开发者用决策模型做元宇宙教学质量控制 agent
作者在 Project0 元宇宙里建免费大学，用 Levanto 决策模型改造出质量控制 agent 巡视虚拟空间并给改进反馈，称 Sage 比 Jev 更好用且成本大降。
Having a lot of fun with the new metaverse project launched yesterday by @Stefan_3D_AI. Proud owner of lot #1, where I am starting Ludus, a public forever free university. It will take a few more weekends to complete and polish, but the visual lessons are really coming along. There are going to be 12 subjects to learn about, the video below is the start on the AI track. There is also a track to learn about blockchain technologies and protocols using realtime data from @CambrianNetwork API. It took some time to build good systems for building these, and I'll eventually share a url to my own more complete engine experience, but this is fun to start sharing for now in a new cool metaverse worth checking out. One tool that has been particularly useful is @levantolabs decision model, they recently shared excellent code for an agent playing WoW, and it was easy and effective to adapt to quality control agents exploring physical space and providing feedback on areas to improve on. I tried Jev and a few others but Sage seemed a lot better. Has also become extremely cheap compared to their first version in just a few months. So scalability in making improvements is becoming better than I thought already. You can go try it out here: https://project0.city/#1