The org chart was a compression artifact
Hierarchies were never the point. They were the cheapest way to move information through an organisation when moving information was expensive. That constraint is lifting, and the structures built around it are starting to look like what they always were.
Consider what a manager actually does for a large part of the day. They take what is happening below them, compress it into something their own manager can absorb, and pass it up. They take decisions made above them, decompress them into instructions, and pass them down. The org chart is, in a real sense, a diagram of where information gets squeezed and where it gets unpacked. It is a compression scheme drawn as a tree.
This was a reasonable design. A person can hold a working relationship with maybe seven or eight others before the overhead of staying in sync eats the day. So organisations grew in layers, each layer summarising the one beneath it. The cost was paid in distortion. Every summary drops detail, and the detail it drops is chosen by whoever writes the summary. By the time the ground truth reaches the top, it has been through five rounds of lossy encoding, each one shaped by what the encoder wanted to be true.
What the structure was hiding
The interesting question is not whether hierarchies are good or bad. It is what they were compensating for. And the answer is consistent: they compensate for the fact that no single point in the organisation can see the whole thing at once. Layers exist because attention is scarce and context does not travel for free.
A summary is a decision about what not to say, made by someone whose incentives you cannot see.
Once you frame it this way, a lot of organisational pathology stops looking like a people problem and starts looking like an information problem. The reorg that solves nothing. The strategy that is sound at the top and incoherent three layers down. The risk that everyone on the ground could see coming and no one above the fold heard about until it landed. These are not failures of character. They are what happens when the compression ratio gets too high and the loss starts to matter.
What changes when context becomes cheap
The thing that made the tree necessary is now the thing that is changing. When the cost of holding context across an entire organisation falls toward zero, the layers that existed only to manage that cost have less reason to exist. Not because management disappears, but because the part of management that was really manual data compression can be done differently.
That does not flatten everything into a happy network. Decision rights still matter. Accountability still matters. Someone still has to choose. What changes is that the choosing no longer has to ride on top of a stack of lossy summaries. The person making the call can have the unsummarised picture, and the people on the ground can see how their work connects to the call without waiting for it to be decompressed back down to them.
We think the companies that take this seriously will not look like flatter versions of the companies we have now. They will be organised around something other than the cost of moving information, because that is no longer the binding constraint. What replaces it is the open question, and it is the question we find most worth our time.
This is an essay, not a roadmap. We publish what we are working through, including the parts we are still unsure about. If it provoked a thought, we would like to hear it.