Attention Is All You Have

In the agentic era, leadership means designing what your machines notice, optimize, and repeat

Now that the machines do the doing, what’s left?

In 2017, Google published the paper that introduced the Transformer architecture, the basis of every modern LLM (the “T” in ChatGPT). Its math is dry, but the title is accidental poetry: Attention Is All You Need.

I doubt the researchers knew they were writing a job description. Yours.

As execution gets cheaper, judgment gets more important. You now sit one promotion up the stack, closer to the work that was always hardest: deciding what deserves the machine’s next move.

AI systems, no matter how capable, can’t prioritize everything at once. They run on data, constraints, and objectives, explicitly or from culture and habit seeping in. The question isn’t whether there is a steering force. The question is whether it’s applied on purpose.

That force is attention.

Tune a customer support AI for speed and the tickets close fast. Tune it for resolution quality and the agent slows down to listen. The system follows your attention, whether you encode it cleanly or by chance. The defining characteristic of your leadership becomes the sheer naked quality of what you choose to focus on. Easy tasks let you skip the choice, but the future will have plucked them all. You’ll have to prioritize.

We spent over a century trying to eliminate complexity in organizations. The whole premise of Taylorism (also called scientific management) was to break work into parts, optimize each piece, and squeeze out as much of the variance as possible. That worked for the low-hanging fruit, the monolithic processes where one size really does fit most.

And then we all reached for more, and more, and here we are: galloping away from standardize-and-scale toward personalization, autonomous action, and systems interacting in ways no single mind can fully map. Complexity is ballooning.

Ashby’s Law of Requisite Variety, an enduring classic from 1956 (the same year that gave us the term Artificial Intelligence), says that a control system must be at least as complex and as capable as the system it governs. Scale up what a system can do and you must scale up what steers it, or the steering stops being steering and becomes weather. You inherited a controller built for a smaller century; playing at tomorrow’s level requires self-amplification.

Attention has always been valuable. It is what turns time into progress. What is new is how fast it now compounds. We know what compound interest does to a dollar set aside in 1956. We’ve yet to learn the same respect for an hour of well-pointed attention today, though the term is shorter and the bank requires no signature. The leader who waits will be steered by the weather.

What is new is how fast attention now compounds.

To manage agentic systems is to design attention scaffolding; the systems you don’t choose will be chosen for you by people whose incentives are not yours. Misapply your attention and you’ll end up busier than before. Build feedback loops and you’ll sharpen your judgment instead of dulling it.

Want to lead hybrid teams of humans and agents? You can’t book-learn that muscle memory. Start practicing systems design on what you know too well: the copy-and-paste drudgery. Resist repetition unless it’s an investment in less repetition next time; teach a machine instead so your attention can work for you while you sleep.

Make your attention work for you while you sleep.

As your AI agents begin to read between the lines of your life and your business, your role is not to do more. It is to see more clearly. To decide what matters. To guide complexity without pretending you control it completely. And to pay attention to your priorities. Literally.

Not everyone will recognize the agentic era as an opportunity for compounding attention. Many will let theirs wander to whatever is loudest. That creates an opening for those who choose differently.

We are early. Very early. Whether that’s good news depends on whether you’re running towards the future or away from it. The leader of the next decade is the one who knows where their attention is going, why it is going there, and what they would have to stop doing to get it back.

Attention is all you need.

Thank you for reading — and sharing!

I’d be much obliged if you could share this post with the smartest leader you know.

👋 On-Demand Course: Decision-Making with ChatGPT

The reviews for my Decision-Making with ChatGPT course are in and they’re glowing, so I’ve opened enrollment for another cohort and tweaked the format to fit a busy schedule. You’ll be able to enjoy the core content as on-demand recordings arranged by topic and then you’ll bring your questions and I’ll bring my answers in a live 2 hour-long AMA with me on Monday June 8 at 8 AM to 10 AM Eastern Time:

If you know a leader who might love to join, I’d be much obliged if you forward this email along to them. Aspiring leaders, tech enthusiasts, self-improvers, and curious souls are welcome too!

🗞️ AI News Roundup?

These days, I’m putting AI news stories (ranked from boardroom to bizarre) in their own separate issues. Subscribe now so you don’t miss the next one!

The Human-Agent Orchestrator Book

A version of this essay appears as the afterword for this book, which launched yesterday:

Consider this book a business-oriented treasure map for being deliberate with building your attention scaffolding and shifting your thinking for human-agent leadership. Here is the link: amazon.com/dp/B0GZDZ178G

Course Promo codes

My gift to subscribers of this newsletter (thank you for being part of my community!) is $200 off the list price of my course with the promo code SUBSCRIBERS. If you haven’t subscribed yet, here’s the button for you:

If you’re keen to be a champion of the course (you commit to telling at least 5 people who you think would really get value out of it) then you are welcome to use the code CHAMPIONS instead for a total of $300 off — that’s an extra $100 off in gratitude for helping this course find its way to those who need it. (Honor system!)

Note that you can only use one code per course, the decision is yours.

P.S. Most folks get these courses reimbursed by their companies. The Maven website shows you how and gives you templates you can use.

Forwarded this email? Subscribe here for more:

This is a reader-supported publication. To encourage my writing, consider becoming a paid subscriber.