About a day after I published the post on Keybase and their inaction against spammers on their platform, Keybase put out this blog post, “Dealing with Spam on Keybase.” They outline changes they’ve been working on, including the ability to block and report people, and what they deem the “nuclear option”: the ability to whitelist people you chat with, which they say “breaks Keybase’s social features.”
What do you think of the post and changes in general? In the post, they also answer the question of “Why didn't you do this sooner?” with “There were other things about Keybase holding it back far more than spam. We were working on those things.” What do we think those things are? That’s a lot of questions, so any of them, or your thoughts on Keybase in general, will do for the open thread. :)
As always, looking forward to a good discussion!
Previous open threads:
Open source tools from big tech companies
Good recommender systems
Ok, I'll start. I think the intention is very good, but the wording of the post really bothered me. It seemed like someone was twisting their arm. "Ok, we'll do this if we have to but it's really lame." I now don't know whether this stems from the Stellar funding of just the general MO of the founders towards social safety on online platforms.
At the end of the day, it misses the obvious point: not all services need to be, or should be "social", and I think experience has shown that if you are developing a service, you have to pick one or the other. You can make a tool for (small groups of) individuals to communicate privately with each other, or you can make one for large public forums. I am increasingly convinced that you can't effectively make one that does both. Unfortunately with their VC funding they are now committed to being the latter and so have blown their chance for being the industry leader in the former.
This doesn't account for people who decided to message other people (and in particular women) that they don't know, but when I joined keybase I followed a bunch of R developers that I don't know personally because the docs mention that "older follower statements are better ones." https://keybase.io/docs/server_security/following
On keybase more generally: I'm very enthusiastic about the "public keys for the masses" vision even though there's nothing I strictly depend on it for at the moment. I use the private filesystem a bit. I'm in a few public slack threads that are at their limits and I consider suggesting moving over to keybase teams but don't want to be 'that guy.' As the Android/Windows user in groups of mostly iFriends, I have already done that badgering to get people to use Telegram, and keybase seems like it would be far too much to ask since it surfaces a lot of workflows they would have no use for.
Keybase's social features have completely failed to get traction. But they seem to think that allowing anyone to message anyone else is what will help these features become popular?
By analogy, since I can't DM pewdipie on Twitter, but I can on Youtube, I'm going to use Youtube for all my communication needs!
> I think the intention is very good, but the wording of the post really bothered me. It seemed like someone was twisting their arm.
Very much this. Apparent from the misalignment of what people complain about and what they said. They spent as much time spinning it as they did actually explaining. The lack of detail is meaningful. There's no standards for behavior. There's no actual _rules_ here. It's the same playbook given in social media. Milquetoast rules, whatever the minimum acceptable to an affluent white techie crowd is. Likely to be a bit anti-sex, not particularly anti-harassment. Unaware of power structures.
At the end of the day, the other things they've been working on have been "a viable business model". There wasn't one, so they got in bed with cryptocoins.