Cultivate your creativity to survive AI
I had just finished recording the final video of my personal productivity mini-course when I turned on the camera and ad-libbed from the heart. I was unprepared for what came out.
Update: this post lead to an interesting conversation on LinkedIn, followed by another on the AI coaching forum, created by Peter Kaminski (check it out!).
Skip the actual video if you want — I cover the important ground after it:
So where did that outburst come from?
Clearly that had been building for several months. It probably started during my in-public experiments in integrating GPT into MyHub.ai, which ended last February (TL:DR; I was underwhelmed). And it built further as I created my course, revisiting my 10+ year old approach to personal productivity to boil it down to a Framework that anyone could apply.
And by the time I was putting the finishing touches on my course, I had recently listened to Artificial Creativity, a recent podcast episode from Douglas Rushkoff which puts the current moment in a historical context of industrialisation stretching back to the first printing press (my notes).
Rushkoff’s piece echoed something I first heard many years ago in CGP Grey’s Humans need not apply: that a robot which is 10x worse than a human will still get the job if it’s 100x cheaper. And LLMs generate text many thousands of times cheaper than humans. Which is a problem, because creating text is how we both develop and share our new ideas.
The first creative work being passed to the robots are, unsurprisingly, the entry-level work given to young graduates as they enter their industry. As Rushkoff points out, however, those jobs are where today’s experts started: “when we relegate all this engineering, reading, and writing to machines instead of human apprentices, how does the next generation arrive at a level of mastery?”
how does the next generation arrive at a level of mastery? Human expertise is being hollowed out
Human expertise is therefore being hollowed out, leaving tomorrow’s humanity reliant on AI tools that mine the wisdom of previous ages where humans actually nurtured their creativity and skills. The only people who’ll profit are those who own the AIs.
Even worse: unlike young humans learning their craft, LLMs will not grow. They cannot master a craft and then innovate, pushing their field forward the way human artists and scientists stand on the shoulders of those they learnt from as they create something original.
Instead, AIs can only restate and recombine past innovations: you can replace a session musician with an AI and tell it “to come up with a solo “like Slash” would play… the AI will dutifully listen to all of Slash’s solos and construct the most average Slash solo ever. More typically Slash than Slash himself” — but it can’t innovate like Slash, or any other artist.
So not only might human expertise disappear, innovation itself might grind to a halt. As Rushkoff pointed out in a much earlier post, an LLM’s response is “based on an “average” of everything already said … pure reversion to the mean. Only a human has the ability to … [generate] Novel ideas that challenge existing orthodoxies unrecognizable to a language model” (Let Them Eat AI, May 2023). Relying on LLMs — as we will when we no longer have skills or expertise — will make “our most probable futures even more inevitable”.
Unless, of course, humans still manage to develop their skills and keep innovating despite the economics of LLM-generated content.
What does this have to do with personal productivity?
All that, above, was the context stewing in the back of my mind as I polished the “TTK Framework”, which integrates several techniques in Task, Time and Knowledge management.
In Knowledge management, in particular, there’s a pipeline to help “tame the firehose” of content coming at me every day (see Manage the Firehose or it will manage you, June 2015). The pipeline ensures you:
- are intentional about what you read, ensuring you spend 80% of your time on the best 20% of the content entering the pipeline
- learn more from what you read: you take notes as you read to better embed the knowledge in your brain (not just your second one) and relate what you’re reading to your experience, perspective and problems
- store that knowledge in your library so that you can both easily find it and connect it to everything else you’ve read or thought
- final step: revisit that library and reflect on its contents to write something original.
In the lesson itself I actually say that the final step is optional: while personally I love writing, it’s not for everyone. But a few days later, speaking to camera without notes, I contradicted myself. Why?
If you follow the above process for a while you’ll find yourself sitting on an entirely unique library of ideas (here’s mine). Nobody else in the world will have that particular library, with those particular thoughts about those particular resources. It’s as unique to you as your DNA (assuming you’re not an identical twin).
So if you do take that final step and mine that library to write your own content, it just might be something noone else could have come up with. Maybe only you could have had that thought. You don’t have to publish what you write, although it’s likely you’ll impose better quality control on what you share with the world, but you do need to create it.
Maybe only you could have that thought
The entry-level jobs, after all, are only the first that will disappear. AI will eventually come for yours if it mainly consists of rehashing past experience, something AIs can do 100x cheaper and faster than you (and maybe not even 10x worse).
While we cannot out-produce LLMs in terms of speed and scale, however, we can out-create them in terms of innovation and value if we learn to innovate consistently. And that, I now see, is probably why I launched the TTK Framework this year, over a decade after I developed it for myself.
So if you want to stay relevant professionally, you’ll need to be unique, original and creative. You need that pipeline.
Followups
This article, of course, is an example of the above process, building principally on the content I’ve read, annotated and stored on my Hub with the tags “creativity and ai”, and “productivity and creativity”. Medium mangles these collections’ URLs when I use them as hyperlinks (which is nuts) so you’ll have to copy and paste the following into a browser:
- creativity and ai: https://myhub.ai/@mathewlowry/?tags=ai&types=like&types=do&types=think&timeframe=anytime&quality=all&tags=creativity
- productivity and creativity: https://myhub.ai/@mathewlowry/?quality=all&types=like&types=do&types=think&tags=productivity&tags=ai&timeframe=anytime
I wrote about the TTK Framework in this post, while my Hub’s About page includes my newsletter subscription form (editions are tagged #newsletter).