[REQUEST FOR FEEDBACK] Allo Arenas

[afo]

@owocki

Bittensor evaluates the quality and value of machine learning outputs.

In what context for machine learning?

@owocki

Develop a consensus mechanism that evaluates capital allocation strategies based on predefined metrics (Value Flowed, ROI, distribution efficiency, community satisfaction).

Would this be applied to all subnets are each can have different metrics and levels they’re focused on?

@owocki

Alternatively, have each subnet tokenize and judge success by how many that token performs.

Does this have any legal ramifications if a subnet tokenizes?

@owocki

Create simulated environments where allocation strategies can be tested using historical data or synthetic scenarios

Would love some examples here.

@owocki

Enable successful strategies to build upon one another through a knowledge-sharing framework

What could this knowledge sharing framework look like? There’s many potential frameworks to pull from that the community has proposed.

@owocki

Create incentives for continuous improvement and adaptation to changing market conditions

Can you provide an example of some incentives?

@owocki

Identify friction points for Allo builders and reduce the friction, enabling faster evolution.

What do you feel are the current friction points? Would love to reconcile this with conversation we’ve been having as Allo builders. The call this upcoming Tuesday to discuss the protocol I hope will provide clarity and alignment here.

@owocki

Once strategies prove successful in simulated environments, provide pathways to deploy them with actual capital (perhaps via GG)

Why is it GG not Allo Capital?

@owocki

Fee-based Model: Charge small fees on successful capital allocations, which fund the system’s ongoing development.

Not sure what this means, who’s paying the fee?

@owocki

The most practical approach might be a hybrid model:

100% The initial experiment will help ground the token in strong data and alignment with the goals of Allo.

Overall I feel the Arena approach is intriguing and with a strong architecture and more clarity in a couple areas seems like a viable path to capital mechanism building and validation.

A couple high level thoughts, one is how subnets are created, updated and destroyed. Don’t see details there and would self plug my Fa-Fo-Fu-Fi proposal with an emphasis on the Fo (Form) and Fu (Function) aspects of the proposal and how they have a lifecycle.

Fa-Fo-Fu-Fi (Fan, Form, Function, Finance)

Yep good thought will create some this week!

The last one is around the token, I think taking another look at Paul Gavin’s proposal for 2 token model might be interesting where this initial token is pure governance.

2-Token Model: Building a Community Battery + Microprocessor

Problem Governance tokens + “price goes up” mentality = inherent conflict between project contributors and investors. This isn’t a problem for organizations that build private goods since both groups share a common goal: profit. But for organizations that build non-private goods, maximizing the project’s success at its mission is often at odds with maximizing its revenue. Solution A 2-token model solves this by separating these functions: a governance token that powers the community’s micro…