In this post, we kick off our second-half 2026 series on human thinking in the age of AI. Stay tuned for more.
Intro: AI is not the elephant in the room; it is an elephant that can fly
AI is a fundamentally different force.
I have been following AI progress diligently, and I still find myself—on a weekly basis—standing in front of my computer, yelling at it: “You are a genius.” I’m lucky to know enough about programming and have been using it intensively, averaging 2-3 hours a day for the past year. I feel I am experiencing the future of work, since programming is 3-5 years ahead of most other professions.
We are dealing with an elephant. It was in the room, it has left the building, and now it is flying. The “challenges” and the “opportunities” are immense.
When generative AI entered the mainstream about four years ago, we were in awe. The elephant was in the room, and people felt differently about it. For most, it was a smart search; for some, an email assistant; for others, a visual creator, etc. Later, the elephant left the building, generating 10x value in some tasks (say, in programming, analytics, or digesting large amounts of data).
This year, with the rise of Claude Cowork, OpenAI Codex/Work, and other agentic platforms, capabilities that once felt out of reach are becoming real. It’s not just about moving faster — it’s about working in entirely new ways.
A Thinking Harness: The Key to Cognitive Flight
To safely fly a giant, we need a support system, a saddle, and a steering mechanism to ensure long-term stability. In short, we need a harness.
The elephant is massive, can work across vast amounts of information, and possesses immense raw thinking power. If you simply point it at a task without guidance, it might crash through boundaries, make up facts out of thin air, or wander off entirely.
You do not command an elephant with rigid rules and switches; you have to learn how to guide its colossal cognitive force without crushing its creative spirit.
For a human thinker, a harness should feel like a neutral extension of the mind. It is about intellectual leverage. When built correctly, the harness does something spectacular: it makes the heavy lifting feel weightless. It lets you fly in the direction you choose.
Towards a Thinking Harness
While researching this post, I had the chance to explore the design of elephant (and horse) harnesses, as seen in {1}. It turns out that the design of a harness is meticulous and deliberate. As visual 1 shows, a harness consists of many interconnected components—a true engineering feat designed to effectively link the animal’s strength to the rider’s control.
At its core, a robust harness manages three essential physical dynamics: moving, stopping, and steering. Yet, for a harness designed to manage intelligence rather than mass, these physical requirements translate into cognitive ones. To effectively command the AI elephant, we need a structure that organizes raw output into actionable human insight.
Let’s Get Airborne
This is exactly what we are building with MindLi: a harness for the AI elephant that empowers humans to soar. We have already identified the three core capabilities of human thinking: Add (adding new knowledge items), Box (organizing them by projects, domains, and subjects), and Create (turning them into outcomes, decisions, or further thinking). Of course, we are also working on the details that support these capabilities and their relationships.
We cordially invite you to join us in co-designing the harness. Feel free to explore our website at mindli.com {2} and discover our comprehensive help center {3}. I look forward to hearing your suggestions, ideas, and feedback.
More Info
{1} More about horse harnesses: https://en.wikipedia.org/wiki/Horse_harness
{2} Our website: https://www.mindli.com
{3} Our help center https://docs.mindli.com
