0 reactions
Built a blockchain where miners earn by running real AI jobs instead of burning energy on pointless hashing — here's how it works
Been building this for a while and finally published the whitepaper and open-sourced the code this week. Posting here because I think the economic design is genuinely interesting and want feedback from people who think critically about tokenomics.
**The core problem with PoW:**
Bitcoin mining consumes electricity equivalent to mid-sized nations to produce exactly one output — block security. The computation has zero utility outside the network. That always bothered me as a data scientist watching AI compute costs explode simultaneously.
**What Obelyth does differently:**
Proof-of-Useful-Work. Miners earn OBY tokens by completing real AI inference jobs submitted by developers. The same GPU cycles that run your inference job also secure the chain. No wasted energy — the work is the proof.
**The economics:**
* 90% of every developer fee goes into a diversified stablecoin AMM reserve (USDC/DAI/USDT/EURC)
* 5% Creator Share — permanent protocol fee, disclosed upfront
* 5% DAO fund — governance controlled
* 5% of all mined OBY automatically goes to the DAO vault at the consensus layer
* 92% of total supply is mined — zero VC allocation, ever
**Why 92% mined matters:**
Most tokens launch with 30-50% held by VCs and team members who dump on retail. Obelyth has no VC allocation at any tier. Every OBY beyond the 8% pre-mine is earned through real work on the network. The 8% pre-mine breaks down as 3% founder (4 year vest), 3% pre-mainnet community pool for early miners and developers, 2% Year 1 DAO discretionary.
**Where it's at:**
Pre-testnet. Python reference node is open source. Rust production node in development. Looking for early validators, GPU miners, and protocol engineers.
Whitepaper and code: obelyth.io http://obelyth.io
**What I genuinely want feedback on:**
* The 90% AMM liquidity design — does locking fees into a constant-product AMM strike you as a reasonable approach to protocol liquidity management or do you see attack vectors, I'm missing?
* The PoUW verification problem — how do you prevent miners faking inference results? We're using optimistic challenges with slashing right now, moving to ZK proofs later.
* Anyone who has seen similar designs — what failed and why?
Not a token sale. Pre-testnet. Testnet OBY has no monetary value. Things will be subject to change after legal review.
#cryptocurrency
Been building this for a while and finally published the whitepaper and open-sourced the code this week. Posting here because I think the economic design is genuinely interesting and want feedback from people who think critically about tokenomics.
**The core problem with PoW:**
Bitcoin mining consumes electricity equivalent to mid-sized nations to produce exactly one output — block security. The computation has zero utility outside the network. That always bothered me as a data scientist watching AI compute costs explode simultaneously.
**What Obelyth does differently:**
Proof-of-Useful-Work. Miners earn OBY tokens by completing real AI inference jobs submitted by developers. The same GPU cycles that run your inference job also secure the chain. No wasted energy — the work is the proof.
**The economics:**
* 90% of every developer fee goes into a diversified stablecoin AMM reserve (USDC/DAI/USDT/EURC)
* 5% Creator Share — permanent protocol fee, disclosed upfront
* 5% DAO fund — governance controlled
* 5% of all mined OBY automatically goes to the DAO vault at the consensus layer
* 92% of total supply is mined — zero VC allocation, ever
**Why 92% mined matters:**
Most tokens launch with 30-50% held by VCs and team members who dump on retail. Obelyth has no VC allocation at any tier. Every OBY beyond the 8% pre-mine is earned through real work on the network. The 8% pre-mine breaks down as 3% founder (4 year vest), 3% pre-mainnet community pool for early miners and developers, 2% Year 1 DAO discretionary.
**Where it's at:**
Pre-testnet. Python reference node is open source. Rust production node in development. Looking for early validators, GPU miners, and protocol engineers.
Whitepaper and code: obelyth.io http://obelyth.io
**What I genuinely want feedback on:**
* The 90% AMM liquidity design — does locking fees into a constant-product AMM strike you as a reasonable approach to protocol liquidity management or do you see attack vectors, I'm missing?
* The PoUW verification problem — how do you prevent miners faking inference results? We're using optimistic challenges with slashing right now, moving to ZK proofs later.
* Anyone who has seen similar designs — what failed and why?
Not a token sale. Pre-testnet. Testnet OBY has no monetary value. Things will be subject to change after legal review.
#cryptocurrency
| 4,000 mlx | |
| 11 | |
| 0 |