Goodbye Google

I have resigned from my position at Google. It was not an easy decision for me to make. Although I have enjoyed working with my colleagues and being rewarded for solving challenging problems, I believe that the rapid pace of AI development is not in the best interest of people right now. I firmly believe that AI progress is happening too quickly, and I have concerns about where it may lead.

Moving to a different project at Google would not have a significant impact, but I am still able to contribute my unique skills outside of Google. I am continuing to maintain Pernosco and rr, and I plan to investigate how AIs debug code and how debugging tools can assist in this process. My future work will focus on being unambiguously pro-human.

As a Christian and elder in Auckland Chinese Presbyterian Church, I take my responsibilities seriously and believe that addressing existential risks related to AI is morally imperative. While the potential benefits of AI are unclear, I am concerned about issues such as cognitive surrender, AI-induced psychosis and loneliness, power concentration, economic disruption, cybersecurity, and lack of accountability.

I am confident that AI capability will continue to increase, but I believe it is happening too rapidly for humans to keep up. The current rate of change is far too fast, and people cannot plan for the future the way previous generations could. I have had this decision in mind for some time, as I planned to go on a ten-day backpacking trip with friends.

I have worked on improving tools for hardware chip design, which I initially thought was relatively harmless, but now understand the significant impact these tools have on accelerating AI development.