I tried it a few weeks ago and found it completely unusable. I do not get how they claim to be using it. I tried what I thought was the intended workflow and it's state got corrupted right away without any warnings.
In my work with high-performance engines, 'on-the-fly' graph resolution is usually the first thing to hit a performance wall compared to simple snapshot-based lookups. Pijul is a brilliant experiment in Category Theory applied to VCS, but until it can demonstrate that it doesn't degrade linearly with history size, Git's 'dumb' but fast snapshots will likely win the network effect battle.