However, after playing with it, checking out their board of directors, and deconstructing their app design, their vision is not really "cloud storage", at least, not the way we typically think of it.
Their long-term mission is preserving a digital legacy, oriented around relationships, families, and organizations. You don't use permanent.org to store things in the cloud that people normally think as "cloud storage", not for the day-to-day stuff. The kind of things you want to store in there are the things you want the world and your descendents to have access to after you die. They won't have to (directly) pay upkeep to keep that legacy preserved. I think that is convincing enough for me to see it as a critical piece of free and open web, even if this doesn't seem obviously connected to the idea of preserving a legacy.
For example, an indie musician wouldn't have to rely on SoundCloud to keep their recorded music around. SoundCloud is not in the business of preserving the creative work; they are in the business of aggregating users and they use user content to do it. Placing those music files in permanent.org has a much better shot of preserving that creative legacy for future generations than leaving it on SoundCloud.
PGP sucks.
PGP (GnuPG at least) is lightweight. I don't need an Electron dependency or a multi-megabyte chat room in the same application (address space too?) that supposedly keeps my private keys safe.
PGP is spoken by everyone, every programming language, having implementations on even ancient operating systems and architectures. Every email client worth its salt can use PGP. Emacs can decrypt and encrypt GnuPG-encrypted files seamlessly; other editors have plugins to do the same.
Even when they do use it, it’s easy to mess up.
The biggest flaw though is that in person key signing parties were never a viable or realistic thing for identity verification and web of trust based on that works poorly as a result. The use of multiple signed public social media accounts for identity instead as a way to fix this was Keybase’s main innovation.
For UX, even Snowden couldn’t get Greenwald to set up PGP and after multiple attempts Snowden eventually gave up and tried Laura Poitras with better results, the burden on the user is too high.
https://pgp.mit.edu/pks/lookup?search=Satoshi+Nakamoto&op=in...
Not to mention it's notoriously slow and has been shown to be an insecure method of distributing keys (due to the fact that anybody can upload any key).
The whole point is that you don't just use it to upload a key. You link various verified identifies of yours across the web to your Keybase account so people know the PGP key there is the one of the verified person. It's a way to tie all your verified identifies together.
If someone would manage to compromise a bunch of identities of someone on the internet, and then create a Keybase account with them and then upload a compromised PGP key that would be a problem if you don't verify the key. But that's a bit of a stretch.
My real wish is that keybase supported ssh keys and would provide them as an agent.
I do agree on the ssh key feature being nice, and, here's a link: https://keybase.io/blog/keybase-ssh-ca You could also just use the keybase file system to keep ssh keys around.
See https://keys.openpgp.org/about for why.
Their original purpose — tying identities to keys — could have been a nice small non-profit. But there aren't fortunes to be made from managing GPG keys, so they had to pivot into shark jumping.