][11]
She specifically points to:
Her claim isn’t that AI will leave humanities work untouched. It’s that once the technical floor is raised by machines, the ceiling for human value shifts somewhere else.
There are a few different threads in this argument.
If an AI can ship passable code and content on command, the question stops being “can you produce?” and becomes “can you judge, integrate, and steer?”
That’s where humanities skills show up:
Media coverage of Amodei’s comments emphasizes that as machines take over more technical tasks, the emphasis is moving towards critical thinking, creativity, effective communication, and insightful questioning.[1][3][6][5][8][9][11] Those are all classic humanities capabilities.
AI is strongest when the problem is narrowly defined: “sort this,” “optimize that,” “generate X given Y.” Real-world AI work inside companies is messier:
Amodei highlights that reading, arguing, and dealing with ambiguous texts and ideas is directly relevant to designing and governing AI systems.[3][6] The tools don’t make those judgment calls for you; they just make it cheaper to explore options.
For founders and operators, this means the bottleneck moves from “can we build it?” to “should we build it, and how?” That’s not an engineering question alone. It’s a question about values, incentives, history, and human behavior—territory humanities have always lived in.
Every AI product lands in a social, legal, and cultural context:
Amodei talks about “understanding ourselves, understanding history, understanding what makes us tick” as critical in an AI-driven world.[3][5][6][8][9][11] That’s not philosophical decoration; it’s practical situational awareness.
Anthropic itself is described as hiring for communication skills, emotional intelligence, and a desire to help others, not just raw coding talent.[7] That hints at a broader pattern: AI companies now need people who can translate between engineers, policymakers, customers, and the public—and anticipate how systems might be misused or misunderstood.
In that sense, humanities degrees aren’t random ornaments on a tech company org chart. They can be pipelines for alignment, governance, and communication roles, where the job is to think hard about how powerful systems meet real people.[6][7][15]
There’s also the simple signaling value of who’s at the table.
Clark argues that the most relevant future degrees will involve “synthesis across a whole variety of subjects and analytical thinking about that.”[15] That’s more or less a working definition of a liberal arts education.
Their biographies don’t prove that humanities degrees are a magic ticket. But they do show that high-leverage roles in AI aren’t exclusively reserved for people with traditional STEM training—especially when the work is about deciding what these systems should do, not just how to make them faster.
Still, there’s a lot of healthy skepticism around claims that humanities will be “more important than ever.”
A few of the main concerns:
Even if critical thinking and empathy are more needed, that doesn’t guarantee well-paid jobs for humanities graduates. Humanities degrees are still associated with lower average earnings and less linear career paths in many labor markets.
There’s a risk that statements like Amodei’s are heard as “don’t worry, your degree is safe,” when what they really mean is “the skills you could develop with that degree matter.” Without clear data on hiring and pay, it’s easy for students and early-career professionals to over-interpret the optimism.
Most core AI research and engineering roles still require serious technical training—math, computer science, statistics. Humanities can be complementary, but rarely sufficient for writing model architectures or deploying complex systems at scale.
From a founder’s perspective, that reality shows up in job descriptions. Even if you value ethics and communication deeply, you still need people who can ship infrastructure and models. Humanities alone aren’t going to cover that.
Some of the first tasks AI took on are routine writing, summarization, and content generation, which overlap with entry-level humanities work. There’s no guarantee that non-technical roles are safer just because they’re “human-centered.”
This forces a more nuanced view: not all humanities-type tasks are defensible, and not all STEM tasks are doomed. The safest ground tends to be work that combines multiple kinds of expertise and requires judgment under uncertainty, not just output on demand.
It’s also possible for something to be vital for a healthy society and poorly rewarded economically. Humanities clearly matter for democracy, ethics, and culture. That doesn’t mean employers will suddenly start paying more for philosophy degrees.
“More important than ever” might be truer as a statement about what we ought to value than about what the labor market currently prices in. That distinction matters for anyone making long-term education or hiring decisions.
The real takeaway isn’t “everyone should major in history now.” It’s more practical than that.
If you’re already in a humanities track, the message here isn’t “relax” but “supplement.” Keep the deep reading, argumentation, and human insight—but add:
If you’re coming from STEM, do the reverse: take seriously the idea that synthesis, ethics, and communication will matter at least as much as your ability to optimize an algorithm.[15]
The common thread is simple: the AI economy will likely reward people who can both understand powerful tools and understand the humans they affect. Humanities alone or STEM alone is a narrower bet than a thoughtful mix of the two.