Previously shared 4 projects in the decentralized AI computing power space, including comprehensive AI blockchain projects like 0G Labs @0G_labs, which focus on decentralized computing power as their core module. There’s also PrimeIntellect @PrimeIntellect, which spans the Web2/Web3 computing power market. A few days ago, I attended an online live session from an AI community about RL frameworks. One of the teams, http://gradient.network, also participated and shared updates on research progress in decentralized training and inference. Took a closer look, and it seems like an interesting project. They’re building on Solana, focusing primarily on decentralized RL training and inference. They raised a $10M seed round led by Pantera Capital and Multicoin Capital, with HSG (formerly Sequoia Capital China Fund) and other notable partners participating. Adding this one to the watchlist for this space. Like 0G @0G_labs, they’ve also shared some research papers, but I didn’t see a GitHub repo. Their main direction seems to be heterogeneous RL. One of their papers: *Echo: Decoupling Inference and Training for Large-Scale Reinforcement Learning Alignment in Heterogeneous Populations* https://arxiv.org/abs/2508.05387v3. There are a few more papers, but I don’t have time to go through them all. If you’re interested, check them out!