We propose to create an onchain emergent narrative game called Godmode Chronicles. In the game, players are tasked with telling compelling stories about an AI that once trained to recognise objects in a simulation and that occasionally indulged in cheating. Stories must be constructed using only data captured from that simulation and made available to the players, but the narrative style, order, and what to include or exclude is entirely up to the players. The best stories will stand the test of time, the worst stories will be burnt.
The data players will use was captured from Godmode Epochs, an AI training simulation game developed by dmstfctn in 2023, played by over 1000 players. As data from that simulation over time does not make for a story, but rather a historical record of events, players of Godmode Chronicles will work their way through it, find and arrange the captivating elements, and collectively tell stories from it. Similarly to the AI once tasked with recognising objects from image data, players are tasked with recognising stories from historical records.
Godmode Chronicles aims to be the first prototype of a ‘Large Lore Model’ – a tool for decentralized onchain narrative building theorised by dmstfctn, Eva Jäger and Alasdair Milne in an essay of the same name, published in Autonomous Worlds N1 (0xParc, 2023 https://aw.network/posts/large-lore-models). The prototype combines data stored on a blockchain describing core events from Godmode Epochs with a “para-ledger” of stories retelling those events.