"Talk to me about Chinese food." ChatGPT spits out an answer and it's pretty good. And then, crazily enough, I ask it more about Szechuan food and it answers. The darn machine answers my question. I talk with it for about 10 more minutes about more dishes before stepping back. The single emotion that consumes me is none other than wonder.
That was back in Nov 2022 when ChatGPT first launched and now my feelings around AI (and maybe a lot of folks) have shifted compared to back then. LLMs are powerful beyond measure and we've seen the impacts on it already in the everyday. ChatGPT alone is the fastest company to reach 1 billion users in a mere 3.5 years [1] and it goes without saying that the world is obsessed with AI and how to use it.
I titled this anxiety but after looking it up, it's a stretch for me personally to say I'm experiencing anxiety as defined by the American Psychological Association [2] as:
An emotion characterized by apprehension and somatic symptoms of tension in which an individual anticipates impending danger, catastrophe, or misfortune.
There is a line later on that I find to resonate:
Anxiety is considered a future-oriented, long-acting response broadly focused on a diffuse threat.
While my initial encounter with AIs was one of wonder, when I examine my headspace and what consumes my mind, it's really just one thing now, our good friend AI. The thoughts most often in my head are:
These thoughts aren't bad, but the real emotion inducing thought is probably the last point. What's going to happen with AI in the future and how can I keep up?
In software, change is a given. The nature of the craft is that you can use your knowledge to create tools to solve the pain points you encounter while coding. At a building block level this is true as new languages, software and tooling are constantly being created. What's different this time around is the pace of change that AI has enabled. [3]
Getting personal here, change is hard. Even if it's good, it always takes personal effort to overcome the inertia of the incumbent. I in particular have a hard time with constant change; I love going to the same restaurant and ordering the same exact thing because I have predetermined knowledge that the result will be great (hopefully). The nature of change is that we have to venture into the unknown without any guarantees of the outcome.
As it relates to AI, the pain of a constantly changing landscape is even more painful than traditional software because AI isn't purely deterministic. Improvements or changes are generally incrementally better, but hard to prove for each and every person's application. To truly know if something is better, you have to try and test things yourself on a wealth of tasks to know for your use case whether it'll truly help. And with new models coming out every few weeks, this can be exhausting to keep track of.
So how can I handle this change better? I think part of it is grounding myself on the simple truth that change is outside of my control. This is true for life as much as it is for the industry and if I try to place on myself the expectation that I fully understand every change of the wind as it comes, I'll probably find myself sleep deprived and paranoid. The other strategy is to focus on what's important in my life. I recently had my second child join my family and it's been a great joy though tiring. And while I'd love to keep up with all the happenings in the industry, ultimately I'd rather have spent the quality time tending to my newborn even if it means being behind.
At the end of the day, I just need to try my best in the time I give to these academic pursuits while reminding myself that there's more to life than AI.