Remember 2020? So many of us thought that was going to be the big one for the way we work. I thought it would be the year that changes productivity forever. And in many ways it was; it forced us to spring clean our dusty attic of beliefs about what makes people productive. Decades of in-office 9-to-6 culture blown out of the window! It was a year that showed us just how resilient the human spirit is and how robust our ability is to get things done. The big discovery that year was how little productivity could change if we put our minds to it.

And that’s how 2022 snuck up on us. You might not have seen the flash of lightning if you weren’t looking out of your window… 2023 is when you’ll hear the boom. That’s when the thunderclap will be too loud to ignore.

Image created by the author with Midjourney.

The year 2022 was remarkable for the sheer amount of showing off we’ve all witnessed on the frontiers of productivity.

I usually write about data science and AI from the point of view of leaders and professionals who use data as raw materials for solutions. But from the perspective of everyone else, many of those solutions simply looked like a slightly smoother or more efficient application. The mathy old guard would casually poke fun at all the funding-grubbing neophytes who felt the need to tell everyone their solutions were “AI-fueled” when it really shouldn’t matter. Does it work x% better than the traditional solution? Great. That’s pretty much all anyone needed to know about it.

They’re not AI tools, they’re proofs of concept in a productivity revolution.

But as 2022 comes to a close, there’s something every data non-professional does need to know. There are a lot of “AI tools” popping up like mushrooms in your social media feed. If you’re ignoring them because you’re not “an AI person” or not excited by tech, you’re missing the bigger picture. They’re much more than AI tools. They’re proofs of concept in a productivity revolution.

So, if you’ve been ignoring all this AI stuff, let’s catch you up.

Phase 1 — Research AI

AI researchers invent general-purpose algorithms and approaches for others to use. They’re not here to solve your specific problem using your specific data, they’re making solutions possible—theoretically. Before 2015, almost every AI course out there was focused on research AI. After many years of study (not just one course!), students who got all the way up to a graduate degree learned how to push the envelope on the kinds of problems that might be solved with AI, whether or not it was feasible to actually use these solutions in a business setting.

Image created by the author. Learn more about it here: http://bit.ly/quaesita_fail.

Phase 2 — Applied AI

Once the AI researchers had created methodology that could — in theory! — be used to solve business problems, it was time for applied AI engineers to enter the ring. Applied AI involves a focus shift to using research inventions to automate something in a business at scale. When I started at Google in 2014, I hadn’t seen a single course that focused exclusively on applied AI and intentionally left research AI to the AI researchers… so I made the first comprehensive one, Making Friends with Machine Learning. Check it out on YouTube if you’re interested:

The applied AI revolution was about taking a new approach to programming, supplementing traditional software engineering to automate a wider range of tasks. But from an outsider’s perspective, not all that much was different. As unsexy as it is when it’s spelled out, what everyone in tech was excited about was that programmers could now tackle a wider variety of tasks more effectively.

Programmers now could tackle a wider variety of tasks more effectively.

And if you were skilled at applied AI engineering, you could automate all kinds of things in your personal life too, giving yourself an incredible jetpack of personal productivity. Which is exactly the same story that traditional software engineers tell about all the ways they made their own lives better by applying a bit of code here and there.

Phase 3 — Productivity AI

And then 2022 swept onto the scene. Progress in both research AI and applied AI came to a boil. All that horsepower could finally be applied to building all kinds of consumer applications including one that’s in a class of its own because of how it feels to the user: personal productivity.

If you don’t work in a tech/data role, there’s no particular need for you to bear in mind that there’s any AI involved here. Too much focus on the term “Artificial Intelligence” might send you to science fiction for guidance and set your expectations so high they’re in outer space. Instead, forget the term “AI” and simply focus on the “productivity” bit. Think of it like this:

By 2022, the tech got good enough that there was a profit to be made in offering consumers personal productivity tools that were better than anything a yurt full of traditional software engineers could have built.

