AI: automation, gratitude and flexibility

I'm deep inside the AI bubble and for a while I have had an occasional urge to “warn” others, here in this channel, about the scale of upcoming changes in the way we live, work, etc due to AI. Also a couple of people reached out, expressing anxiety about the uncertainty of the future, especially young people who chose to study software engineering, but now worried that it might get automated away, and they will end up without a job / in a much smaller job market. I've had some kind of existential crisis too.

I've been grappling with these questions for a while, but writing something helpful on this topic is hard for a few reasons. It's clear the changes will be significant, but how exactly things will change is unclear. Could you predict the modern times when electricity was just invented? Probably not. The bolder the prediction, the harder it is to accurately predict the time it would happen, because it's unnatural to think in exponentials, but also there may be bottlenecks that we normally don't think about. If a prediction is accurate, but far in the future, then it might be unactionable because there isn't much to be done now. For example, some predict that AI will erase humanity, and this might be accurate, but if it takes 100 or 1000 years, most people reading this probably won't have any influence on the outcome or be impacted, so why would I suggest you to worry about something you can't control. Besides, I'd sound like a lunatic and people wouldn't listen to me, so what's the point of writing such a post.

I expect our experience would be more like falling into the black hole (I'm not a physicist, so my analogy might be inaccurate): even if you see what's coming, it takes an infinite amount of time for this to actually happen, or it might happen only when it no longer matters (say, you die from stress). It would be a shame to waste your best years worrying about something that ends up not mattering. I spent a few months in such a state, and it wasn't great, but I did come out on the other side with a lesson that I'd like to share.

—

I have a decent imagination, so after thinking really hard about long-term AI implications about 10mo ago, I got into a state where I was thinking about “the end” a lot. Also, during that period, I had a few great moments with my family/friends, and under the influence of my state, I thought those moments will end one day too and not much time is left. Incidentally, this led me to experience those great moments stronger: an apple tastes better if you think that you have 10 apples left to eat in your life. You start appreciating things that are otherwise regular.

AI will change our lives. Some things will get much better and some others will get much worse. Some good things you have today will be gone and you will miss them, but you don't know which ones. Do you enjoy crafting quality code by hand? You have maybe 1-2 years left before models learn how to write good enough code faster than you, without sleep. Are you composing music for a living? You have a little more time, but the day will come that AI will disrupt that industry too. Some changes are indirect: remember when twitter was free of bots? That's over. A lot of things will change so you better enjoy good things in your life while you can.

Of course nobody will stop you from doing what you love, like coding, but it might become an unpaid hobby, so depending on your financial situation, you might not have much time for it. Yes, I know, my advice to appreciate what you have sounds like a clique, but this change in my attitude had an actual, noticeable positive influence on my life and got me out of the deep sad state.

—

On a more positive note, there is Jevons paradox: as technology makes a resource more efficient to use, the consumption of that resource tends to increase rather than decrease. For example, more fuel-efficient cars led to people driving more, not less. With software engineering becoming cheaper thanks to AI writing code for you, there will be more software, including software created on the fly. Some companies will choose to produce same amount of software with fewer people, and others will choose produce more software with the same number of people. While not everyone at Anthropic thinks this way, and the future has never been less clear, I'd like to think that engineers will continue to be in demand, as long as they can adapt to new ways of work.

Tractors automated farming, but farmers are still here. A farmer who rejected tractors probably went out of business because their competition adapted and became much more productive. If you associate yourself with a very particular thing that you enjoy doing, but it might get automated, like coding, then you might perceive that AI is making YOU irrelevant. Instead, I suggest you get mentally ready to be flexible and to adapt to new ways of work/joy. Moreover, AI will continue to evolve and change the way we work, so I expect each year to be different, so you will need to adapt continuously. I suspect open minded hustlers with adaptability/survival skills will come out on the other side of this evolution period much better than those that got stuck in their ways.

—

I will end on a philosophical note. The point of life is to live: if you spent a day worrying about the future and not experiencing the present moment at all, consider that you were dead that day. I suspect what will matter in the end is not the final destination, but the choices you made along the way, including those today. 

Life changes a lot. The way you think will be different in 5 years. AI will accelerate the wind of change. Get ready to evolve.