The entire cryptocurrency airdrop thing has caused lots of noise, triggered campaigns to try and hack/social engineer accounts that would qualify the attacker to grab more cryptocurrency, ... It makes it vastly less likely I'll recommend Keybase with those associations, which diminishes the value of the "key part". (Not recommending it both because it makes me question the long-term priorities of the product and because I don't want to explain why I'm recommending "weird cryptocurrency-stuff", which is the impressions others could have)
Isn't Keybase built exactly for this though? To be able to look at the connected accounts/proofs, and know that a given Keybase user has proven control of those accounts? An attacker looking for crypto may have hacked someone's HN or GitHub. However, with Keybase, you can establish someone you want to talk to has linked their Twitter and their web domain and the like, whereas an attacker probably does not have access to all of their various identities around the Internet.
Keybase promised to give out free cryptocurrency to everyone who linked an existing Github and HN account, which lead to an attack wave on such accounts from people wanting to claim that: https://essays.suryad.com/hnhack/
If you are an identity service and do things that paint a target on others online identities, causing attackers to want to link my account with their fake accounts on your service for profit, you have an image problem. Similarly, that's not something I particularly want important services to be funded through long-term.
