Original source: Grass X account
Author: Grass
Compiled by: TechFlow
Information Express
Yesterday, we announced that users can now link their Solana wallet addresses to their GRASS accounts. This is a significant moment not only for GRASS, but for the entire AI industry. Our plan has always been to revolutionize the development of artificial intelligence, and the importance of this update lies in:
Users who contribute resources will receive corresponding rewards
The Grass network is able to track the data that each node has scraped from the network. When this data is used for AI reasoning or training, there will be a clear link linking the scraping behavior to the dataset. When each node is associated with a wallet address, we are able to assign the value of the AI back to the specific node that provided the data.
This is a step toward a fully decentralized method of tracking resource usage. These records cannot be manipulated or hidden by any company, which means AI will be more fair and equitable, and individual users can actually get a share of the benefits.
Considering that you’re already providing these resources to companies for free, the only difference is that now you’ll actually be rewarded.
Laying the foundation for transparency in AI training data sources
With this feature, the network will not only track which node has grabbed which dataset, but also the specific source of the dataset - that is, where the data came from on the network. The moment a dataset is grabbed, the network will automatically compile metadata about its source and the node that grabbed it, and record it in batches on the chain, forming a permanent record of the data's provenance. This is a feature that other data supply solutions do not have, and we believe it is a key link in solving the biggest problems in AI today.
Making AI more transparent and fair, and less centralized
Think about the biggest issues affecting AI right now. For developers, there’s no way to ensure that the dataset you’re using isn’t contaminated unless you re-scrape it yourself to confirm. For users, there’s no way to tell if the agent you’re interacting with is biased unless you can see the data the developer chose to train it with. And for everyone, we face the risk of a handful of companies dominating AI and creating a monopoly over the most important technology.
Grass aims to solve all of these problems. When developers can verify beyond a reasonable doubt that a dataset comes directly from the intended source, they no longer have to worry about data contamination. When this information can be made public for individual models, users can know without a doubt whether they are being presented with real information or fake news. Most importantly, when a system is created to distribute the value of AI to the millions of people who collectively operate the network, rather than a handful of large companies, we can work together to make it more fair and less exploitative.
Other developments
In addition to the positive impact on training data acquisition, Grass also breaks technical barriers by introducing Live Context Retrieval. This refers to fetching data in real time at inference time to provide context for LLM. This is a feature currently under development and does not exist elsewhere yet, and we will release more information soon.
We believe a better world is possible and that the future is now. There is still time to save AI and we are committed to it. The mission is ongoing. Thank you to all our users for your vision and support of Grass.
