I watched Brendan's "Why Qri Failed" (listed above as "Building Apps on IPFS v3") talk and took some notes on it. (For background, Brendan started a company called [Qri](https://qri.io/) that provided archival services for large scientific datasets on [[IPFS]]. That company announced it was [shutting down in January](https://qri.io/winding_down). This talk is about why it failed.) - App developers have different types of concerns: - speed of development cycle - stability - ease of getting started - simple abstractions - IPFS (Kubo and other libraries) not designed to allow app developers to iterate quickly - Abstractions of the library are too low-level for app developers (having to care about file encodings for textual data, for example) - There are not enough app developers building on top of IPFS for the IPFS library maintainers to be able to iterate quickly with user feedback - Maybe we (as in Fission) should "pay our first users" and incentivize the first app developers on IPFS/WNFS/Fission tech. - This matches with the idea of [the "on-site customer"](https://martinfowler.com/bliki/OnsiteCustomer.html) from Extreme Programming.