The most widely used digital platforms for communication — social media, messaging, forums, content networks — were built to connect people at scale. They have succeeded at this. Billions of people are connected through them. Connections that would never have formed in a pre-digital world now form instantly, across geographies and demographics and interest groups.
And yet the quality of collective understanding — of the shared comprehension of complex questions that communities need to navigate them — is not improving. In many measurable ways, it is getting worse. The connection is real. Something else is also happening.
What optimising for engagement produces
The design logic of social platforms is engagement: the number of people who interact with a piece of content, the time they spend on the platform, the volume of connections they form and maintain. Engagement is what these platforms were built to maximise, because engagement is what their business models require.
Engagement and understanding are not the same thing. The content that maximises engagement is almost never the content that maximises understanding. Emotionally activating content drives more engagement than epistemically careful content. Simple narratives drive more engagement than complex ones. Content that confirms existing beliefs drives more engagement than content that challenges them. The algorithm that maximises engagement will therefore systematically promote the emotionally activating, the simple, and the confirming — and systematically suppress the careful, the complex, and the challenging.
The platforms built to connect us at scale are optimised for a property of content that is nearly perfectly inversely correlated with its value for building collective understanding.
The epistemic consequences
The consequences for collective understanding are structural, not incidental. A community whose information environment is shaped by engagement algorithms will, over time, develop an information diet that is high in emotional activation and low in epistemic diversity. The ways of knowing that are quieter, more careful, more uncertain — the lived experience that does not fit a shareable format, the expert knowledge that requires context to be understood, the philosophical challenge that complicates rather than simplifies — are consistently disadvantaged by engagement-optimised environments.
This is not a problem of individual choices. People participating in these platforms are not making bad epistemic decisions. They are responding rationally to an environment that rewards certain kinds of contribution and penalises others. The problem is the environment — the design of the platform that determines what gets amplified and what disappears.
What the alternative architecture looks like
The design principles required for collective understanding are almost precisely the opposite of those required for engagement maximisation. Understanding requires: persistence of contributions over time (not a feed that scrolls away); preservation of epistemic ground (not just what someone said, but the type of knowledge behind it); structure that enables perspectives to develop in relation to each other (not sequential disconnected posts); and honest representation of uncertainty and disagreement (not a system that rewards confident assertion).
These are the design principles that Hunome's SparkMap architecture implements. Contributions persist. They carry their knowtype — the epistemic ground beneath them. They connect to other contributions in a visible structure that shows how the understanding is forming. Uncertainty is preserved rather than summarised away.
What communities can build with different infrastructure
The question is not whether social platforms should be designed differently. The question is whether communities and institutions that need genuine collective understanding have access to infrastructure that can produce it.
Groups working on genuinely complex shared challenges — climate adaptation, organisational strategy, policy design, community futures — need tools that work on different principles from engagement-maximisation. They need tools that treat epistemic diversity as an asset, that preserve the reasoning behind perspectives, that allow understanding to develop over time rather than expire with the news cycle.
The infrastructure for that kind of collective understanding building exists. The design principles are opposite to the platforms that have become the dominant information environment. The difference in what they produce is correspondingly significant.