We hired a team of hackers to break our own AI. It was scary and eye-opening. I'm taking you inside our first AI red-teaming exercise—what we did, the crazy vulnerabilities they found, and how we're fixing them. This is a raw look at what it takes to build responsible AI.
The Reality Nobody Talks About
A lot of the advice on AI red teaming rests on assumptions that made sense five years ago. The world has moved on.
The first thing to understand is that simplicity beats complexity every time. I've seen this play out across dozens of companies. The pattern is unmistakable.
Many teams spend months on the wrong path before realizing that the market doesn't care about your roadmap. Once they make the switch, everything changes.
Why Most Approaches Fail
Many common approaches to AI red teaming are flawed at the root.
The root cause is usually one of three things:
Copying what big companies do without understanding why they do it. What works for Google doesn't work for a 10-person startup.
Over-engineering the solution when a simple approach would work better. I've seen teams spend six months building something that could have been done in two weeks.
Ignoring the human element. Technology is the easy part. Getting people to actually use it is where the real challenge lives.
The AI Angle
Any discussion of AI red teaming in 2026 has to include AI. As someone who's invested in Anthropic, OpenAI, Scale AI, and Hugging Face, I have a front-row seat to how AI is transforming this space.
The short version: AI makes good practitioners better and bad practitioners worse. It's an amplifier, not a replacement.
Some teams use AI to get far more out of AI red teaming. Others spend heavily on AI that solves the wrong problem. The difference comes down to understanding what you're actually trying to achieve.
This connects to broader themes around EU AI Act, deepfakes, AI regulation 2026 that I've been thinking about a lot lately.
What's Next
Best practice in AI red teaming is changing fast. What worked last year might not work next year. That's both the challenge and the opportunity.
My advice: stay curious, stay humble, and stay close to the people who are actually doing the work. Read less thought leadership and do more experiments. Talk to fewer consultants and more practitioners.
If you're building in this space, the best time to get AI red teaming right is before you need to. Don't wait for a crisis to force your hand.
I'll keep sharing what I learn. This stuff matters too much to keep to myself.
Frequently Asked Questions
Q: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.
Q: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 234 startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
Q: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.