Shout it from the rooftops: One of the things I’ve learned in my many, many conversations with people who swear by AI is that for most people, AI adds only an incremental amount of value to their work.

This means that, especially in these, the earliest days of the productization of transformer-based and other AI models, using these tools is optional.

Companies would like us all to use AI as much as possible — and we will. But for most of us, in most instances, the way we’ll end up using it is that it will become an invisible part of the tools we already use.

The most obvious example of this is our increasingly algorithm-mediated news / media / entertainment environment.

Here’s a great piece from my colleague Meghan Bobrowsky about how Instagram’s Reels is now bigger than TikTok by revenue.

The secret to its success? Meta’s engineers cracked the code on showing us what we most want to see next — just like TikTok’s did, before them.

Here’s the relevant passage:

“Instagram’s algorithm previously had been built on a following graph, meaning it primarily showed users posts from people they followed—either friends, celebrities or creators. TikTok upended the idea of the following graph, instead showing users videos from accounts they didn’t follow and figuring out what they liked based on how long they lingered on each video.

Instagram had to figure out how to do that much harder task, too…. As people spent more time scrolling Reels, the algorithm got better at predicting what users wanted to see.”

AI companies need a return on their massive spend on data centers. All the coding tools + agents in the world aren’t going to pay for them, unless every developer on the planet is going to start spending thousands a month on Claude Code and its ilk.

Which means these companies are working furiously to make their tools more accessible and genuinely useful to you, the end user.

Early adopters and AI hustlebros want you to think you’re falling behind already.

You’re not.

Indeed, the Third Law of AI, which I outline in my book, is that in the future, AI will be a feature, not a product.

If you want to hear about the other 23 laws of AI, preorder my book How to AI, out January 27.

How to AI is the book on AI for the rest of us — a blueprint for the next decade of the development of this technology and our relationship with it.

Instead of a conventional book tour, I’m doing a “listening tour”. The concept is simple: I want to spend as much time listening to people as I want to spend lecturing them. If you’re a company, nonprofit, group of investors, book club of any description, hit me up! I’m going to be giving as many of these — virtually — as I can. I’ll also be up and down both coasts starting in late January, so there will be opportunities to talk in person.