Everything will change/nothing will change

I am out in the real world this week (on vacation!) so I don’t have my usual exposure to engineers and engineering teams. I didn’t think I would write but this habit is about 15 years old now, hard to break! (As an aside - it’s an old adage but very true, that if you do a little of something, like writing, every day, it adds up. I looked back at my letters and I have roughly 3 books worth of writing. That’s not even every day! It’s just once a week. So, bonus life lesson: if there’s something really big, and hard, you want to have done in your life, start doing it, just a little, today. And some tomorrow…it’s hard to get the habit started but if you can do just a month, you will likely get it established and get going. Best time to plant a tree is 20 years ago, second best time is today, and all that).

I had a conversation with someone last week that’s been on my mind. They asked whether I thought AI was “real” and would “change everything” like the internet did. It’s an interesting (and timely) question! Thinking about it, I reflected on both the promises of the internet, and what has changed and hasn’t, in daily life. It’s weird to say, but the internet obviously changed a whole lot of things in our lives - we have connected mobile phones, search engines, social media, a very different cultural landscape, ecommerce, all kinds of things are different.

And yet, most of life is the same: we have houses, and jobs, and friends, and we eat at restaurants, and go to schools and hospitals and churches that are pretty recognizable to someone from 25 years ago. Parts of the culture would be disorienting and some parts even exotic and nonsensical (memes!) but most of it would still make a lot of sense to my pre-internet self. I suspect I’d be really excited about a lot, and unhappy about some.

I’m not a forecaster. My writing here is mostly about observations, patterns, system thinking. AI is probably as hard to predict as the internet was. But I remember a friend’s father, when we asked if the internet was “real” saying: “absolutely! The thing humans love to do the most is communicate, and this makes communication much better”. I think the same could be said, and more, about thinking - and AI is (so far) all about thinking and communicating.

So, it’s probably safe to say that AI will look like the internet did - everything will change, but lots won’t, somehow. Most predictions will be wrong. The big changes will probably be more surprising than we think. Things that we expect to happen (like we expected ecommerce) will happen but probably not the way we think. Money (investment) can distort the behaviors for a while (hello pets.com!) but the working patterns will emerge eventually (hello, Amazon). It will be an interesting ride, and no one really has a clear map right now - all you can do is pay attention, be flexible, and use the tools as they emerge (this, btw is another interesting recent observation - there are a large number of people I know who have strong opinions on AI but have never touched an LLM. It’s really weird, though I do remember folks like my parents having strong opinions about the internet before they had ever used it. Most of those early skeptics are as glued to their phones as anyone today).

One last observation: overwhelmingly, folks liked my writing last week vs the AI writing (thanks!) The sentiment seemed to be around the human connection, that I could write from experience, but the machine wrote platitudes and generalities. That’s an interesting data point. You could take it as “machines will never have experiences and be authentic” (which strikes me as roughly equivalent to the early complaint that ecommerce wouldn’t work because shipping and returns were hard and expensive), or you could take it as “we need to find a way to give machines authentic memories and experiences before they can do some tasks (which, in our analogy, would be like Amazon deciding “ok, let’s fix the cost of shipping”). Remember, we all like to ask “why not” questions. The better ones to ask are “what if”. “Why not” always points at a problem to solve. What problems do you see in the writing from last week?