Creating Social Knowledge Graphs by networking second brains via the Fediverse

I’ll be writing a chapter of an upcoming book on Personal Knowledge Graphs (PKGs) over the summer. Provisionally entitled *A decentralised social network of second brains”, it will describe the future of MyHub.ai, the publishing platform alpha-launched in 2020.

Given the book’s subject, my chapter will focus on the PKG at the heart of each Hub’s content management system (“thinking management system”), and the Social Knowledge Graph created by networking them together via the Fediverse. There’s a lot more: Solid hosting, the AI algorithms used and trained by its users, the data union the users form to benefit from monetising those services, etc.

I’ve just published a “brain dump first draft” of my chapter’s Abstract and Introduction. It’s long - I’ll probably move some of it into other parts of the chapter in later drafts - but right now I just want to get what I’m thinking out there to elicit ideas for me to pursue as I develop my chapter over the next two months.

So I’d love your comments and suggestions.

1 Like

There’s a lot to like, and I think in general your approach puts the finger on where there’s something lacking in all those many - often quite innovative - PKG projects: the space where your own knowledge touches that of others and where collab / cross-pollination may happen. And federation + linked data + personal data vaults may be the ideal technological foundation.

I wrote a small braindump of my own on this forum:

The solution you are working on would be strongest if the linkded data / AP extension vocabs would be well-defined and documented, so they describe a general Knowledge Management ‘business domain’, and can become part of a growing ecosystem of apps and projects that deal with it.

I have one point of caution, and an advise to that: Anything blockchain, cryptocurrency and DAO’s is highly controversial in Fediverse culture. Though there are fedizens that are into that space, if your solution is only perceived to be blockchain-based (whether that’s justified or not), you will lose the passion and potential contributions of a large part of fedizens, your projects be shunned and getting negative feedback (no matter whether that’s informed or not), plus the fedizens that you do deal with will be like an echo chamber with the ‘crypto lovers’.

So my advice would be to separate the parts. If you want to go DAO and blockchain, then make that a completely separetely-positioned thing. Where they can still be combined to constitute your overall diagram. Then there is another opportunity to have a second version of the diagram where the ‘blockchain parts’ are instead (platform) cooperatives, people-driven governance, and interfacing with common payment providers, etc.

I don’t necessarily disagree with the sentiment, but these are opinions stated as facts.

I for one will be interested in those things but don’t think of myself as a part of the cryptocurrency echo chamber.

For all this to work we need to operate as if ‘crypto’ means cryptology and not just cryptocurrency, and cryptology is a required part of building a fediverse. We can’t shy away from it.

Fair enough. Observed on many occasions as very active fedizen, but that is anecdotal and echo chambers depend on following and what you expose yourself to, so my perception may be shifted the other side. Still, I’d take this advice in consideration. ‘Crypto lovers’ referred to aforemenetion cryptocurrency, nothing wrong with cryptography. Unfortunate overloading of terms (like web3).

Thanks to you both, @aschrijver & @bengo, re: crypto. If you read the full post - and don’t worry if you have, as who has 18 minutes to read one blog post these days? - you’ll see

" and managed ( maybe ) by a DAO in which all users will stakes in, proportionate to their activity."

The "maybe " is there because I’ve never been entirely convinced about the need for a coin or even a DAO. What I do want is a system that rewards users proportionate to their [AI training] activity. An “in-game currency” is probably required, but if I can build that and get the requisite level of trust without blockchain/DAO, that’ll be easier.

I’ll comment direct to the Semmy post as soon as I get the time (I’m just about to head off on a 10day trekking holiday). Suffice to say now that there’s so much in there that finds an echo in what I’ve been immersed in these last 10 years, so thx for sharing.

1 Like