David Roon · · 4 min
When intelligence becomes abundant, perspective becomes the edge
Why distinct points of view matter more in the AI era—and how Haoman can help organizations give them a voice.
AI makes it easier to find information, develop an argument, and turn a rough idea into something coherent. These are extraordinary capabilities. They also raise a question: if everyone can access increasingly capable intelligence, what will make one organization’s thinking different from another’s?
I believe perspective will matter more than ever.
By perspective, I mean a particular way of seeing the world: an opinion about what matters, what is worth pursuing, and which assumptions deserve to be challenged. A point of view that leads us to make choices someone else would not make.
AI can help us explore these differences. But in everyday use, it can also smooth them away.
The pull toward the middle
Bring an unusual idea to an AI system, and it will often make it more reasonable. It adds qualifications, balances competing arguments, and translates unfamiliar language into recognizable categories. The result can be clearer and easier to communicate. It can also lose the quality that made the idea interesting.
This is not an unavoidable property of AI. Models can be prompted to challenge assumptions, defend unpopular positions, and explore radical alternatives. Yet the ordinary pattern of asking for a good answer often rewards something plausible, polished, and broadly acceptable.
Repeated across an organization, that pattern could have a real consequence: more articulate thinking, with less difference between the thoughts.
Organizations already have their own pressures toward conformity. Hierarchy, incentives, and the desire to reach agreement all influence which ideas survive. AI risks accelerating that process if we use it mainly to summarize discussions and produce a common position.
A summary can accurately represent what most people said while losing the most consequential thing anyone said.
Why the edges matter
Some ideas deserve attention precisely because they sit far from the current consensus.
An apparently unreasonable ambition can expose a constraint everyone else has accepted. A minority opinion can reveal a problem that the majority does not experience. A proposal that initially sounds impractical can force us to reconsider how the work is organized.
Being extreme does not make an idea valuable. Plenty of unconventional ideas are wrong. But distance from consensus should not automatically count against an idea either.
The challenge is to give unusual perspectives enough room to become understandable and testable before we make them acceptable.
That requires more than asking AI to “think outside the box.” It requires people who have convictions, experiences, and judgments of their own—and an environment where those differences can influence decisions.
An opinion matters when it changes what we notice, what we choose, or what we are willing to try. Its value comes with a responsibility: to explain it, expose it to evidence, and revise it when necessary.
What I want to build with Haoman
This is part of what I am trying to build with Haoman.
Organizational intelligence is often understood as the ability to know what an organization knows: its documents, decisions, conversations, and accumulated experience. That matters enormously. But I want to go further.
I want to enable different perspectives within an organization and let them speak.
Two people can agree on the facts and still see entirely different possibilities. Someone close to customers may recognize a shift that leadership has not noticed. A founder may defend a direction that current results cannot yet justify. An operational objection may reveal exactly what an ambitious proposal needs to become workable.
An intelligent system should help us understand those positions on their own terms. It should preserve the reasoning behind a disagreement and help people examine what would follow if a different perspective were taken seriously.
Agreement can still be the outcome. It should not be the default imposed by the system.
The choices intelligence cannot make for us
As AI makes competent analysis more accessible, having a distinct point of view becomes more consequential.
What do we believe that others do not? Which tradeoffs are we prepared to accept? What future do we want to help create?
AI can sharpen our answers. It can challenge them and help us act on them. But we still need to take a position.
The opportunity is to build organizations that use intelligence to expand the range of ideas they can seriously consider—and have the conviction to pursue some of them.