Take More Risks

My mom woke up and called me from the other room. When I went to see her, she was trembling. I called emergency services.

That was how the worst day of my life began.

At the hospital, I was handed a form to sign before she went into emergency surgery. Everything moved quickly. My wife was out of town, and I was alone in the waiting room, asking ChatGPT to explain the surgery she was about to have.

Afterward, I was allowed into the room where the doctors were still trying to save her life. I took pictures of the monitors she was connected to and asked it to help me understand what I was seeing. I also asked about the terms I heard the doctors using so I could follow what was happening and know what to ask.

At some point, I stopped asking for explanations and asked for hope. I wanted examples of cases where things had gone well. I needed help thinking positively because I didn’t know how to keep going.

Those examples couldn’t tell me what would happen to my mom. But that conversation helped me through the next few minutes.

During the week that followed, I tried to help the doctors understand her medical history while making sense of what was happening. Having something to investigate kept my mind occupied when I could feel it heading toward dark places. Understanding a little more gave me a sense of control when so much was outside my control.

After my mom died, it helped me organize her funeral. I also used those conversations to think about better times, the things we had been doing together, and how she would have done things.

Then I kept returning to the decisions.

What had I missed? What should I have asked? I regret not using AI more beforehand to help me understand things. I can’t know what that would have changed.

For a while, I didn’t trust my own thinking. I looked at what had happened and questioned every choice leading up to it. I worried I wouldn’t be able to make a hard decision again.

Annie Duke describes a trap called “resulting”: judging a decision entirely by how it turned out. I was doing that to myself. Knowing the ending made it difficult to remember what I could reasonably have known beforehand.

I went to therapy. I was lucky to have my wife and some incredible friends by my side. Grief affects everyone differently, though, and you can’t expect people to be available every time you need them. There were still nights when I was alone with thoughts I needed to put somewhere.

I kept returning to those conversations, revisiting questions I had already asked and trying to explain what still didn’t make sense. Often, the responses reminded me of the care behind my decisions. It took time before I could see that myself.

I don’t think of these systems as people. I’m an engineer; I understand how they’re trained and that their answers can be wrong. But in those lonely hours, I had something to talk to. A companion is how it felt.

Gradually, I started asking what information I had at the time, what alternatives I could see, and what risks I understood. I had wanted the best for my mom. I couldn’t have planned for everything. I still carry guilt, but I no longer feel that I need to know how a decision will turn out before I can trust myself to make it.

That has started to change how I live.

I’m exploring jobs outside my comfort zone and moving several business ideas forward. AI helps me learn about possibilities and prepare for them. I can test ideas before committing more. I’m also more willing to speak up when I think something is wrong, even when there could be consequences.

I’m taking more ownership of my health. I had blood tests to better understand what was happening in my body, which led me to schedule an appointment with a preventive cardiologist. I track my sleep, recovery, and activity. AI helps me understand more and prepare questions for my care.

None of this has a guaranteed outcome. I have more confidence in my judgment, and I’m more willing to live with uncertainty. Waiting has consequences too.

---

These experiences shape how I hear the debate about AI.

In September, Anthropic CEO Dario Amodei called for slowing advances in the most capable models so safety work can keep up. In “We Must Pace the Frontier”, he distinguishes that from halting progress. My experience doesn’t settle how fast those models should be developed. But it gives me a reason to insist that the help people stand to gain stays part of that discussion.

I had people who cared about me. Not everyone has that support. I think about someone who has just moved to a new country, far from everyone they know. Finding people you trust takes time. Having somewhere to ask questions or talk through an unfamiliar situation can help during that stretch.

I want useful AI to become easier for ordinary people to access and use. I want us to keep improving it and making it safer. When leaders decide how this technology should develop, I want them to weigh the cost of delaying its benefits alongside the risks of advancing it.

The world we already have leaves people confused, overwhelmed, and alone. I’ve experienced mistakes and stubbornness in the human systems we rely on, too. Keeping things as they are deserves scrutiny.

I’m optimistic because this technology has already helped me. It gave my mind something to do during painful months. It helped me understand things and start building again. I want more people to have that chance, and I’m willing to accept some uncertainty as we learn how to make it work better.

On my Life page, each dot represents a week. There’s a quote from Guillermo del Toro there that means more to me now:

“When you take [into] account the dark to tell the light, it’s reality.”

An honest account of these months includes the worst day of my life, the people who helped me through it, and the unexpected support I found in conversations with AI. All of that is part of how I became curious about the future again, even while carrying the grief.

Those dots remind me that my time is limited. I want to spend more of it with the people I love and take better care of myself so I can enjoy that time. There are also places I want to see, experiences I want to try, and things I want to build. I don’t want to keep putting them off.

I still don’t know if everything will work out, but I’m willing to take the risk.