MyceliumWebServer Update | ActivityROS Demo

Shows decentral AI fungus that can be browsed like a chat bot connected to other chat bots - and it can also be interacted with through Mastodon and its connected knowledge graph. It's basically three views of the same thing, just from different webs!

Its a decentralized AI bot based on volunteer computing and decentralized, federated learning.

You can find it under: @[email protected]. It will recommend songs when you post under #babyfungus. You can also insert your own songs. Happy home-grown ai!

(Please keep in mind: right now, its just a POC)

You can now access the fungus' chat bot via the browser:

This way, you can actually travel the chat bots of the mycelium web, by entering their url in the browser.

They are also connected to ActivityPub and the underlying knowledge base can be browsed in the browser, too, of course.

Next milestones:

• Add Activity**Pods **integration - solid pods are small storages of data. This way, every mycelium server/fungus can have its own data storage. Right now, they share a jena fuseki server together. There already exists a bridge to ActivityPub called "ActivityPods". With it also comes its own small Activity-Server, which is nice, too
• Be able to actually "browse" the chat bots of the mycelium web based on linked data (shouldnt be that hard, because they already communicate over shared data).
• Generalization of functionality: right now, the fungi can only recommend songs. They should be general AI agents
• Overall improvements

You can think of this whole project basically as huggingface based on WWW technology. Spoiler alert: hugging face will enshitifiy too.

A fungus sits at the intersection of the social web (Mastodon, Pixelfed, Lemmy, etc.), the semantic web (knowledgraphs like Wikidata.org) and decentralized federated learning.

This means it represents the "computation web"-aspect in the above picture. Together with other similar agents, it result in a decentralized, federated web of AI agents that work on open, shared data and are open to communities. Everybody should be able to set up their own fungus service and help to grow an AI model of their choice. I call this the "fungiverse" or "mycelial web".

A fungus web-service ...

• answers user requests over the social web
• users are also able to insert knowledge via posting to a hashtag the fungus listens on
• writes and reads data from the semantic web to collaborate with other fungi agents (this would ideally done with decentralized technology like solid pods, or other knowledge graphs, e.g. like wikidata.org or an own fuseki server)
• develops a shared AI model (which is also written to the semantic web) based on decentralized federated learning (which would be ideally be based on something like FlowerAi, but isn't at the moment).

The three parts can be thought of as a social network, a chat bot network and a knowledge graph (like a wiki for data), with all of the three aspects interlinked - the fungus drawing its data from the knowledge graph and being available in the social web as a bot, the social network to give access to the knowledge base and AI bots through bots, and the knowledge base, which can be viewed as a wiki with the bots that currently work on it and the comments from the social network. From a users perspective, the fungus appears as a chat bot interface (with the ability to view all bots its connected to and to which they are connected to etc.), the knowledge base would appear as a wiki like wikidata.org and the social network as a network like Mastodon.

The idea is that every user can install all of these three aspects on their own local PC to participate in the overall network (meaning having access to the social network, accessing global knowledge graphs and participating in decentralized learning to train their own bot).

Behaviour

In its behaviour its similar to that of a fungus (hence the name):

The shared model data can be thought of as the spores, which are also used by other fungi to adjust their models. The resulting AI chats available to the users are the "fruits" of the fungi.

Roughly, a fungi's behaviour is defined by a protocol, for example SPORE.

The following fungus is able to make song recommendations to the user: https://github.com/bluebbberry/MusicRecommendationFungus

Here is a more technical description of a fungus and its incentives for human-fungus-interaction.

Hi everyone! I’m building Seamantic, a Mastodon client that introduces the "semantic feed" — a way to interact directly with the Semantic Web.

Here’s how it works:

• Ask Questions: Post queries to the semantic feed. Bots like SeBridge (which I introduced in an earlier post) connect to knowledge bases to provide answers by listening on Mastodon hashtags for queries.

• Contribute Data: Insert data into the feed by posting insert-queries, helping bots grow their underlying knowledge bases.

• Sea-Level: Track your balance—querying raises the "sea-level," and contributing lowers it, encouraging collaboration. When the sea-level goes over a certain level, posting queries is blocked until the sea-level is lowered by contribution.

By connecting users and knowledge bases, the semantic feed creates a dynamic flow of high-quality, consensus-driven data.

What do you think of the idea? Feedback is always welcome.

The semantic web, for example through projects like dbpedia.org, tries to make the internet machine-readable. By making it accessible in Mastodon, users could quickly access and share data from the semantic web.

Theory

A new decentralized social media paradigma, in which data of posts is used to gain some addional features (opt-in).

I would additionally differentiate between central and decentralized bots.

Collective posting means posting to a bot (centralized or not), which then determines which posts go out and how, based on data collected by the bot through earlier iterations of this process, determined by some agreed upon guidelines.

Anarachist posting means posting to the Fediverse, while at the same time collecting all of these messages via a bot (centralized or not), which sends the data back to the client, who can use it again for some calculation based on chosen processes. Examples

Let's see what that means if I want to post my cooking plan for the week.

Anarchist Posting

I post my fridge-content and what I want to cook. This information with all the other participants is saved by the bot and later provided to me as suggestions for my next cooking plan based on a process that I picked.

Centralized bot

The bot posts global suggestions what each participant can do better (for example to eat healthier), based on their chosen settings.

Decentralized bot

Its all done in my bot, which displays suggestions individually to me.

Collective Posting

I post my fridge-content and what I would like to cook. The bot takes this message in with what all the others want to cook and have and then decides what I get to cook and post it based on a distribution-algorithm that I have agreed to.

Centralized bot

One centralized bot calculates the perfect meal for everyone; or every participant but the message is posted globally.

Decentralized bot

Perfect meals are calculated decentrally based on the agreed upon guidelines.

In Sidekick

Anarchist posting: When posting through the Dolphin-bot, all posts are collected and then used on the client-side to provide suggestions to Dolphin-users.

Collective posting: With the Buzz Lightsting-bot, all posts done via Buzz are not sent immediatly, but collected at the central Lightsting-bot, shuffled and then sent randomly over the participating profiles at specified times.

(This is just an outlook, its theoretically implemented, but the client currently only supports being used by a single person ... it would also be nice to attach this functionality to hashtags ... its all a bit work in progress ;))

Currently, there are six bots available:

• Larry: Classic text posting: just post text
• Spark: turns all text into uppercase
• Ennui: produces low-effort posts (removes punctuation, makes it lowercase, introduces typos)
• Jea: schedule posts by starting post with a hashtag and the number of minutes to postpone
• Hamlet: adds a random Shakespear quote to the post
• Legion: send out multiple posts by starting the message with a hastag with the number of repeats

You can select three of them and have them then available next to the post-button as a quick-select.