owned this note
owned this note
Published
Linked with GitHub
# Product discussion
Angela: Place to view all requests and progress against those requests. From my understanding where I would put it is Phase 2 for clients under the user story submit research requests / review outcomes, I would put this under view research requests, but MVP version of this is just sending this to you two
- Jordan: To the contributors of Diamond DAO
Jordan: Block type where user can query graph and view results in the block. Probably not MVP. It would kind of go under developers right?
- Christian: This would go into "analyst tool" but not Phase 1.
- Angela: Requires new user story, fullfill research request.
- Angela: Most of the research will be on Chainverse right?
- Jordan: A lof of research will be obtained outside of Chainverse
- Angela: Anything about searching Chainverse?
- Jordan: Not really. I guess you should be able to view metadata about the types of information associated with an entity.
- Angela: Leverage existing Chainverse data.
- Christian: User should have a wallet for each user.
- Jordan: Why do we want to custody their wallet?
- Christian: It will be stickier. They can use their own. Idk if it should be custodial or non-custodial?
- Jordan: Can we compromise and say it's an account? Two different things. Auth and managing assets earned on Chainverse.
- Angela: Depending on the claim process, they may need a custodial wallet?
- Jordan: Does this become a compliance or liability issue?
- Jordan: We need wallet-based authentication.
- Angela: Show a user or analyst a block within the graph.
- Jordan: View credentials for area of expertise.
- Christian: MVP version of this is like for completing tasks on the platform
- Jordan: This is useful for us, for filtering/sorting data
- Christian: Some way for users to indicate interest in making a request.
- Jordan: Yeah so like a button that's always there to solicit requests?
- Angela: This is probably phase 1.
- Angela: How do show usage for blocks? Like block metrics?
- Christian: Like how many blocks cite this block. Reminds me of what Boris said. Like knowing this block is useful and has been used in different contexts.
- Jordan: Start with number of edges created off of that node. One number. Strange that we've stopped trying to calculate.
- Christian: would be nice to reward people whose work is referenced a lot.
- Angela: The ability to link blocks within blocks
- Jordan: Block types to structure information on things relevant to organizations. Get people to create blocks that address specific inforamtion requirements. Would really like to limit block creation to those things.
- Investors
- Partnerships
- Etc, et
- Angela: Then we need to consider permissioning block creation, right? Is Chainverse a closed platform?
- Jordan: Consumption no, creation yes
- Angela: Add feature that allows user to create "draft" block before publishing to Chainverse as finished block
- Jordan: This could be an interesting opportunity for us. They can't publish unless it's a block type we are looking for.
- Angela: Process of distilling stuff into insights is an important part of reasearch. Do we want that stuff in Chainverse?
- Christian: I don't care.
- Jordan: I think it should be in Chainverse; if people have to switch back and forth wrisks information being left out of Chainverse
- Christian: Chainverse borrows components from other knowledge management platforms. For instance, we can build upon Logseq. We can see if using components from these, make an informed choice.
- We can get a huge lift forward? We enter open source communities?
- Angela: I would prepend that with "maybe"
- Jordan: Client can define request for information. State what they want on the object, steps for validating.
- Angela: Ability to traverse from a block. Phase 1 or Phase 2?
- Jordan: Phase 2.
- Jordan: Maybe we should say analysts? Analyst builds a brand? Got feedback that some of the terms we used before didn't sound legit. You can say "I was a research analyst at Diamond DAO".
- Jordan: Add profiles entities maintain on various platforms, like DAOhaus, Snapshot, Etherscan.
- Angela: this solves a lot of problems
- Jordan: probably a good principle for the product, don't use resources on traversals to display information that can be displayed through purpose-built tools
- Christian: Publish open-source standards for Chainverse data models
- Angela: View examples of quality research
- Christian: Part of a larger need to provide training and Q/A for analysts that receive credentials for Chainverse
- Angela: Adding more structure to blocks
- Jordan: Focus on schema, like block types, versus formatting, formatting later
- Angela: What is the difference between a block type and tagging a block as a type?
- Jordan: Block Type would mean enforcing a schema, tagging a block doesn't necessarily result in change in block schema
- Angela: So we are going to add attributes, like "Partner", for instance, so they can be indexed and searched?
- Jordan: Yes
- Angela: So we will create edges?
- Jordan: Yes and No. Through the block, not directly between the entities. Graph would get messy if users could create edges between entities.
