We can see that AI computing is more applied to cloud computing platforms. What is a cloud computing platform? We modern people often use some face recognition data, or some identification data involving behavioral characteristics. You may upload this data from the terminal device to the cloud computing service platform and add it through a series of complex artificial intelligence algorithms. Some computing clusters calculate the corresponding data and eventually return it to individuals. This is the cloud computing model that is often encountered.
Edge computing is a very recent field of fire. For example, after the birth of Apple, faceID is actually a chip that is very industrialized or has achieved commercialization and can already reach the face recognition level. The main function of this chip is that it is actually Instead of uploading any data to the cloud service, it stores face data on its own terminal and its own mobile phone chip. We can call such calculations as edge calculations. Putting all computing resources and storage data on the ground, edge computing is the future trend of post-AI development.
Today, I'm talking about examples of how blockchains can be combined with edge computing and AI computing. Everyone may have questions. Later, I can elaborate on how the three are combined.
First of all, let's talk about what a blockchain is. Of course, the blockchain that I talk about is not an ordinary person or a newcomer who just understands the industry. Babbitt has been working in the entire blockchain industry for four years. We believe that the blockchain is divided into four layers. The first is that when the company first entered the industry, it believed that the blockchain was a decentralized protocol and information could not be tampered with. It is the previous cognition. Secondly, the blockchain is an economic ecosystem and is a multi-role community. From this level of understanding, many people who do technology may only think that the blockchain is a technology. Finally, they will find that the blockchain is inefficient. The problem of high cost is not a practical, highly concurrent, high-throughput system. We believe that blockchain is not just a technology, but also an economic ecosystem. Third, from the business model level, blockchain is actually a new model of data commerce. Finally, blockchain is a new idea that does not require third parties to endorse. All operating protocols are transparent and the participants will have their own roles. If there is no ideological connotation in the blockchain, it is easy to be considered by people. It is a Ponzi scheme.
Blockchain can be applied to intellectual property, big data and other fields. We believe that blockchain technology has the greatest potential for physical assets in financial assets and has the highest ceiling. Many people may have realized that this change is coming. Former Paypal COO-David Sack once pointed out: If there is a certain amount of space, then all assets can be promoted to the commercialization process and promote the efficiency of the entire industry.
This is also the focus of the original chain. The positioning of the original chain is aimed at the exclusive public-chain agreement in the asset sector. Although Bitcoin only accounts for a small part of the global financial market, if it counts its derivatives, this is a very large market.
I briefly introduce the architecture of the original chain, firstly, the boundary between ownership and uniqueness; secondly, it is based on the UTXO model; again, the contract system that designs unique financial assets transactions; and finally, the original consensus algorithm of the design. Called "Tensority". This algorithm belongs to POW mechanism and really opened the border line between artificial intelligence and blockchain technology. We believe that Bytom is an agreement and the public chain is not licensed. For other DPOS, or some other PBFT-like algorithm, it is a permissive system, and other people can agree to join it. An agreement that is more neutral than the original chain should be added without any permission. We think this is a POW that is eligible.
Computational barriers are the best way to create a public chain ecology. Currently, Bitcoin and Ethereum are two important public chains. It is very difficult or even impossible for Ethereum to cut into the overall process of POS. Computational barriers are the biggest barriers to security stability from other public chains, including POS, which is the simplest way to copy your system to run such a blockchain, the normal logic, if I don't make the currency, just Create applications on the blockchain, then I will choose the chain with high efficiency and low cost. From this point of view, POS is a very easy to copy chain. In addition, the protocol still has the problem of self-renewal iteration, and POW is almost impossible to have such problems.
Many people will ask why do you want to invent a new algorithm? This seems to me a more interesting thing. We discuss edge calculations, and let all of our nodes run as far as possible on some devices in these terminals. Many of our SHA256 nodes are not well equipped. In this case, universal results should be taken into account.
There is a problem of wasting the miner's machine in terms of iterative speed of the miner. If the next generation of mining machines goes on sale, the previous generation of mining machines will become scrap metal. The whole node of the public chain community is an important indicator. The network is strong and whether it is accepted by more people. This is the convenient condition it needs.
How can we solve the common problem? Under the current background, the country has issued a new generation of plans for the development of artificial intelligence and intelligent planning. Artificial intelligence will occupy a very important position in the development of the entire country. At the second tensor calculation level, what we are doing is a component of deep learning. It is a matrix multiplication operation. It is a matrix of three 256 matrix multiplications. The final workload is calculated by SHA256. This is the core of our calculation. The algorithm to prove is to apply a random, or random, result. We can already think of it as a general-purpose algorithm. Any GPU or everyone's mobile phone, smart phone, and all smart devices can be applied to our algorithm to participate in our consensus and get rewards more than the original currency.
There are also equipment and computing power reuse, AI's computing power sharing, has become a more fashionable business model, Vectordash is a company in Silicon Valley, there is a fortune-seeking deep learning company or scholar, no one can rent, you can The use of force to mine will produce a commercial balance. Those who use the chip can choose different tasks under market adjustment. After the birth of the next generation chip, if there is a higher computing power to dig more than the original currency, the machine that is eliminated can completely do these accelerated services. It has greatly improved the ecological robustness of the original chain, and has greatly accelerated the development speed of chips such as artificial intelligence. In addition, there are a lot of convenient operating conditions for mass AI. How to make your equipment machine enter the network faster than the original chain, AI edge calculation is the focus of major semiconductor manufacturers after faceID comes out.
Chips for various kinds of smart devices are expected to be shipped by 2020. The operating nodes may not be PCs or servers, but are all kinds of routers in homes, including AI devices such as cameras and mobile phones. In such a situation, there will be a large share of the market than the original chain. Everyone can participate and the community will become bigger and bigger.
We put forward such a slogan called One Camero One Vote. The surveillance cameras are more natural than the original equipment and have a large base. Its great characteristic is that it can reach the standby state of 7*24 hours, there is storage in the network, and in the case of running more than the original chain, the power consumption only increases by 1%.
This trend will also iteratively develop, and more and more cameras will increase such artificial intelligence attributes in the future. We can promote the creation of Bytom network machines while allowing more people to engage and participate in the original chain.