Internet 2010

On August 15, 2011, Lu Zhou's Notes - "As a Last Resort":

It was a very difficult decision, but as before, I had to push the button.

At this time last year, the scale of WeChat and Menggu Cloud had not yet reached such a large scale. In a strict sense, only the business of Menggu Advertising had entered the commercial stage. Zhou Wen was in charge of the data processing of the entire Menggu advertisement at that time, and this became the prototype of Menggu Yun's later D+.

Of course, in the future development, even Dream Valley Advertising has migrated related business data processing to D+ of Dream Valley Cloud.

Everything runs smoothly. Dream Valley Advertising has recruited a group of employees in the data post, internally called data engineers. They process data through the visual graphics or data interface provided by D+. This should be called empiricism, and it relies on the empirical knowledge of engineers.

In the beginning of WeChat, Cheng Xuyuan did not overly consider the problems in the middle. It turned out to be a wise choice, under the attack of the penguins, I think pushing the product into iterations is a higher priority than optimizing the data. Before the Spring Festival this year, when data processing was put on the agenda of WeChat, everything became almost out of control.

Obviously, even if WeChat recruits thousands of data engineers, it will never be able to process the data. Because as the amount of data grows exponentially, the amount of data to be processed remains above 90%. Manpower is limited, but data is limitless.

Therefore, artificial intelligence is inevitable. It can replace more manpower.

To extract some concepts from the whole thing, I think, there are only three: data, algorithm, computing power. Even if we don't talk about artificial intelligence, I think, we also need to face these three problems.

Data is constantly expanding, whether WeChat or other companies. Its size is so huge that I can't imagine it. And it is still produced in large quantities every day, constantly consuming the storage hardware purchased by Menggu Cloud.

Algorithms have been around for the most part since the last century, and what we're doing is re-architecting and engineering. This is the question that Wolong, WeChat and Mengguyun of Seven Stars are discussing together.

computing power, computing power. On the road of evolution, it has also reached the point where it has to be considered. General-purpose computing CPUs are destined to be a poor choice in this field. Just as I see people discussing FPGA mining machines in the Bitcoin community, I think higher-level hardware, such as FPGAs and ASICs, is also required for artificial intelligence and data processing within companies.

In the evolution of hardware, artificial intelligence and blockchain are somewhat similar here. Of course, the correlation is not as great as imagined. Or am I still unable to see far? Maybe.

Coming to this point, the following discussion will not be too difficult. I divided it into several layers: application platform, algorithm implementation, framework interface and hardware.

The hardware machine is the main part of the computing power and the physical foundation; the framework interface realizes the channel between the hardware and the software, and exposes the interface that communicates with the hardware to the upper layer; the algorithm is implemented on the basis of the existing theory, Maybe we will try to do basic recognition of graphics, speech and text; the application platform, here should refer to Menggu Cloud, which encapsulates the entire bottom three layers to provide developers with computing power.

If cloud computing is to virtualize computing power into a stream, and users can turn on the switch to obtain computing streams, then, I believe the same is true for artificial intelligence, and it can also be understood as a new extension of the cloud.

Users flip the switch to take advantage of artificial intelligence, regardless of hardware, interface, or algorithm, just like the cloud, plug-and-play. I think this is a good vision.

What about plugging it in?

I think it's up to you to decide.

......

Little Leaf: I have read it, but I don't understand it. Please advise~

https://

Genius website address: . Mobile version reading website: