3D metal printing
3D printing itself has been around for a very long time, but it was used mainly for working with plastic. All other materials – especially metal – were too expensive for this technology. Now 3D printing of metal products has become affordable, widely applicable and will soon change the very system of mass production.
In particular, the missing part can simply be printed on a printer in a single copy, instead of organizing large-scale production. For old cars, televisions and other rare equipment, the problem of lack of spare parts will disappear. Factories will be able to diversify the range of their mechanisms, adapting them to the specific needs of customers.
So, in the UK, a patient was fitted with ribs printed from titanium. The French company Stella Aerospace printed a fuselage panel with a self-reinforcing effect. And Dutch shipbuilders presented the first fully printed ship propeller. And these are just a few of dozens of examples of successful applications of metal 3D printing over the past year.
Artificial embryos
Scientists from Cambridge are completely changing the very concept of creating a new life. They managed to grow a mouse embryo in vitro using stem cells alone. Neither eggs nor sperm were required for this. Stem cells somehow self-organized into an embryo, which, however, is still unable to develop further. Nevertheless, this is the first step towards the creation of viable human embryos by such a non-standard method.
Such research will help scientists understand how to grow artificial organs and edit genes, prevent hereditary diseases and study the development of embryos at the earliest stages.
It will be located along the East coastline of the city, and the local division of the Internet search engine Google Canada will move its head office there. The company will become an anchor resident there and will invest $50 million in the development of the area.
The project will cost Sidewalk about $1 billion, with an additional $1.25 billion to add to the Toronto government. The plan includes a complete overhaul of the transport system (all cars will be unmanned and ride-sharing).
There will be stringent requirements for the protection of the right to privacy, especially with regard to government interference. In addition, autonomous transit lines and the most environmentally friendly energy systems will appear in the “smart” area.
An extensive network of sensors will help improve the infrastructure by collecting information on everything from air quality and noise levels to the daily activities of citizens. Sidewalk Labs will make all its software free so that IT companies can build new, interesting applications based on it.
If the Toronto project is successful, smart neighborhoods will be built in San Francisco, Denver, Los Angeles and Boston.
AI for everyone
Today, the AI market is dominated by Amazon, Baidu, Google, and Microsoft. But the tech giants have begun to put some of their developments on cloud platforms so that other companies can use machine learning for their own purposes. This will allow the industry to develop much faster and help revolutionize medicine, manufacturing and energy.
The problem is that few people can afford to hire AI specialists. Therefore, even the availability of information does not guarantee that it will be able to properly use it. Google and Amazon, realizing this, offer consulting services to help prepare new talent in the field of AI.
They simply had no imagination. Ian Goodfella of the University of Montreal came up with the solution by creating a Generative Adversarial Network (GAN).
Then one of them, which is called a generator, adds a new feature to the already familiar image – for example, a pedestrian has a third hand. The second neural network – the discriminator – must understand whether it has seen a similar image before.
or whether it is a fake. That is, she must understand whether a three-armed man can be real. Over time, the generator learned to fake images so cleverly that the discriminator could not distinguish them from real ones. It turns out that the neural network imagined and created something new and realistic.
GAN has become one of the most breakthrough technologies of this year. Developers from Nvidia, having compiled hundreds of thousands of images of celebrities.
created the faces of people indistinguishable from real ones – despite the fact that these people never existed. Another group of studies were able to produce paintings indistinguishable in style from Van Gogh’s paintings, which he never painted. The technology opens up space for high-quality fakes, but at the same time allows AI to give the imagination it needs to better work with the surrounding reality.
And yet, it’s better than talking through Google Translate, since each of the participants in the dialogue can control the microphone in the headphones and on the phone, zooming in as needed.
Zero emission gas power plant
Startup Net Power has built a zero-emission gas-fired power plant in Houston.
The combustion chamber produces carbon dioxide and some water vapor. This high-temperature, high-pressure mixture is then sent to a gas turbine, where the pressure energy turns a shaft and generates electricity. The cooled gas mixture leaves the turbine, where it is divided into parts.
