Resurrection of Developments: Atmatrix To Utilise Blockchain To Create A Global Shared AI Network
Most people who meet Feller Gao for the first time are usually impressed by his passion and vitality. Feller’s daily routine has become irregular for most of us. In the past few months, he would wake up at 3 or 4 am to fly to another city to attend meetings and present at science and technology conferences.
As founder and visionary leader of Atmatrix, Feller, born in 1974, always tells people around him, “I feel like I’m a Post-90”. Feller graduated from Shanghai Jiao Tong University with a Master’s degree, majoring in Artificial Intelligence. He joined Huawei Technologies Co., Ltd. in 1990s and worked there for 18 years. Feller was mostly responsible for the innovation and incubation teams. He is also an active investor within the Angel Circle.
In 2013, he learned about Bitcoin for the first time. In 2015 an opportunity by chance helped him “understand blockchain intimately” when Shenzhen Huawei headquarter invited Vitailik, the founder of Ethereum, to share his experience. Since then, Feller started to work and innovate onto blockchain.
In his view, the development of blockchain is happening in three phases. Firstly, blockchain technologies of the first generation (represented by Bitcoin) brought about a decentralised cryptocurrency, which helped solve the problem of building trust without any third party. We call this the digital asset phase. Secondly, blockchain technologies of the second generation (e.g. Ethereum) brought us virtual machines and smart contracts, thus removing the ambiguity of Bitcoin blockchain and nurturing a large number of decentralised applications. Thirdly, the next generation of blockchain technologies, in Feller’s opinion, will be represented by AlaaS (Artificial Intelligence as a Service) such as DBot distributed robotics and Oracle cross-blockchain interoperability technologies. A large number of distributed robots will empower different blockchain to interact and interoperate.
Feller says he “has a passion for AI”, but why does he want to combine AI with blockchain? As for this question, Feller says:
“AI will become the main productive factor of our society in the future. It will release humankind from miscellaneous work and blockchain is designed to solve problems related to productive relations which will change the distribution, transformation and management of assets.”
After working on AI for many years, Feller firmly believes that the fundamental problem of AI industry is the separation and lack of collaboration between key stakeholders.
“There are competitions between AI infrastructures and business entities. For example, Google has Alpha Go, Baidu has Baidu AI and Tencent is also working on its own AI Lab. In fact, this is a kind of repetitive work that is wasting resources. All DApps stay on online blockchain and cannot connect with offline, traditional AI infrastructures, which means they cannot be used directly.”
This is why Feller created Atmatrix; he hopes, through this network, to solve problems already noticeable in AI industry from two aspects by:
- Using blockchain to create a shared platform of AI,
- Channelling the power of AI into blockchain.
Standing on the shoulders of giants: Leveraging blockchain to connect AI worldwide.
Simply put, Atmatrix is comparable to how Ethereum changed the blockchain but for the AI industry. It provides a consensus-based, distributed and virtual AIaaS (Artificial Intelligence as a Service). With the help of blockchain economics, it manages to use AI and robots worldwide, thus creating a shared AI platform.
However, how can Atmatrix solve problems? Our founder believes that we should ‘stand on the shoulder of giants’, by taking advantages of current technologies related to AI and blockchain. Regarding AI, we should use bots, both at home and abroad, to establish an API store of software robots and DBot platforms, and ensure they are open for collaboration. All types of AI open-source technologies and relevant communities can connect with each other to build more. Relating to the blockchain field, we should use blockchain Oracle technologies to provide cross-blockchain API services for Bitcoin, Ethereum and DApp services. Furthermore, we should use and improve Graphene technologies to establish a principal blockchain that is compatible with EVM (Ethereum Virtual Machine). Lastly, we should keep technologies open to connect different types of blockchain open source technologies and related communities.
Solving technical problems one by one
Atmatrix blockchain technology includes the main blockchain and other services that enable the main blockchain to communicate with Ethereum, Qtum, Metaverse, Hyperledger and other global blockchains. The main blockchain means Graphene technologies with smart contacts are compatible with EVM as the basis. It also uses Oracle technologies to provide AIaaS and the blockchain with interoperable services.
As the blockchain develops in general, its performance and scalability have always questioned. However, a shared AI needs blockchain of high performance to operate effectively. So how to ensure the efficiency of transforming data on those chains? Feller says that this problem is also the one that Atmatrix needs to solve. At present, the solution is “to leave”.
“Blockchain focuses on accounting information. It invokes data and algorithms and separates them. As a result, two systems appear. One is for data, and the other is meaningful information. This is also a technical challenge.”
A fierce combination: adding private blockchain into public blockchain
Both public and private blockchains have existed for a long time. However, Atmatrix chooses to combine them: starting from a public blockchain and then adding private blockchain into it, thus creating AI private blockchain as an industry first.
Feller says he chooses to do it this way because the development of these two kinds of blockchains can bring out the best of both worlds,
“Actually, the most advanced technologies always come from public blockchain programs. While private blockchain can apply advances on public blockchain into blockchain industry. Therefore, blockchain can become more commercialised, which makes it easier for people to accept and adopt blockchain. Public blockchain drive technologies to become more and more advanced while private blockchain puts these technologies into real-world applications, and further support public blockchains. However, if those applications don’t work well, several years later, people will know that maybe this is just the business within a bubble.”
Therefore, Atmatrix chooses to start from a technically advanced public blockchain and then add a business-centred private blockchain within it. “This is just like how Ethereum appears at first sight, but then EEA (Enterprise Ethereum Alliance) gets formed. At present, some of the top 500 companies have already signed strategic cooperation agreements with us to create an industry alliance of AIA Atmatrix.” Feller added.
Atmatrix can play an active part in the application and development of AI. Feller provided us with two short examples: Using a current application to empower the normal application to become an AI application. The other example is to help existing robot manufacturers to improve their AI abilities.
“When hardware manufacturers cooperate with AI companies, they may meet difficulties such as small production and complicated customization. AI Market promoted by Atmatrix can help those companies to make qualified robots to become artificially intelligent.”
AI is becoming increasingly popular in the past two years. Moreover, how can Atmatrix, a new platform, deal with combat and competition coming from giant companies? Feller says that Atmatrix is designed to establish an open community. Feller said, “Atmatrix is not intended to gain exclusive advantages. It is open and shared. All AI manufacturers are welcomed to join in the future.”
As for his next step, Feller tells us that as Atmatrix develops, its technical difficulties are mostly resolved. He will launch AI Market next. The first series of demonstration programs based on virtual Ethereum machines will be provided shortly.
What is the biggest challenge for Atmatrix in the future? Feller says it is the industrialisation and standardisation of Atmatrix.
“It may be the establishment of customized AI for different industries. The industrialisation of Atmatrix is not an easy task. So we need continue our trails and explorations in unique and innovative business models to release our workload and keep costs on a manageable level. For example, such a model may appear in the future: we provide a basic platform and let other teams industrialize AI for a certain industry.”