The prestigious American magazine MIT Technology Review revealed on 21 February its annual list of the ten disruptive technologies for 2018. Artificial intelligence, smart city, cryptographic tool … We have selected five digital innovations from this list. The prestigious MIT Technology Review published on February 21st its list of the ten disruptive technologies for 2018. Each year since 2001, the American magazine compiles innovations which it judges decisively in the coming years for the companies and our ways of life. Industry, medicine, digital … The sectors are varied. We selected five digital technologies from . the list 1 — THE SMART CITY It’s an amazing choice as the smart city covers different meanings. The review uses the Quayside project by Sidewalk Labs and Waterfront Toronto as an example. We presented this project in October 2017 to his announcement. Quayside will rely on an extensive network of sensors to collect data on air quality, residents’ activities and noise levels. Modular buildings, autonomous vehicles, robots delivering the mail … The promises are numerous. The really interesting part may be elsewhere: Sidewalk Labs should leave its software open access, allowing other companies to create services from the infrastructure. “In the same way that we build applications for mobile phones”, describes the MIT Technology Review. 2 — ARTIFICIAL INTELLIGENCE FOR EVERYONE The fashion around artificial intelligence also reflects a reality: its democratization. Artificial intelligence services based on the cloud make the use of this technology less complex and less expensive. This break is already in place according to the MIT Technology Review with technology giants selling their tools to smaller companies. In January, Google launched, for example, to popularize machine learning through pre-trained systems. In October 2017, Amazon Web Services and Microsoft created an open source interface, Gluon, to enable developers to build machine learning models for the cloud and mobile applications. Cloud AutoML 3 — GENERATIVE ANTAGONISTIC NETWORKS If we want to be more technical about AI, the MIT Technology Review highlights the performance of a new method of deep learning: generative adversarial networks (GaN). When two artificial bits of intelligence clash, they arrive at a better result and do it faster and more efficiently. In October 2017, for example, Nvidia led an AI to create fictional human faces by feeding celebrity photographs (see video below). With this type of process, autonomous vehicle AIs could create images of pedestrians themselves to practice recognizing them without having to go on the road. 4 — REAL-TIME TRANSLATION Understand any foreign language in real time through headphones. It is a recurring fantasy of science fiction that would already be within reach of the general public. In October 2017, Google announced the commercialization of its Pixel Buds: $ 159 earphones capable of translating more than 40 foreign languages in real time thanks to Google Translate. However, they still have to work with a smartphone. Equipped with a microphone, the headphones pick up the voice of the first user while the other party, with a smartphone, listens to the translation in real time. This is also played in the connected headphones. 5 — A CRYPTOGRAPHIC TOOL FOR PERFECT PRIVACY In a majority of blockchains, including Bitcoin, transactions are anonymous but public. A user can be “tracked” by cross-checking information about his payments. At the end of 2016, developers of the virtual currency developed a new cryptographic tool to get around this flaw. The method is called zk-SNARK for “zero-knowledge succinct non-interactive argument of knowledge” … More vulgarly, it allows a user to prove that he holds information without disclosing it, so as to limit the identity theft or security breaches. It was adopted in 2017 by the blockchains of Ethereum and JPMorgan Chase. Zcash 5 Misconceptions About Blockchain There is no doubt that blockchain technology is a source of value and companies are increasingly interested in possible applications of the blockchain. But there are still many misconceptions about technology. 1. The myth of immutability The term “immutable” which, in the case of the blockchain, means “unalterable or cannot be changed” is technically impossible, the researchers said in their report, citing two ways of modifying the blockchain. “The first is to recalculate the chain, either in its entirety or upstream of the point where an adverse event occurred. The operation clears and recreates the history. An event of this kind had occurred in the early days of bitcoin, “they explained. “The alternative is to duplicate the chain. This time, the historical code and transactions are kept. But after this operation, the software works differently. The The researchers also note that, from a technical point of view, block chains subject to authorization are easier to modify and have much fewer nodes than public block chains, especially in the early stages. But as a result, “they are also technically more vulnerable and more likely to end up under the influence of criminals or fraudsters who manage to seize the network’s credentials.” In practice, the security and governance mechanisms that apply to the network help to control risk. “Ecosystem participants need to understand that it is not the technology as such that protects the blockchain records against the changes, but the way the network is designed, implemented and exploited. That’s 2. The myth of disintermediation and decentralization According to the researchers, “cost reduction and efficiency gains are the main reasons why a company is moving towards a blockchain solution.” Often this involves deleting an existing intermediary. “The company is wondering why it is a third party to trade while it can deal directly with its business partners. Some situations will allow you to do this. But it is a mistake to believe that there are no trusted intermediaries in blockchain networks or that these networks are entirely decentralized. As the researchers point out, in practice, blockchain networks never work completely without intermediaries. These are distributed networks and, as such, they maintain a certain level of centralization. In addition, the chain can bring new intermediaries and existing intermediaries can disappear without the knowledge of all. 3. The myth of zero trust “The two main block chains — Bitcoin and Ethereum — have proven that it is possible to exchange” value “between people and entities who do not know each other or who do not trust each other.” said the researchers. But they also showed that it was a myth. Neither network is totally devoid of trust. “Participants need to trust the continued functioning of these networks on many levels. For example, they must trust math and cryptography and they must believe that the code will always be executed as intended, “the researchers added. 4. The myth that chain blocks are “machines of truth” “Many solutions based on blockchain networks claim that they can prevent fraud and guarantee the provenance of goods in the physical and digital world,” the researchers said. To a certain extent, this assertion is not wrong, since blockchain transactions are extremely difficult to manipulate, and any attempt is quickly spotted. “But no technology, whether blockchain technology or another, can be considered a complete deterrent,” they said. “It’s important to keep in mind that a chain of blocks cannot protect against all fraud and that only a blockchain network cannot guarantee where the physical goods come from. According to them, it is necessary to distinguish the pure traceability from that which provides the proof of the source “. 5. The myth of transparency “Making transactions more transparent is one of the biggest benefits of blockchain networks,” said the researchers. But for most businesses, transparency is as much a curse as a blessing. “Aside from scale issues, the biggest technical challenge developers face is confidentiality,” they said. CIOs should keep in mind that in a typical blockchain stack, all channel content is visible to all participants. According to the researchers, it is one thing to provide proof of provenance or to guarantee integrity. But, it is another to preserve data confidentiality, commercial transparency, and tractability.
