I’ve sold two AI companies and invested in over 200 more, including some of the biggest names in the space like Anthropic and OpenAI. I’ve seen the entire lifecycle of an AI product, from a wild idea on a whiteboard to a billion-dollar acquisition. And I’m here to tell you something the rest of Silicon Valley is too scared to admit.
Everyone is obsessed with the technical side of AI safety. They talk about alignment, catastrophic risk, and complex algorithms. That’s all important. But it’s not the real cost. It’s not even close.
The true cost of AI safety—the one that nobody is pricing in—is the human cost. It’s the talent, the culture, and the brutal, unglamorous work of building systems that don’t just work, but work safely.
The Myth of the Safety Algorithm
For years, the prevailing wisdom was that we could solve AI safety with a clever algorithm. Just find the right mathematical formula, plug it in, and the AI would magically behave itself. This is a dangerous fantasy.
I remember back when I was building RemoteTeam. We were using early AI models to help automate HR tasks. The pressure to ship was immense. Our investors wanted growth, our customers wanted features, and our competitors were nipping at our heels. The easy path would have been to grab an off-the-shelf model, fine-tune it, and deploy.
But we knew the risks. An AI making a mistake in a chat application is one thing. An AI making a mistake with someone’s payroll or employment status is a life-altering disaster. We had to build a culture of safety from day one, and that was far harder than writing any line of code.
The Six-Figure Cost of a Single Hire
Let’s talk numbers. The market rate for a top-tier AI researcher is astronomical. But what about an “AI Safety Engineer” or an “AI Ethicist”? These roles are even harder to fill. You’re looking for someone with deep technical expertise, a strong ethical compass, and the backbone to stand up to a room full of engineers and say, “No, we can’t ship this.”
Finding that person is a six-month, six-figure recruiting process. And you don’t just need one. You need a team. At a minimum, I tell my portfolio companies to budget for:
- 2 AI Safety Engineers: $250,000 - $350,000 each
- 1 AI Ethicist/Policy Lead: $200,000 - $300,000
- 1 Technical Program Manager for Safety: $180,000 - $250,000
That’s a million-dollar annual commitment before you’ve even written a line of safety-specific code. For a seed-stage startup, that’s a massive chunk of your runway. It’s a cost that most VCs don’t want to hear about, and most founders don’t want to admit they need.
Culture is Your Last Line of Defense
But even with the right team, you’re only halfway there. The real challenge is building a culture where safety is not a checkbox, but a core value. It has to be part of the DNA of the company.
This is where most companies fail. They create a “safety team” and then isolate them. The rest of the engineering organization sees them as the “no” department, a roadblock to shipping features. This is a recipe for disaster.
At MovieLaLa, which was acquired by Gfycat, we were dealing with massive amounts of user-generated content. The potential for harmful or biased content to be amplified by our recommendation algorithms was huge. We didn’t just have a safety team; we embedded safety champions within every product pod. These were engineers who were passionate about the topic and were empowered to raise red flags.
We also made one of our most controversial decisions: we tied a portion of every employee’s bonus to safety metrics, not just growth or engagement. It was a radical idea, and some people hated it. But it sent a clear message: we are all responsible for the safety of our products. It’s not someone else’s job.
The Coming Regulatory Tsunami
If the internal costs aren’t enough to convince you, the external ones will be. Regulations like the EU AI Act are just the beginning. The days of the Wild West of AI are over. Soon, having a robust, documented, and audited AI safety program will not be optional. It will be the price of admission to the market.
I’ve seen the draft regulations. They are complex, they are expensive to comply with, and they have teeth. Fines are calculated as a percentage of global revenue. For a company like Google or Meta, that’s billions of dollars. For a startup, it’s a death sentence.
This is why I’m so bullish on companies that are taking this seriously now. The ones that are building the internal expertise, the ones that are creating the right culture, the ones that are treating safety as a product, not a tax. These are the companies that will survive the coming regulatory tsunami. The rest will be washed away.
The Real Conversation We Need to Have
So, the next time you read a headline about a new AI breakthrough, ask yourself: what’s the real cost? What’s the human infrastructure behind it? Who are the people making the hard calls, the ones who are trading short-term growth for long-term safety?
These are the invisible heroes of the AI revolution. And it’s time we started talking about them. The future of AI depends on it. It’s not about building god-like intelligence. It’s about building human-centric systems that are safe, reliable, and worthy of our trust. And that is a cost I’m more than willing to pay.
Frequently Asked Questions
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.