Why AI Companies Are Racing to Recruit Philosophers and Domain Experts

Anthropic, the AI ​​security company behind Claude, has been one of the most visible companies in this regard. He spends some of his trillion-dollar valuation on people who have never written a line of code in their professional lives. The question worth asking is why.

The Talent Gap That No One Advertises

For years, the dominant narrative about AI has been that it will first replace routine cognitive work and then threaten more complex professional roles. What the narrative overlooked was a more pressing problem: The systems built were extraordinarily capable of pattern matching and text generation, but surprisingly poor at knowing what was important, what was true, and what was worth saying.

This is a question of knowledge. Knowledge, on the other hand, is field-specific, contextual and accumulates over years of practice.

An internet-trained language model knows that Shakespeare wrote Hamlet. A literary scholar knows what Hamlet means, why it endures, how it has been misread over the centuries, and what a contemporary production reveals about the concerns of our moment. These are not the same type of information, and the difference between them is crucial when building AI systems that humans must rely on to make important decisions.

What Do Domain Experts Actually Do?

The role of a subject matter expert in an AI company varies by field, but the common thread is this: They provide the ground truth that the system cannot establish on its own.

A medical expert reviewing a clinical AI tool evaluates whether the rationale is sound, whether edge cases are handled appropriately, and whether a clinician under time pressure can use the tool safely. This requires clinical judgment resulting from the treatment of patients.

A constitutional lawyer working on an AI paralegal asks whether the system understands jurisdiction, precedent, and the gap between what the law says and what courts actually do. They catch seemingly plausible errors that a non-expert would miss precisely because these errors seem so plausible.

A philosopher who works on AI alignment—making AI systems pursue goals that people actually want rather than technical representations for those goals—does the groundwork of figuring out what we want in the first place. This turns out to be one of the most difficult questions in the field.

Software Engineer Paradox

AI companies are also aggressively hiring software engineers, although their CEOs suggest that AI will soon write most of its own code. Dario Amodei publicly stated that artificial intelligence will soon be able to do the job of an entire team of brilliant researchers. OpenAI’s Sam Altman also made similar statements.

The recruitment of engineers continues regardless. Domain expert recruitment continues. The number of employees is increasing.

It’s simple when you see the resolution. As AI handles more routine tasks, it is shifting towards tasks that require irreducible human input, genuine understanding, contextual reasoning, and the ability to identify the right questions. More capable AI increases the risks for the remaining humans; it doesn’t drop them.

What does this mean?

The most durable professional asset in the age of AI is real, hard-earned expertise on a real subject. Experts on the subject are not competing with artificial intelligence. They are his evaluators, his conscience and his compass; They are people who can distinguish between an answer that sounds right and an answer that is actually correct.

Trillion dollar companies already understand this.