In a 2024 discussion with Tyler Cowen in Miami hosted by the Mercatus Center, Peter Thiel, famed founder of PayPal, Palantir and Founders Fund and the first outside investor in Facebook. was invited to speak on the complexities of political theology. When the conversation turned to AI and its impact on Silicon Valley, he gave his thoughts.
In the interview, Cowen asks Thiel what he thinks the large language models’ (LLMs) – computational models designed for generation of text – impact will be on “wordcels”, a term used to refer to those who write and revel in the space of ideas as well as subject matter experts who are called upon to give their opinion on a specific topic: “My intuition would be it’s going to be quite the opposite, where it seems much worse for the math people than the word people. What people have told me is that they think within three to five years, the AI models will be able to solve all the US Math Olympiad problems. That would shift things quite a bit.”
As expected, his response is contrarian. For years, the narrative has been that majors in science, technology, engineering and math (STEM) were the most promising fields to venture into as a prospective high school student about to embark on college education. In effect, the liberal arts were put on the back burner in favor of more analytical, marketable and practical majors. In other words, a major that will lead one into a lucrative job.
This advice was not misguided. In a 2025 analysis conducted by Georgetown University’s Center on Education and the Workforce, it found that the median salary of workers with a bachelor’s degree in STEM was $98,000 compared to $65,000 for humanities and arts majors.
As the argument went, a Computer Science degree was more valuable from an economic perspective than an English or History degree. But the tide is shifting. Tools such as Devin AI launched by Cognition Labs provide a look into the future of AI software engineering where coding agents respond to English language requests to conduct complicated code migration, refactoring, data engineering and analysis and bug resolution.
The need to structure text inputs – instructions, context, examples – to guide LLMs like ChatGPT, Claude and even Deven AI toward accurate, relevant and safe outputs has given rise to the field of prompt engineering (a job with a yearly median salary of $126,000). As the world continues to leverage LLMs to write software, generate images and video and complete complicated tasks, it becomes clear that the future belongs to those who can quickly and accurately find the write words to capture what they intend to achieve. In other words, the “wordcels”.
In the Fall of 2026, the Stevens Institute of Technology will launch a new AI degree program offering a bachelor’s degree in artificial intelligence. The program would start with a “foundation in computer science and mathematics, building essential skills in programming, algorithms and probability”, progressing into “more advanced AI topics like machine learning, natural language processing, computer vision and human-centered design” and concluding with a senior design capstone where students tackle real-world AI challenges with industry sponsors.
But the “AI fluency” the program aims to provide seems to be missing curricula that builds the reading, writing and communication aptitude that the industry is telling us is currently needed and will continue to be needed in the age of LLMs and AI. Skills which put English, History, and Philosophy students – liberal arts majors – at an advantage.
This isn’t to say that understanding the linear algebra behind reinforcement learning isn’t valuable, but it certainly doesn’t take precedence over the clear and important use of the written word with tools that are becoming more and more crucial for jobs across industry.
If we are to brace ourselves for Thiel’s warning and prepare the countless students that graduate from technical universities across the country for employment, we should make an effort to bring the liberal arts back into our professional curriculums (a step already being taken by universities such as Columbia College and Harvard College). Stevens’s AI degree, while pioneering and noteworthy, fails to accomplish this and would benefit if it were to incorporate more reading and writing classes. In addition, the degree claims it will teach students the ethics of AI to better prepare them for the safety and alignment issues surrounding AI tool development. But historically, the humanities were responsible for addressing such questions of justice, fairness and equality – major touch points in AI research.
Unless we take a more temperate approach to technical education and make room for
the humanities, it seems quite clear that the future belongs to the liberal arts.
