Text | Joey
Musk is ready to open again.
Just a few days ago, Musk's brain-computer interface startup Neuralink officially opened its recruitment for a clinical trial director. Since Neuralink's inception in 2016, the company has been working to implant chips into the human brain to propel people who are unable to walk due to trauma to have access to possible recovery opportunities.
In fact, Musk's innovation in this field of human-machine collection does not stop there. As early as the middle of last year, he began to dabble in the field of artificial intelligence robots, hoping to create truly intelligent humanoid robots. Obviously, this man who wants to retire on Mars is eager to truly change the shape of human society by himself, to explore the field of human-machine intelligence and break through the bottleneck of science and technology.
There is no doubt that Tesla has always appeared in front of the world as an innovator in technology and a reinventor of the car. The end point of its innovation and the goal of reshaping is not to create an electric car, but more importantly, to build the car into a robot. This is the same as Musk's dream of a humanoid robot, a humanoid robot, a model robot, and ultimately a smarter robot.
So, what are smarter robots? Reading the instructions and completing the corresponding operations can only be the threshold for entry, and being able to have a deep emotional interaction with the person who issued the instructions seems to be able to truly meet the standard. Of course, such a goal is not only what Musk and Tesla want to achieve, almost all automakers who want to dabble in the field of intelligence are moving towards this goal.
Developing smarter cars is Roewe's tradition
There is no doubt that new energy vehicles are now the trend of the times, and almost all new car brands and established giants are sprinting towards this new field. Just in the past 2021, the three major BBA of traditional German luxury brands have initially completed their product layout in the field of new energy and formed a more complete product matrix.
These products, without exception, have the performance performance closest to the ceiling in the mainstream performance of pure electric vehicles. However, in terms of human-computer interaction, compared with Chinese car companies who dare to break through the traditional shackles and find new forms of software and hardware, BBA seems to be taking a different path.
In these automotive products, human-computer interaction is almost still stuck in the earlier thinking mode, in fact, it is not difficult to understand. After all, building a new energy vehicle that is more like a traditional fuel car seems to be one thing these European car brands are more keen to do.
In China, many car brands are now more concerned about how to make cars smarter. Whether it is a new car-making force or an established automobile company, whether it is a new energy vehicle or a traditional energy vehicle, they all hope to create products with a more valuable sense and a better interactive experience in the field of human-computer interaction.
In this field, SAIC Roewe seems to be a well-deserved domestic automobile brand that leads the trend of human-machine intelligence. As early as 6 or 7 years ago, SAIC Roewe was equipped with early human-computer interconnection technology on the RX5 model. As a result, the RX5 became China's first mass-produced Internet car.
Since then, SAIC Roewe has continuously promoted on this road, and through a series of intelligent systems such as Zebra Zhixing, it has opened a new era of domestic automobile brand application ecology and human-computer interaction mode. Nowadays, a series of car applications launched by new car manufacturers, such as online shopping, car machine payment, and establishing their own ecosystem, have long been proposed at that time.
The times are progressing, and the car machine system has also evolved a more exquisite UI interface, smoother operation logic, faster computing speed, and even deeper command understanding. In such a situation, can Roewe, which has led in this field for a long time, be able to come up with a more advanced vehicle and machine system?
Yes, they already have.
High computing power supports intelligent cross-domain integration
Just in the last quarter of last year, SAIC Passenger Cars released the new Roselle Intelligent Cockpit System. As the first mass-produced multi-core heterogeneous fusion intelligent cockpit, it supports microkernels and macrokernels and SOA-based cloud fusion frameworks, with the ability to converge across domains.
As the second part of the AliOS trilogy, the Roselle Intelligent Cockpit is different from the first generation of in-vehicle operating system, only allowing the car to be online, but can be upgraded through the underlying layer, integrating interactive intelligence, service intelligence and driving intelligence. So what is cross-domain fusion capability?
In fact, in simple terms, cross-domain integration capability is the ability to perfectly integrate multiple functional modules on the car. Through it, previously unrelated functional modules, such as intelligent driving and entertainment, can be integrated into one system. The user will implement the invocation and modification of all functional modules through a unified command system. In short, the operation of the car will become easier.
The lower nanometer count means that the number of integrated circuits that can be accommodated per unit volume will become more, and the performance of the chip will naturally become better. Obviously, Roewe is "bloody" on the chip, after all, as far as the current market is concerned, there are basically few car companies that choose this top-level specification chip.
Of course, this also shows that the Roselle intelligent cockpit system needs a chip with higher computing power to support. But the chip is not all intelligent, after all, the strong computing power chip is more important to dedicate a more intelligent human-computer interaction experience.
How advanced is the semantics of cultivated AI
How do you make the machine understand your words? The current mainstream approach is to release commands through voice. At this time, the performance of a car in the process of instruction understanding and execution will intuitively form a sharp contrast between the advantages and disadvantages of the system in the user's mind, but this is only a comparison on the low-level platform. Because "being able to accept execution" and "being able to understand and choose the most appropriate way to execute" are two concepts.
Such an intelligent command receiving mode is no longer to open the system input instructions, but to be on standby at any time, capture the language and can be analyzed autonomously, and the degree of intelligence has gone up to a higher level.