On AI, fig trees, and choosing the next branch

When I moved to California nearly a decade ago, I was bright-eyed, bushy-tailed, and excited to be at the mecca of technology. I have never regretted that choice, and it feels fortuitous that I’m here for this AI awakening. Living in San Francisco reinforces my notion that we're approaching the vague outline of a massive monument through the fog. We can’t make out the details yet, but there's an indelible sense of a new world order being created.

It is wise to be cautious—we're changing the course of human lives as we know them; however, I believe AI is a seismic shift that unleashes abundance. As Ezra Klein asks in Abundance:

What is scarce that should be abundant? What is difficult to build that should be easy? What inventions do we need that we do not yet have?

I think AI can help us answer these existential questions.

Until recently, I didn't have an intimate understanding of how AI even worked. Now that I do, it's almost comical how these LLMs we've anthropomorphized are essentially doing what computers have always done—manipulating bits and probabilistically predicting the next bit. This might sound obvious to most people, but it took me a second to grok that "intelligence" is not intelligence in the way we attribute to humans. Melanie Mitchell wrote in her book Artificial Intelligence: A Guide for Thinking Humans:

We should be afraid. Not of intelligent machines. But of machines making decisions that they do not have the intelligence to make. I am far more afraid of machine stupidity than of machine intelligence. Machine stupidity creates a tail risk. Machines can make many, many good decisions and then one day fail spectacularly on a tail event that did not appear in their training data. This is the difference between specific and general intelligence.

That isn't to say AI and LLMs aren't improving by leaps and bounds each day. There is a level of intelligence they will inevitably reach that will far surpass most humans. But this isn't a future I fear, for a few reasons. When AI automates the mundane tasks inherent to every field and industry, smart people can focus on solving higher-order existential problems like curing cancer or solving climate change, as opposed to creating yet another SaaS CRM.

AI may also spur a redistribution of wealth and status from the "knowledge class" to the "working class". I'm excited for a future where a nurse, a plumber, or a teacher is paid more than those of us staring at a screen all day. AI has been trained on data from books, videos, and everything on the internet, but there is still so much knowledge that hasn't made it online, particularly in the third world. Given how hungry AI and LLMs are for data, I'm hopeful that there are more incentives to digitize and preserve languages and cultures that might have otherwise languished.

The possibilities enthrall me. I have a tattoo of a fig on my arm, immortalizing a line from The Bell Jar that has always stuck with me:

I saw my life branching out before me like the green fig tree in the story. From the tip of every branch, like a fat purple fig, a wonderful future beckoned and winked.

The siren song of AI was too strong for me to resist, and for that reason, I’ve left my role at Temporal to work in AI. I'm deeply grateful for my time at Temporal. It’s a rare experience to work on a product that users adore, and it’s not just the product I’ll remember, but the people who made it so special.

For the next two weeks, I'm reading books, playing tennis, and being a lady of leisure because what's coming next is going to be a rollercoaster. I cringe as I say it, but stay tuned...