Back in 2020, while federating more than 100 service meshes across on-prem and AWS for a big hybrid cloud project, I had an idea: what if we could "split" the CAP theorem in a way that flips its limitations, enabling massive scaling far beyond traditional consensus protocols? Fast-forward five years:I started prototyping with libp2p, but the networking layer was always just a means to an end; the real goal is that CAP inversion/split for extreme distributed scaling. I think the timing is perfect given current geopolitical pushes.
Super curious to hear thoughts from folks. Any pitfalls I'm missing? Open to feedback or collaborators.