the most important thing AI can do
helping humanity see the same reality at the same time
AI’s most important use case isn’t content, productivity, or creativity.
It’s coordination.
There’s a kind of problem that refuses to be “solved” the usual way. You can’t just gather more data, hire smarter people, or throw more money at it. These are the aptly named “wicked problems” – and most of humanity’s biggest challenges live here.
A simple problem has clear boundaries.
You can define success, collect the right information, test approaches, and ultimately converge on a fix.
Fixing a bridge is simple in this sense: engineer the structure, source the materials, do the thing. The problem can be held in one team’s head.
Wicked problems don’t work like this.
They span domains. No single discipline owns them. They involve many actors with competing incentives and they change as you try to address them.
Climate change is wicked.
It’s environmental, but it’s also economic, political, technological, and cultural. Every intervention creates ripple effects. Progress in one area can create setbacks in another. You don’t “solve” it and move on. You can, at best, navigate it.
AI governance is the same deal.
Technology moves faster than policy. Companies compete globally while regulations happen locally. What looks like progress today can create new risks tomorrow. No single body can control the outcome.
As I often highlight, there’s no shortage of people working on these problems. Researchers, policymakers, technologists, activists — all making progress in their lanes. The issue is that wicked problems generate information faster than any human or institution can process it, let alone integrate it.
Climate research alone produces thousands of papers a year across dozens of disciplines. AI development is splintered across labs, companies, and countries. The challenge isn’t just gathering data. It’s making sense of it collectively, fast enough to actually act on it in relative alignment.
That’s the bottleneck: shared understanding at the speed and scale the problems demand.
Most of the conversation about AI focuses on tools: writing, coding, design, research. All useful. But I think the bigger opportunity is upstream.
For the first time, we have a technology that can sit in the middle of all this complexity and help us see a shared picture:
Synthesizing research across disciplines in real time
Continuously integrating policy moves, technical progress, and implementation barriers
Surfacing how solutions interact and compound
Tracking cross-domain patterns no single person could hold
In other words: AI as a kind of collective working memory. Not replacing human judgment, but giving it a better, cleaner input.
We’ve seen what real coordination looks like.
The Manhattan Project brought physicists, engineers, and military planners into one coordinated effort to build the world’s first nuclear weapons. There was shared information, shared priorities, and a clear sense of where they were in the process.
The Montreal Protocol tackled the ozone crisis through structured international coordination. Scientists built consensus. Policymakers negotiated together. Industry adapted in sync.
Both efforts depended on infrastructure for synthesis and shared reality. But the scale was still human. A dedicated group of people could track the problem, manage the updates, and maintain a common picture.
Today’s wicked problems have blown past that threshold. The domains are too interconnected, the actors too distributed, the pace too fast. No team, no matter how brilliant, can manually synthesize the full landscape anymore. The information complexity has exceeded human bandwidth.
The technology now exists to do what the Manhattan Project and Montreal Protocol did manually, but at the scale our problems demand.
AI can process the thousands of papers, track the policy changes, monitor the technical progress, and present a coherent picture of what’s changing and why it matters.
What’s missing isn’t another report or another conference. It’s durable infrastructure that can synthesize what’s happening, keep it updated, and make the picture clear to the people doing the work.
Wicked problems only move when groups can see the same reality and act from it. AI can build that shared picture at the scale we now need. The question is whether we build the infrastructure around it.
Because AI will reshape the world regardless — the issue is what we point it at.
If not us, who?
j