Blockchain Project Review: Fetch.AI:6.9 An AI Driven Smart Distributed Ledger

Original article by EVALUAPE

For more reviews and information,please visit https://evaluape.io/share to download EVALUAPE APP.

Fetch.AI

A. Introduction

Fetch.AI combines machine learning (ML), artificial intelligence (AI), multi-agent system, and decentralized distributed ledger technology to build an economic network. With tangible guidance and prediction, digital agents for economic active units (such as data, hardware, service, staff, and infrastructure) can do their jobs efficiently.

Advantage:

1. With a better artificial intelligence development background, the team is well established, and its connection to the DeepMind project is the core highlight.

2. Excellent business cooperation capability; the team has reached cooperation with Mobi, European blockchain alliance, etc.

Disadvantage:

1. The team is lack of experience in the field of blockchain.

2. The project development schedule has been overdue, and there are some hidden dangers in the implementation of its technical products.

3. There are no focusing scenarios of the early stage, thus it is hard to establish a data ecosystem.

4. The value discovery based on the combination of token economy and AI data remains at the conceptual stage.

5. The project needs too many variables such as blockchain technology, AI technology, as well as b-end BD implementation, the project mode is too complex , which makes the team strategy and implementation of a great test.

6. Capital pressure: There are 35 members with high academic qualifications in the team. The cost is high, and if the short-term results and financing cannot be achieved, the capital pressure is huge.

B. Industry

High concept combination

The concept of the project involves AI, blockchain, big data, M2M where are all of huge area of imagination in the future. Among them, the AI industry has problems in data acquisition, high cost, data islands, and so on, and the data acquisition equipment of the Internet of Things is experiencing an increase in geometric data which makes the data scattered, fragmented, and there is a problem that a large amount of sedimentation value is not found. Fetch hopes to use the distributed ledger technology and the massive data of sensors, by using economic incentives and AI technology to intelligently match data supply and demand, to improve data automation efficiency and value discovery, thus form an economic network, and then provide prediction and other functions. Among them, the problem solved by the blockchain is the data security and privacy protection, and the AI solves the intelligent matching of the data supply and demand, and the economic incentives reduce the data collection cost. But in the real world, whether these concept industries are mature enough to require coupling is still open to question.

Competing products:

This concept is not uncommon in blockchain projects, and is accompanied by a lot of competing products. For example, the Internet of Things and M2M focusing project Iota, IOTEX and MXC which uses LPWAN for data transactions, and decentralized artificial intelligence autonomous system Cortex and so on. (7/10)

C. Mode

What Fetch wants to solve is the data automate transaction and then to form an AEA incentive network, so as to reduce the transaction friction in digital currency, increase the data value discovery function in the digital world, activate the precipitation of massive data and thus build a digital economic environment. For this purpose, the project combines blockchain, AI/ML and other technologies to achieve its goal.

The architecture of Fetch is divided into three layers: FSL (Fetch smart ledger layer), OEF (open economic framework layer), and AEA (autonomous economic agents) from bottom to top. FSL: the bottom layer of weight confirmation, using the uPOW consensus mechanism and DAG design as the underlying layer. OEF: together with FSL, they form a full node that itself provides autonomous agent support for distributed software entities. AEA: autonomous economic agents, which is the top layer of the digital economy.

Token Economy:

FET token is the main value exchange medium of Fetch network. Any transaction in the network requires an FET as a refundable registration deposit. Collateral is also delivered in the form of FET, which constitutes a mechanism for rewarding network operators.

Usage of token:

1) To help agents and nodes access to the network: similar to the role of deposit, users can mortgage a certain token so as to access to the network. When there are evil nodes, corresponding tokens will be deducted.

2) Medium of exchange: as the settlement token of the network, users can use the token to pay the advertising fee, the usage fee of AI algorithm, and the handling fee of using the network, etc. (6.5/10)

D. Technology

Fetch uses a three layers architecture, where the bottom layer is an uPOW+DAG layer and is extended by sharding. The project claims the platform is able to achieve the characteristics of 10000TPS, low transaction costs, and low resource waste. In the whitepaper, only consensus mechanism and performance target are proposed, the specific details are not described in detail. This situation is similar to many projects that only come up with a variety of consensus concepts, but are still unable to be implemented. In Fetch whitepaper, how can UPOW define “useful”, how to solve the double-spend security problem of DAG, and how to remove iota as the role of the coordinator on the premise that the node is not of a certain scale, all these questions have no answer. In addition, the blockchain development experience of the project team is not enough. If they develop the underlying layer by themselves, its security and maturity need time to obtain market and ecosystem recognition. Probably, they can cooperate with other mature public chain with high performance.

Fetch is now open source, and its main libraries include ledgers, SDK libraries of python and C++, and OEF core libraries. AEA is not yet open source. Among them, the ledger library has the most submission, but the number of forks and collections is only 6 and 18 times, with low heat. In addition, the github link on the official website only has 6 code libraries, while Binance has published other internal code libraries, among which the Fetch.AI Ledger has been updated frequently, but it is unknown why it has not been open source yet.

According to the latest roadmap on the official website, in the first quarter of 2019, the test network will be open only to invited users. Q2 will publish OEF, ledger, multi — party auction ledger computing function. Q3 will release Alpha version and the test network of which all the major features will be launched. In Q4, the main network will be released.

Since there was a quarterly delay in the release of virtual machines in the third quarter of 2018, we hold a wait-and-see attitude on whether Fetch can be completed on time. (6/10)

E. Team

The main highlight of the team is the close connection between the founder and DeepMind (acquired by Google at a valuation of $400 million in 2014). The CEO is an early investor, but he has little to do with technology. He has also founded two artificial intelligence companies, itzMe and uVue, with good leadership. The CTO, an early developer at DeepMind, had 30 years of research and development experience but had left the company before the 2014 acquisition. At present, the announced members number is 35, in addition to a few business promotion personnel and economic researchers, the rest is basically artificial intelligence, software development, engineers and other technical personnel. (8/10)

F. Ecosystem

The community is generally maintained, but the project is primarily targeted at class B users, with traffic coming from Launchpad.

Search Index

The peak of the project search index was in October 2018 and January 2019, when the project conducted its financing. (7/10)

G. Conclusion

Fetch.AI combines machine learning (ML), artificial intelligence (AI), multi-agent system, and decentralized distributed ledger technology to build an economic network. With tangible guidance and prediction, digital agents for economic active units (such as data, hardware, service, staff, and infrastructure) can do their jobs efficiently.

The highlights of the project are DeepMind’s early investors and developers, well-equipped team, and business cooperation ability. However, the concept of economic network of the project is too advanced, which requires the maturity of different industries to reach the resonance level before it can be implemented. Moreover, in the whitepaper, the design of its underlying blockchain and the data supply and demand matching algorithm of artificial intelligence only remain at the conceptual level, which leads to some doubts of the implementation of its products. In terms of data acquisition, whether to use trusted equipment or not and how to choose the initial scenario to do the case are not elaborated, so the implementation prospect is not optimistic.

Hype Score: High

Risk Score:Medium High

Expectation:Medium

Total Score: 6.9

All information in this article is provided for reference only and does not constitute investment advice.


For more reviews and information,please visit https://evaluape.io/share to download EVALUAPE APP.

About us

EVALUAPE is a platform for demonstration and evaluation of blockchain projects. We provide comprehensive database and professional evaluation of global blockchain projects with our extensive experiences, while adhering to the values of community culture, ultimately decentralizing the platform. We desire to become an international, professional and remarkable project evaluation platform.