The storage idea in Solidus's 2017 founding letter, before \"Solid\" was in the vocabulary

Nine years ago the proposal was to stop storing data on rented hosting. Today we store it on rented hosting. That is the honest shape of this page.

The paragraph

From a letter written on 11 September 2017, before any code existed:

"Ben farzı misal sürekli çevrimiçi olması gereken datamı amazon aws ya da hosting firmaları üzerinde değil de bu ağ üzerindeki kişilerde depolayabilsem."

In English: suppose I could store the data of mine that has to be online all the time not on Amazon AWS or hosting companies, but on the people in this network.

The word "Solid" does not appear in the letter. Neither does "pod". The idea arrived at the problem, data lives on somebody else's rented infrastructure and that is a structural problem, not a pricing one, without arriving at the vocabulary the field later settled on.

What it got right, stated without inflation

The diagnosis. That personal data sitting on rented infrastructure is a dependency rather than a convenience is now an ordinary observation. In 2017 it was a less ordinary one to write down.

And that is the whole of the credit. The letter names other projects working on distributed storage in the same paragraph, so it was not an isolated insight and does not claim to be. A correct diagnosis widely shared is not a moat, and this page is not going to dress it up as foresight.

What we did with it, which is the uncomfortable part

The mechanism in the letter, data stored on peers, who are rewarded for providing capacity, is not what we built.

What runs today is a Solid server on one rented machine. One origin, one jurisdiction, behind a commercial edge network. There is no peer network holding anybody's data. The check is above and takes one request.

Our own archive says so, in its own words. The 2026 footer on that letter records the distributed compute and storage thread as deprioritized, and notes that the validator network is distributed for consensus, not for general storage or compute. We are quoting our own file contradicting our own founding document, because deleting either would be worse.

Why the swap is defensible, and why that is not the same as vindicated

Identity-first was a deliberate choice, not a drift: a storage network with nothing to say about who is asking is a commodity, and the identifier question turned out to be where the unsolved work was.

But be clear about what that choice cost. The 2017 idea's whole point was removing the landlord. We kept the landlord and changed the subject. A pod you run yourself is the version that keeps the original promise, and it is Solid's design that makes that possible, not ours.

The question the letter could not answer, still unanswered

The letter flags one problem as unsolved: how people who contribute to the network are rewarded. Its author wrote, in effect, I have not been able to make anything work here.

Nine years later that is still true, and the footer says so.

A founding document that names its own hardest unsolved problem, and a company that still has not solved it nine years on, is the most honest thing on this page.