By: Barbarian
Bittensor’s MANTIS project is making waves with its innovative approach to signal processing and prediction markets. Following a recent AMA (Ask Me Anything) session hosted by project lead Barbarian, we’ve compiled the highlights.
This recap draws directly from the discussion, shedding light on MANTIS’s progress, future plans, and technical edge.
Whether you’re a miner, alpha token holder, or Bittensor enthusiast, these answers provide a clear roadmap for Subnet 123 and beyond.
Q: How far away is MANTIS from generating revenue?
A: Revenue generation on Subnet 123 involves three key steps: Generation, Verification, and Monetization. We’re currently tackling trustless verification as the primary bottleneck, with a solution expected in the next approximately three weeks.
Q: What will Subnet 125 focus on?
A: Subnet 125 will be an architecture search subnet, targeting language models and long-horizon forecasting tasks. In the future, it may split into specialized sub-subnets to allow for more distinct and focused innovations.
Q: How will alpha token holders benefit from holding positions in Subnets 123 and 125?
A: Alpha holders will directly benefit as all revenue generated from the subnets’ outputs will flow back to token holders. For Subnet 123, there’s a clear path to revenue. This could include buybacks, burns, and potentially direct TAO deposits into holders’ wallets or claimable via a dashboard in the longer term.
Q: How will the ‘mini-Bittensor inside MANTIS’ actually function? Is it a step toward sub-subnets?
A: This feature will enable users to generate custom targets, such as forecasting the expected volatility of RUB/USD in the next hour. Holders can then vote on emissions allocation per challenge, creating a mini-Bittensor ecosystem within Subnet 123.
Q: What makes MANTIS superior to existing financial prediction tools?
A: Unlike other prediction networks, MANTIS rewards miners solely based on improving overall network performance. Duplicating existing signals, information, or strategies adds no value—if a miner enhances the network’s outputs, they’re rewarded proportionally to their contribution.
Q: Why is XGBoost better suited for MANTIS signals than transformers or neural networks?
A: XGBoost has proven ideal for interpreting the outputs of linear regressions trained on miners’ embeddings. While we may evolve the model and hyperparameters in the future—potentially even running challenges on Subnet 125 to optimize this—XGBoost currently delivers superior results in this context.
Q: Why should miners start mining on MANTIS?
A: Miners are incentivized to validate and submit signals that genuinely improve network model outputs. It’s in their self-interest, as rewards are tied directly to profitable contributions that elevate the subnet’s performance.
Q: If everything goes according to plan, what’s the ideal outcome from the NumerAI appearance?
A: The NumerAI event should serve as a platform to showcase MANTIS’s vision, design, and architecture, attracting high-quality miners. Since network outputs depend entirely on miner quality, boosting participation and refining evaluation methods is the subnet’s core goal.
This Q&A captures the essence of the AMA, emphasizing MANTIS’s role in advancing Bittensor’s decentralized AI economy. With trustless verification on the cusp of resolution and expansions like Subnet 125 in the works, MANTIS is poised to drive significant value for the community.
Stay tuned for updates as these developments unfold.
Be the first to comment