Sahara AI has completed a $43 million financing, led by Binance Labs, Pantera Capital and Polychain Capital. This investment amount and investment institutions are top-notch. Partners include Sean (Xiang) Ren, an associate professor of computer science at the University of Southern California who graduated from Zhejiang University. @xiangrenNLP
and Tyler Zhou @tz_sahara, former investment director of Binance labs who graduated from the University of California, Berkeley.
What does Sahara AI @SaharaLabsAI do? The official website says that it is a decentralized AI blockchain platform. The details are as follows:
1. Data layer: Provides digital watermarks, basic data record traceability and permission management to facilitate data ownership confirmation.
2. Transaction layer: Use AI technology for POS consensus algorithm.
3. Execution layer: points out that traditional ML algorithms and LLM algorithms can be used for AI training.
4. Application layer: A platform for generating, trading, and accessing valuable data sets, models, and agents. Agents can be generated without code.
After collecting news reports, interviews with the founder (https://chaincatcher.com/article/2120262), official website introductions and some information searches on the founder, there are still many uncertainties. The evaluation of this project is as follows:
1. In the interview, it seems that the founders do not have a clear understanding of what they are going to do, and there is no white paper or economic model introduction on the official website. For example, Tyler Zhou described that "to build a "Personalized Agent", there are two prerequisites. The first is to have your own database, and then train the AI Agent based on your own data to enable it to achieve some desired capabilities; the second is to have the relevant infrastructure and tools to build the Agent." It seems that we know what they do, but at present we only have a vague idea of what capabilities they provide, but it is still unclear what tools they provide, and what infrastructure they provide. I haven't seen it.
2. Tyler Zhou said that "Sahara will launch its first consumer product in April or May." Currently, there are no products available for trial on the official website.
3. Sahara said it has served more than 30 companies including Microsoft and Amazon. Considering Sean (Xiang) Ren’s background, he also worked on NLP and related data sets, and there are “200,000 AI-related staff”. So these promotions may have been about cooperating with these companies or contributing data sets, and copying some existing things in web2 to web3. But I didn’t see any work in web3.
4. Data watermarks at the data layer are some concepts of data rights confirmation that have been overblown in the web2 field, and the execution algorithm layer is those things of autoML. They are not new things in themselves, so don’t make them a patchwork. It is necessary to further clarify the specific innovations and application scenarios of the project in the field of Web3 and AI integration. Subsequent product releases and actual application cases will be the key indicators for evaluating the potential of the project.