- Jordan: In summary. Blocks should be labeled by type. Type enforces structure. Structure lets us aggregate & query
- Jordan: this will, among other things, help us assemble reports without needing to spend as much time formatting/organizing. Just aggregate blocks into relevant sections
- Christian: So like "backslash" /partner block, which will prompt them through a form designed to justify the partner block
- Jordan: Form is way to encourage people to justify their claims without formal validation
- Jordan: Can tempalte it out; can be text, multi-select, etc
- Christian: Some thoughts on general usability and accessebility
- More text focused workflow
- Lack of mobile functionality
- Jordan: I think this type of functionality becomes more important as user base grows
- Angela: So what's the function of the graph view?
- Jordan: I think it's useful for general public, people who are doing exploration more than adding research, right now the only important use case is breaking down objects that have a lot of data to make it easier for people to navigate them
- Christian: So it's kind of like a navigation tool
- Jordan: Yes it's a way to situate users within the context of a node, not necessarily the best way to communicate information about a node to a user. People get really excited about graph visualization but generally it's more of a novelty thing than a utility thing
- Christian: Serotonin release
- Angela: Why are we dropping the hypothesis that immersive tool is important?
- Jordan: We're not. It just seems like it will take a long time to build something general users can use & add value with. We will get to product market fit more quickly by encouraging users to add content
- Angela: So what's the ultimate vision?
- Jordan: For people to make a ton of money by contributing content on the platform, to level up from data collection to designing research to deciding what research gets funded
- Christian: Template for users to deploy their own Chainverse instance
- Angela: How should search work?
- Jordan: I think that the ability to search from a workspace is helpful. Right now block search is our weakest spot but not really sure how helpful individual block results is helpful. Also need to be able to search for entities to add them to workspace and use them as a pivot point to drill down. Ultimately we should work towards something closer to semantic search to actually answer user questions instead of just returning results
- Christian: Being able to search blocks is helpful, I do this in Obsidian, so I can link stuff to each other.
- Jordan: I agree that it's helpful to reference, I don't think we need to enable full fledged block exploration. So search blocks should be designed to fit that specific use case versus like returning a table with metadata
- Angela: We need to figure out the terminology here. Blocks, workspaces, etc.
- Angela: More block attributes for formatting, like titles, headers
- Christian: Change logs
- Angela: Versioning would also be good
- Christian: Integrate stripe / non-crypto payment options
- Jordan: Maybe when we have more customers
- Angela: How do we give users more insight into what's in Chainverse?
- Jordan: Dashboards, better communication of what our research priorities
- Angela: On the product side, what Wikipedia does is serve recommendations based on what it knows about the users
- Jordan: This gets back to your ideas about recommendations on the "Home" page
- Christian: Even if it's not super dynamic at first
- Jordan: Recommendations can be based on wallet affiliations, etc
- Jordan: Need to flesh out how governance plays in.
- Angela: How does governance fit in?
- Jordan: approving block schemas, voting to approve research proposals, etc
- Jordan: This is really complicated.
- Christian: We should get tacos
# Protocol discussion
Christian:
- So right now there's data out there, we pull it into AWS, structure it, ingest into Neo4j, and then make available through the Chainverse app
- It costs about $600 a month, with not that much data, and we have no revenue, so this is pure cost right now
- This is the situation Amazon was in, where infra was just a cost, but what they did was break out their cost into a service other people could use, and now it's the most profitable part of their business
- IPFS
- data is assigned a content identifier and split between nodes, users poll nodes, nodes with the data in their cache/that have pinned it can serve it to the user
- Ceramic
- Ceramic sits ontop of IPFS to expose that data for queries, w/permissions
- High-level plan
- First step, decide which parts of our stack we want to open source and which we want to operate "privately". We want a lot of stuff to be open source.
- Some features are difficult to implement using exclusively open source software
- For instance, GraphQL exposes our data but we can't do complicated traversals and network analysis without Neo4J
- The first thing we open source is our data models / data structures
- So for instance, there's the concept of a "DAO" Entity, that could have x,y,z attributes and a,b,c allowed relationships?
- You're saying that "this data type meets this format" and benefit is that it makes the data interoperable
- This enables people to query our data through Ceramic
- However, Neo4J still exists to support our "commercial" operations
- Research for clients leveraging Chainverse platform
- Specialized UIs for querying and exploring the data