Soul Interrogation: What are blockchain+AI projects doing now? Everyone in the cryptocurrency industry has gone into AI. The people who work in technology and content have also gone, and I have a question. Isn't there a lot of AI projects in the cryptocurrency industry? Are they okay? What are they doing? DeAI used to be all the rage, why do we have to look at the faces of people in the AI industry and become a new leek over there now? Let's dive in! ⬇️ Firstly, it must be the price of the currency, with the vast majority experiencing a severe decline and over half already approaching zero. Of course, if we look at the entire market, there are also many that are even worse than the concept of AI, so we cannot conclude that AI is a scam based on this data. Then there is the project progress, with over half of Twitter no longer updating or making do with it, and the product itself almost stalled, such as the project where someone wants to eliminate AI illusions. But there are still some projects that continue to thrive in the bear market. When we talk about products alone, we can still get a glimpse of the situation. At least, we can see which projects still have money to build, want to build, and expect to cause trouble. This angle is definitely not wrong. Bittensor is still celebrating the 72B parameter, and @ 0G_labs trained a 107B decentralized model in July last year, which is now open source. They are still competing in this direction, but it is clear that 0G is far ahead. From the perspective of crypto+AI, DiLoCoX-107B is actually a milestone breakthrough in decentralized AI training. If you remember why decentralized reasoning was used, then you should know the significance of this training result. DiLoCoX-107B uses a decentralized node network that trains ordinary people and nodes around the world with GPUs and regular 1Gbps home Ethernet cables, eliminating the need for billions of dollars in data centers. Again， The models you eat for lobster now are monopolized by a single company in terms of computing power, data, and models, which are expensive, closed, and easily subject to scrutiny. DeAI uses token incentives and blockchain verification to organize idle GPUs around the world, mining AI like Bitcoin mining. Nodes contribute computing power to earn tokens, and users use tokens to pay for inference/training fees, forming a closed-loop economy. But it needs to be rigorous, the most impressive aspect of 0G is that it is a full stack infrastructure: EVM compatible L1 chain+decentralized computing+ultra fast distributed storage+data availability layer. Speaking of human language, it not only trains a model, but also provides underlying computing power, storage, and verification base for non AI agents. In addition to decentralized computing power training for large models, the main network used by 0G for AI agents has already been launched, and the ecological accelerator is supporting products with real gold and silver. The official recommendation is promoting Ghast AI, which is interesting and similar to a high-level blockchain version of Lobster. It is still in beta testing. If you want to play for free, you can follow the official recommendation @ Ghast_SAI. Only when the water rises can the ship rise. At this stage of market downturn, many prices are distorted. At least, the product can still be played and tokens can be freely withdrawn. At this point, if a project can survive, not pretend to be dead, and continue to update and iterate products, it can at least be put on the "observation" list. Wish us good luck! / Author: Anymose | A Soft Core Science Popularization Writer This article is for educational purposes only and does not constitute any investment advice. Always remember DYOR!