Clear Intention: The Survival Skill of the AI Age

"I knew I had just seen the most important advance in technology since the graphical user interface [GUI]."

‍
-Bill Gates regarding ChatGPT

"If you don't know where you are going, any tool will do."

‍
-David Allen, author of *Getting Things Done: The Art of Stress-Free Productivity*

‍

Asking the Oracle

"When am I going to die?" he typed, and waited uneasily for the answer. The year was 1999, and this was not ChatGPT, but a search engine called "Ask Jeeves" that tried to respond to questions, rather than search terms, with relevant information. The world wide web was new, increasingly available in libraries, universities, and internet cafes, and it seemed to know everything. Well, nearly.

It has only taken 25 years to give us an internet we can question instead of query. In 2006, Jeeves retired as Google became the world's information butler of choice. Then in 2022, ChatGPT arrived.

Gates' comparison of ChatGPT to the modern GUI is no accident. The ability of these systems to "autocomplete" not just single words, but whole paragraphs, with relevant continuations, had been around for awhile, quietly training in better ways on more and more digital text. It was when OpenAI added the conversation layer, allowing generations that grew up with email and text messages to interact through that familiar interface of typed replies, that ChatGPT had its "iPhone moment" of mass recognition.

‍

AI Hype and Amara’s Law

With it, much like the early days of the internet, has come much anxiety, speculation, and financial investment. Amara's Law, coined by the futurist Roy Amara in the 1970s, seems as prescient as ever: "We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run."

For now, amidst all the overestimation, AI's best uses are being found in just a few domains, including: marketing, to generate and test content; and software engineering, to write code. Lawyers remain uneasy about tasking digital paralegals with contract first drafts. Big bets are being placed however on wearable devices that can record everything you interact with in a day, summarise it, and even "coach" you on what to do next, like a kind of fitness tracker for your psyche.

‍

From Calculators to Co-Pilots: How GenAI is Different

If that comes true, are we done here? Will this vast collective super-consciousness soon obviate our squishy little brains?

More likely, it will supplement them. Whether you call it "augmented intelligence" or "cyborg working", millions of us are already "collaborating" with interactive generative AI. Unlike a tool such as Excel, where the same numbers in one cell produce the same results in another cell based on a mathematical formula, ChatGPT feels more like a co-pilot, assistant, and even friend than any software tool that has come before it. (I'm looking at you, Clippy.)

This is partly because it is based on statistics, allowing its outputs to resemble reasoning, but also making them prone to wild mistakes. In some ways, this is an advance on "deterministic" systems--like spreadsheet formulas or chemical reactions, where the same input produces the same output every time. But we humans are a third kind of system--we are intentional. We want things, have desires, dream up plans.

Because we have evolved alongside other intentional beings, trying to suss out whether they want to help us or eat us, we project intentionality onto other non-deterministic systems, bringing all the baggage of anthropomorphism to the familiar interface of text chat. This is why it is so impressive when AI correctly summarises the action points of a meeting, and so jarring when it gets something simple spectacularly wrong. The bubble of the metaphor bursts, and we are left to ask, "What were you thinking?"

‍

AI Agents and the New Garbage

Much promise lies in AI agents being able to take advantage of other software tools to fulfil our requests, demonstrating even more reasoning-like ability in choosing to search the web or scan our documents folder for answers, providing greater practical assistance than it could in isolation. Beyond a summariser or brainstorming buddy, agents integrated with our productivity tools can start to support some of our mind's executive function--the parts of our brain that deal with long-term thinking, planning, and remembering what matters.

In the end, though, while we can offload some of our reasoning to it, we can't offload our intentionality.

There is a phrase amongst those of us who develop deterministic software--garbage in, garbage out. It means your output is only as good as your input. Increasingly, as we use AI to generate code, it has become clear to us that the "garbage" we have to guard against is not bad data, but bad intentions. Because AI has no desires, our own desires, poorly clarified and articulated, lead to undesirable results.

‍

‍

The Power of Clarity in Fast Fast Times

Getting Things Done turbo-charges executive function by helping you comprehensively clarify, track, and prioritise your commitments--the fundamental units of intentionality that turn ideas into action. Because thinking, and especially clarification thinking, takes effort, the dream of "push a button and tell me what to do" technology has been around since the publication of the GTD book in the Ask Jeeves era of Palm Pilots and Blackberrys. Nowadays, in a haze of anthropomorphism and an Amara-esque hype cycle, it might seem we are getting closer with AI.

The truth is, we're finding out just the opposite. As Atharva Raykar put it in his blog on AI-assisted coding, "AI thrives far, far better in an environment in which a human would also thrive." This includes, "Clearly defined features, broken down into multiple small story cards," which is the software development equivalent of desired outcomes and next actions in GTD. While it can supply information, and even "reasoning", it can't supply clear intentions.

So, is AI coming for your job? Well, it is almost certainly going to change your job. Is it coming for your brain? Definitely. And to cope, you need to understand how to manage what makes you unique--your intentionality, with not just more tools, but a psychologically sound thought process that gives you confidence in your ability to execute, and not let things slip through the statistical cracks.

The internet sped up our expectations for knowledge work, and AI is poised to multiply this yet again. In the internet age, information became cheap and readily available, and many people are still drowning in it. Methodologies like GTD have been a life raft and a rudder for them and will continue to be in the underestimated "long run" ahead.

This is because in the AI era ahead, not only information, but "reasoning" with information is getting cheaper every day. Clear intentions, however, and the means to execute on them whatever the tides, remain priceless as ever.

‍