Lessons from working as an optimist for critical AI

The work you do will influence how you see the world, and visa versa, are you prepared for it?

I have been for all my life what is called a techno-optimist, I see big possibilities and grand futures for products, companies and projects I take on. I then go on to do everything that needs to be done to make this future a reality. No pipe dreams here. A few years ago at the precipice of the newest AI wave, I was working on responsible AI. The field inherently requires you to be a critic, and anticipate problems before you anticipate possibility. This was interesting... This journey of being a “misfit”, and finding my footing taught me 2 big things:

Lesson #1: Innovation and optimism can happen even in the most critical fields.

This was perhaps my biggest influence to the teams I led and the companies I was a part of. I brought optimism and innovation aka my core values to the field of responsible AI. This meant that instead of thinking about how to find problems in AI models and products, I went a step ahead and thought about solutions. And so, my teams and I would not only find the problems but innovate solutions for our product partners. Whether it be configurable AI solutions to make safety customizable / tunable for their use cases or end to end evaluation solutions to help them understand their model performances quicker, with less resources, and machines in the loop.

This enabled us to not be the just people who found problems, but also allies that brought product thinking and help products GTM, much better for their users! We ended up building products with millions of QPS and dozens of launches, we were the collaborators not the police.

So, the lesson being, even if one’s core values are different from what is traditionally perceived to be the values required for a said big bet, it’s worth exploring what you can uniquely bring to the table. Leave a place better than you found it!

Lesson #2: It’s good to practice opposing views to expand your comfort zone through your career

When my son was just born, my husband and I joked that we were training this tiny “large model” with compute demands that were exceeding our quota. To this, my old boss said, “you aren’t just training him, he is also training you!” This is true for anything and everything we spend our lives doing. We are influencing the big bets we are a part of, the projects, the organizations, but they are also influencing us.

So, while I brought innovation and optimism to this bet, the time spent working on this problem that lent a critical lens to AI, helped me build this muscle. I now not only default to thinking about why something is great, but also why it would fail. This helps me build more robust solutions and anticipate failure patterns better than ever before!

So, the lesson being, be open to bringing yourself to the big bets, but don’t resist the learnings that come from expanding outside your comfort zone!