REPEAT AFTER ME
i will not my sell my $ADM under 100M. i will not my sell my $ADM under 100M. i will not my sell my $ADM under 100M. i will not my sell my $ADM under 100M. i will not my sell my $ADM under 100M. i will not my sell my $ADM under 100M.
The $AIUS front end update for v5 voting periods is underway. It is designed to dynamically react to your interactions.
The voting button will be greyed out and provide messages based on these scenarios.
*"Voting has not yet begun." (before voting period has begun)
*"100% of user owned governance power must be contributed before proceeding." (during voting period)
*"You have already voted, create a new stake to vote again with your new balance." (during voting period after voting one time)
Each voting round will determine which AI models receive the "lion's share" of mining rewards on the network and also influence compute provider's contribution.
Revenue will be forwarded to the DAO's treasury. Increased revenue from inference fees generated will boost rewards for stakers. The DAO can vote to fund programs, dev initiatives, burn tokens, or issue buybacks with the funds accrued. AI agent frameworks in time will be the largest consumers of this decentralized inference provided.
Each AI model on the network will have its own hardware requirements, inference fees "which can be adjusted" and rewards rates for successful solutions provided by the mining community.
Later on, tangible revenue metrics will play a big part in influencing network emissions. I believe that, compute provider rewards should go to AI models that are in highest demand and most profitable.
In regards to Arbius Transformers, the dev team has successfully tested its ability to generate a docker container that can run inference from imported AI models inside the @huggingface library. Further alpha stage development to reach a working public beta will include streamlining its implementation so that it is much easier for others to use. I have emphasized to the team that the act of streamlining should not be a stopper for getting the mining network operational. In the early stages, our internal team will curate and add the AI models and eventually decentralize that process more as time moves forward.
After minimum viable product is achieved and the network state is active on @arbitrum, that is when I feel it will be more necessary for Arbius Transformers to be open for public usage. In tandem, the addition of Gauges will be handed off to DAO votes via proposals. We want the community involved in deciding what AI models should make the list!
In my humble opinion, this design is superior to all other attempts at decentralized inference including $TAO.
Remember also that truly uncensored (abliterated) AI models is only possible with real decentralization. Things like $VVV rely on B2B payments to DePIN which is prone to censorship. DePIN does not solve decentralized AI, at any time large corporations could rent out all their available hardware for centralized purposes. There will never be another crypto innovation as meaningful after @arbius_ai is fully deployed.
This is all I have for you now. Take care!