This is part of our series on a year of Bittensor experience, leading up to our first anniversary at the 13th of July. We discuss 9 miner archetypes (and some related types) before digging deeper into incentive landscapes. After describing honest, rogue, calculative, blockchain-savvy, overfitting, devops, (over)analyzing and clueless miners, we turn our attention to the slow miner. Please let us know what you think in our Discord channel!
Some miners eventually do find out the tricks, either through their own research or by carefully analyzing Discord messages and logs, but are too slow to implement them. Subnets evolve fast, and the low hanging fruit is picked in the blink of an eye by the winners – some tricks only work for a day or two. In many subnets you can see how most miners converge to the same solutions, while only the first ones to move make good earnings from them.
Although this may come off as a bit negative, having a solid base of slow miners might be good for subnets, as they provide some stability. Every subnet should prefer a diverse mining base over a monopoly. By asking trivial questions in Discord the slow miners also help document the subnet in Q&A form, and help shape the narrative of the subnet; if the slow miners don’t fully understand the subnet, how could external investors do? Slow miners are also ideal for optimizing log output and monkey testing.
As more and more tricks are eliminated, all miners become honest miners, and the slow miner very often catches up with the miners that found all the tricks. The incentive landscape may need some specific adjustments to keep the slow miner engaged from the start.
Next up is a true archetype: the lazy miner.
0 Comments