The thesis is simple, and it’s not AI-specific at all: you wish you could do something all by yourself but you can’t. Maybe you don’t have the talent or the skills or the time. You could pay money to hire someone who can do it for you — if you can afford their services. No? Well, perhaps you have coding skills to automate it yourself? No? Too bad, you’ll just have to sit this one out…

…Unless a company can offer it to you at a rate you’ll find palatable. The lower the rate, the more people will sign up.

Image created by the author with Midjourney.

Boom! Relative to where 2022 started, they can offer it to you on more tasks than ever.

The productivity revolution has been brewing for a lot longer than 2022.

To be crystal clear, the revolution has been brewing for a lot longer than 2022. Here’s a personal productivity example. This revelation may rattle the 112K+ of you who follow me for my writing (sorry!) but I’m not all that good at writing. Luckily, I’m an excellent speaker. That’s why all my first drafts are spoken, not written. Otherwise, I would be instantly beheaded by the dreaded blank page.

My blogging productivity is so unusually high precisely because I don’t write. I ramble all my first drafts into a recording app and only then do I polish them up. Naturally, I don’t have the time to transcribe all those spoken drafts myself, so I use a speech-to-text AI tool. I’ve been doing that far longer than I’ve been blogging (hope this isn’t too much of a shocker). I started by coding my first personal speech transcription tool up myself, but was relieved to see consumer options show up fairly soon, so I switched to an app. These days I often write by using the voice typing feature on Gboard.

In the past, you’d have a shot at giving yourself a productivity boost… but only if you were code-and-data-savvy.

Which brings me to my main point: in the past, you’d have a shot at giving yourself a productivity boost… but only if you were code-savvy. Then you’d be part of the secret cabal who could automate things for themselves. Today, more and more companies are stepping up to do it for you. And that is the big revolution.

But it’s still a flash of lightning because 2022 was the showy phase: Look! Look what we can do!

The year 2022 was remarkable for the sheer amount of showing off we’ve all witnessed on the frontiers of productivity. The revolution isn’t even on GitHub anymore — the everyone-friendly interfaces for AI-powered productivity have landed!

The endgame here is productivity.

Among the splashiest are OpenAI’s DALL·E 2 and ChatGPT, which audaciously offered millions of users the opportunity to tinker with a graphic art assistant and writing assistant respectively. Other recent darlings that have dazzled social media are Midjourney (which I used to make the illustrations for this post) and Lensa (example selfie below). Hang on, productivity tools? But aren’t these AI art apps? Look past the curlicues and you’ll see it. The building blocks are the same, but the attention the tech is getting is on another level.

Adjusting the author’s selfie with Lensa. Unedited original on the left, gently Lensa-improved version in the middle, all face retouch options set to max (yikes!) on the right.

Google has, of course, been steadily improving our productivity tools each year and it constantly amazes me how quickly people learn to take them for granted. To recalibrate my own sense of wonder, I daydream about explaining pretty much any innovation to someone in the 1950s. For example, Google Translate allows you to have a real time voice conversation across multiple languages, no translator required. Half a century ago people would have laughed at me for suggesting it would become part of everyday life. But today, it’s the most natural thing. We’re so used to it that it’s easy to forget that it’s very much an AI productivity tool. The AIness of it is so quiet relative to ChatGPT.

The everyone-friendly interfaces for AI-powered productivity have landed!

While many writers dissect the most attention-grabbing AI applications from 2022 as revolutions in art and creativity, I suspect that AI art is a means to a different end. If the endgame here is actually a personal productivity revolution, the pieces fit: not only does the tech have plenty of building blocks that could be recycled, but art is the perfect way to make people take notice (and glamorous selfies are a fabulous way to make sure all your friends know about it too). If it’s possible to build a tool that helps someone without much talent write and illustrate their own blog post, then imagine what else is possible (after paying a subscription fee, perhaps). And in the very near future, we’re likely to see it in full bloom.

Mech suits! Mech suits and jetpacks for everyone! Happy 2023! Image created by the author with Midjourney.