Lawyer Lin Shanglun's article: The college entrance exam results are out—it's really time to stop choosing arts majors in college

Humanities and social sciences bear multiple pressures in the AI era. University students will face difficult choices regarding future skills.
(Background: a16z analysis: AI costs halved, usage doubled, America entering an “adulthood delay” era for 30-year-olds)
(Additional context: Don’t blindly follow OpenClaw; crayfish AI is powerful but not necessarily suitable for you)

Table of Contents

Toggle

• 1. The “Hell Mode” of the Past: Structural Dilemmas in Salaries and Industries
• 2. Dimensionality Reduction in the AI Era: When “Text Processing” Becomes a Cheap Commodity
• 3. Taking the Legal Industry as an Example: Dramatic Efficiency Changes and Value Rebuilding
• 4. Not Just Lawyers: The Common Challenges Facing All “Text Workers”

Whenever college entrance exam results are announced, the debate over “liberal arts” versus “science” tracks always resurfaces. On this tech-obsessed island of Taiwan, the saying “Science track is the broad road, Liberal arts is hell mode”—though somewhat teasing—harshly reflects long-standing market realities.

However, in today’s wave of AI technology sweeping the globe, I want to present a more severe perspective: if the past liberal arts was considered hell mode, then in the AI era, this mode’s difficulty will be adjusted to an unprecedented level.

1. The “Hell Mode” of the Past: Structural Dilemmas in Salaries and Industries

Before discussing AI’s impact, we must understand why the liberal arts were once called “hell mode.” This mainly stems from two structural issues:

First is salary ceilings and industry demand mismatch. Taiwan’s economy heavily relies on tech and manufacturing, giving STEM graduates clear career paths and relatively high starting salaries. In contrast, traditional liberal arts fields like literature, history, philosophy, social sciences, and anthropology have limited positions (researchers, teachers, cultural workers), with most industries showing slow salary growth.

Second is skills substitutability and market saturation. Many liberal arts graduates end up in entry-level roles such as administration, marketing, planning, editing. While important, these roles’ core skills are oversupplied, leading to fierce competition, difficulty in standing out, and cycles of low pay and long hours.

The past “hell” was about economic returns and career development. But the coming storm poses a direct challenge to the core values of liberal arts.

2. Dimensionality Reduction in the AI Era: When “Text Processing” Becomes a Cheap Commodity

Many still discuss whether AI will develop self-awareness or dominate the world—like sci-fi movies. But these debates often overlook a more urgent reality: in the fields of “language processing” and “content generation,” AI’s capabilities have already leaped from catching up to surpassing.

Writing marketing copy, organizing meeting notes, drafting legal documents, translating multiple languages, generating news reports, creating stories—these skills once considered the core competencies of “text workers” are now being rapidly replicated, optimized, and even overtaken by AI. AI can digest vast amounts of data in seconds and produce coherent, meaningful text. This means that jobs focused mainly on “organizing, arranging, optimizing” text are rapidly losing their value.

3. Taking the Legal Industry as an Example: Dramatic Efficiency Changes and Value Rebuilding

In the legal field I am familiar with, this transformation is clearer.

Can AI fully replace lawyers? Honestly, not yet. Because lawyers’ work involves not just text processing but also strategic thinking, human insight, courtroom adaptability, and client communication. However, AI has already become the “hired army” of top lawyers.

Previously, drafting complex litigation documents took 8-10 hours of focused work—researching case law, organizing issues, finalizing drafts. Now, a lawyer skilled with AI tools can have AI produce a draft in minutes. The lawyer then spends the saved time on higher-value tasks like strategy, client relations, and business development. What used to take a day can now be done in a few hours with higher quality.

What does this mean? It means the legal service market’s “M-shaped” structure will accelerate. Top-tier lawyers can exponentially increase their capacity, handling more cases at lower costs, boosting their market competitiveness. Meanwhile, those relying on repetitive legal document processing will face severe squeeze—AI can do their work faster and cheaper.

4. Not Just Lawyers: The Common Challenges Facing All “Text Workers”

This efficiency-driven structural unemployment is not limited to law. Journalists, editors, translators, copywriters, research assistants—all professions centered on “producing text” face the same challenge.

In the past, a team might need several copywriters for different marketing needs; in the future, perhaps only one strategist skilled in instructing AI will suffice. Translating a book once took months of labor; now, AI initial translation plus professional proofreading will become mainstream, drastically reducing demand for pure translation labor.

As the marginal cost of “producing text” approaches zero, the market will no longer pay premiums for “simple writing.” The professional skills that once sustained livelihoods will suddenly become cheap or even worthless. This is not alarmism but a brutal reality happening to all text-based professions.

Returning to the initial question: in this context, how should we view choosing liberal arts? I must cautiously say that the risks of choosing traditional liberal arts have never been higher. If the “hell mode” of liberal arts was rooted in industry structure and salary ceilings, then the challenge of the AI era is a bottom-up attack on its core values.

This does not mean dismissing the value of humanistic literacy. On the contrary, in an age of information overload and difficulty discerning truth, critical thinking, historical depth, and understanding of human nature are more important than ever. But we must be clear: these qualities alone are hard to directly translate into market value. If after the college entrance exam you still choose liberal arts, you should know that in the AI era, your goal should not be to become a “text producer,” but rather a “thinker and creator” and a “tool master.” This path is undoubtedly more challenging, but only then can you find an irreplaceable position amid the coming structural upheaval, rather than being ruthlessly swept away by the tide of change.