Over the past week, a single question from Anthropic's interview process has been circulating through the AI talent ecosystem. The question is deceptively simple: "What happens to your commitment if the stock goes to zero?"
It's a cold probe. A stress test. And for anyone who has spent years examining systems under extreme conditions, it's a question that reveals far more about the company asking it than the candidate answering it.
Anthropic is not just asking this question to filter employees. They are building a team that can survive a scenario their own valuation models apparently account for. That's not culture. That's risk management wearing a philosophical mask.
I've audited protocols that collapsed under precisely this kind of misaligned incentive structure. The failure mode is always the same: the team that believed in the mission stayed. The team that joined for the exit was gone before the liquidity dried up.
The question Anthropic is asking isn't unusual in the context of cryptographic systems. It's the same question a DeFi protocol should ask its validators before a market crash. What's unusual is that a major AI company is asking it openly, to every candidate, before the crash arrives.
The Context: A Company Built on Contradiction
Anthropic was founded with an explicit mission: build AI systems that are safe, interpretable, and aligned with human values. Its legal structure reflects this ambition. Registered as a Public Benefit Corporation in Delaware, Anthropic is legally obligated to balance shareholder interests with broader societal good.
That's a meaningful commitment. And the "stock to zero" question is a direct test of whether a candidate believes it.
Based on my experience auditing protocols with mission-driven teams, the structural design matters more than the stated intent. Anthropic's PBC status is not a marketing claim. It's a legal constraint that shapes every business decision, including who gets hired.
The question itself suggests the company has already internalized the possibility of total equity collapse. That's not a standard screening mechanism. It's a confession about what the management team believes the future might hold.
Anthropic's valuation history makes this confession even more pointed. The company has raised over $7 billion since its founding. At various points in 2024, its valuation was reported to be in the $60-70 billion range. Asking candidates whether they would stay when the stock goes to zero at a company worth billions is not a hypothetical exercise. It's a scenario that management appears to take seriously enough to embed in its hiring process.
The Core: What the Question Reveals About Structure
Let's look at the question as a system parameter. Every company has a "retention threshold" — the point at which core team members leave when things go badly. In crypto, we see this constantly. Protocols with strong token incentives attract mercenaries who exit at the first sign of protocol stress.
The core question for Anthropic is: what is the retention threshold of their workforce under extreme downside scenarios? They are attempting to set that threshold as low as possible by screening for candidates who value mission over financial upside.
That's rational. It's also revealing. The company is effectively telling candidates that its equity may not be worth anything eventually. This is not the signal that a high-growth company typically sends when competing for top talent against OpenAI and Google DeepMind.
The signal is consistent with something else: Anthropic is building for a scenario where the AI market contracts severely. Perhaps they believe it will contract. Their internal models may already have stress-tested the collapse of AI investment.
From a capital allocation perspective, this is defensive positioning. Anthropic is not assuming its current valuation is the peak. They are building a team that can survive the trough.
The bottleneck isn't the infrastructure. It's the ability to retain the people who know how the system works when the system is failing.
In the AI industry, that's a different approach than most other companies. OpenAI offers massive compensation packages. Google DeepMind offers stability and brand. Anthropic offers a question — and that question filters out a significant percentage of the talent pool who believe their equity should be worth something.
This is also a governance structure issue. In DeFi, I have seen what happens when token prices collapse and the team structure disintegrates. The protocol doesn't fail because the code is buggy. The protocol fails because the team that knows how to maintain the code is gone. That's why I've always considered the human capital risk to be greater than the technical risk.
Anthropic's approach is to engineer out that risk by building a team that values the mission enough to stay when the equity is worthless. That is the right instinct, but it carries a downside. The question also selects for a specific kind of person. The person who says they would stay when the stock goes to zero may be the person who doesn't understand the difference between equity value and mission value.
The Contrarian Angle: Screening for the Wrong Behavior
Here's the problem with the question. It screens for declared commitment, not actual commitment. Anyone who has interviewed candidates for critical roles knows there is a significant difference between what someone says in an interview and what they do when the pressure is real.
The question also screens out a specific type of talent: the one who is pragmatic about their own financial survival. In a field where top researchers can command high salaries and equity packages, asking them to pledge commitment to a scenario where their equity is worthless is a signal that the company may not be prepared to offer competitive compensation.
In crypto, we call this "denial of the token price." When a protocol starts telling its community that the token's price doesn't matter, it's a signal that the protocol's fundamentals are failing. Anthropic's question could be interpreted in the same way.
There is also a more serious issue. This question may be screening for the wrong type of alignment. If the team is full of people who will stay when the stock goes to zero, it means they are staying because of the mission. But what if the mission itself is wrong? What if the AI safety mission is built on flawed assumptions?
The best security researchers are not the ones who are committed to the protocol. They're the ones who are committed to finding the truth, even if the truth is that the protocol is broken. The same principle applies to AI. The best safety researchers are not the ones who will stay with Anthropic when the stock goes to zero. They're the ones who will tell Anthropic that their approach to AI safety is fundamentally flawed, even if that causes the stock to go to zero.
By asking the "stock to zero" question, Anthropic may be screening for people who are committed to the company's survival rather than the mission's integrity.
The Workforce Component
There's a specific angle here. It's about the distribution of the equity. A significant portion of Anthropic's workforce is likely to be compensated with equity. The company's equity is their savings. When you ask them if they would stay when the equity goes to zero, you are asking them to accept a total loss of their compensation.
This is not a trivial question. For a candidate with a family, the question might be disqualifying. This is the likely outcome for the most financially responsible candidates. They may have obligations that require them to consider the financial side of the equation.
The candidate who can honestly say they would stay when the stock goes to zero is likely to be a candidate who has a lower financial burden. This is a selection bias.
The Signal to the Market
The question to the market is a negative signal. When a company asks its employees to accept that the stock goes to zero, it is implicitly acknowledging that the stock could actually go to zero. This acknowledgment can trigger a downward pressure on the stock price.
The employees are the internal investors. If they are being told that the stock is going to zero, they will sell their stock when they can. They will also not recommend their friends to join the company. This is a self-fulfilling prophecy.
A rational employee would be concerned. They would ask why the company is asking this question. They would see that the company is signaling that they do not have confidence in their own stock.
The market corrects. The code remains. The same principle applies to the stock price.
Takeaway: The Metrics That Matter
The question is not the problem. The problem is that Anthropic is using a single question to determine the alignment of its employees. This is a blunt tool for a complex problem. They need to get the retention rate after a market downturn. They need to get the turnover rate when the stock drops by 50%.
Those metrics are the real test. They are not asking the question. They are measuring the behavior. That is the difference between a company that is preparing for a downturn and a company that is just preparing.
The takeaway for any organization is simple. Do not ask a candidate whether they would stay when the stock goes to zero. Instead, measure the actual retention rate when the stock goes to zero. That is the only metric that matters.
The question is a signal. The market will eventually judge the answer. The code remains, and so does the question.