“At a time when everyone has at least one smart phone, every family is buying a smart home and is constantly shocked by the progress of artificial intelligence, many people may not realize that the driving force behind these is actually software. Mark Anderson, a senior investor in Silicon Valley, has long predicted that software is eating the world. But in the era of mobile Internet, Internet of things and artificial intelligence, how has software gradually changed the industrial logic of mobile Internet, Internet of things and even artificial intelligence? “
At a time when everyone has at least one smart phone, every family is buying a smart home and constantly shocked by the progress of artificial intelligence, many people may not realize that the driving force behind these is actually software. Mark Anderson, a senior investor in Silicon Valley, has long predicted that software is eating the world. But in the era of mobile Internet, Internet of things and artificial intelligence, how has software gradually changed the industrial logic of mobile Internet, Internet of things and even artificial intelligence?
What the iPhone really changes is the global software industry
Before Anderson cheered that slogan, in October 2007, apple made a decision that seemed insignificant at that time, but was as great as the iPhone: Apple promised to release software development tools to all third-party software developers, so that all third-party developers could develop applications similar to Apple’s native applications.
However, when the first iPhone was released in 2007, jobs still looked down on third-party applications. In his view, only Apple’s own applications can be called “desktop level” applications. Of course, third-party application developers can develop applications for apple, but they must follow the standards of HTML and javasript. In other words, jobs just wanted third-party developers to develop gadgets for the iPhone, not native apps.
At that time, there were app stores of a certain scale on Microsoft’s windows mobile, palm and Nokia’s different mobile platforms. But there are some common problems in these app stores. First, the rampant piracy leads to developers selling apps at high prices.
Take windows mobile as an example, a clock application that can provide weather forecast and flight information costs $75. The reason why the price is so high is that piracy is rampant. The platform operators do not have a set of perfect policies to restrict users’ piracy. Developers can only rely on themselves to maintain their income at a high price to offset the pressure from piracy.
Secondly, in order to curb piracy, some mobile platforms make application development more difficult, which makes the whole application development release process complex and affects the enthusiasm of developers. For example, Nokia’s Symbian system, in order to ensure the stability of application operation, Nokia adopts micro kernel architecture, strict control of memory leakage, preemptive multi task scheduling and other measures to restrict the third-party developers. At the same time, Symbian applications are developed based on C, which has high technical requirements for developers and high threshold to exclude a large number of developers.
These problems are negative teaching materials for apple and jobs to learn from. They hope to unify the fight against piracy (closed) and encourage development (open). In February 2008, with the release of software development tools to third-party developers, Apple also launched two mechanisms for application development:
First, the application is signed by apple;
Second, all third-party application notifications are pushed by Apple server.
The former uses Nokia’s signature mechanism for reference, but it is more friendly and convenient for developers. It only costs $25 a year to develop applications for iPhone. The latter uses Microsoft’s exchange information push architecture, which greatly reduces the possibility of malicious push messages by unscrupulous third-party developers.
These two sets of mechanisms perfectly explain Jobs’s closed Aesthetics: creating a stylish app store with iPhone style quality control.
On the one hand, the hot selling iPhones make developers have to develop applications for the majority of iPhone users. Even if they are free applications, the free effect of scale is still very huge. Instagram is an extreme example. The filter app, developed and operated by six people, appeared on the IOS platform for free. In less than three years, the number of users exceeded 13 million, and was eventually acquired by Facebook for $1 billion. On the other hand, due to the influx of a large number of developers into the IOS platform, how to highlight their unique value in the fierce competition of app store has become an important proposition for application developers, which also improves the application quality and quality on the IOS platform.
Apple’s app store went online in April 2008. A year later, the number of downloads exceeded 1 billion. Now, the number of downloads per month is as high as 2 billion. In six years, apple almost subverted the software industry. In 2014 alone, third-party developers made $15 billion in revenue through the app store. Of course, according to the 30% rule, apple received $4.5 billion in revenue, which is 50% higher than that in 2013. Although the data for 2015 has not yet been released, according to Apple’s financial report, the app store brought Apple $3.3 billion in revenue in the last three months of 2015.
Although Google, which only followed apple, also took great pains to build its own app store, Android’s fragmentation and localization problems made its revenue unable to compare with apple. More importantly, while Android manufacturers at home and abroad blindly accumulate hardware to “kill” the iPhone, they do not mention the quality of their own application market. In this sense, Android manufacturers still have a long way to go to subvert Apple’s leading position.
Software is the starting point of Internet of things, and the killer weapon is still software
If software makes mobile phones a new computing platform, then software makes the Internet of things an experience platform. The first step of the Internet of things is networking, which is also the main reason for the emergence of “Networking” and “intelligent” Internet of things devices in recent years. According to the latest report released in October 2015 by Linley group, a technology industry analysis company, the number of Internet of things devices used in the world will reach 1.9 billion by 2020.
The agency believes that although the industrial Internet of things develops faster than the consumer Internet of things on the whole, the consumer Internet of things will eventually become the main force in this market (unless otherwise specified, the Internet of things discussed in this paper are all consumer Internet of things), and the scale of the Internet of things market will surpass the smartphone market in 2023. The sluggish growth of smart phones is also supported by the report of Gartner, a market research company. According to the latest report, the growth of global smart phone market in 2016 is expected to be less than 10%, and it will be the first single digit growth. The agency expects the global smartphone market to grow by only 7% in 2016.
When the Internet of things devices complete this step of networking, the next step is how to improve the functionality of the device. Different from smart phones and tablets, the update frequency of Internet of things devices is not high. In essence, home Internet of things devices, from TV to audio, are durable goods, not consumables. From this point of view, how to make an IOT device, such as your audio, maximize its function and even continuously improve its performance has become a big problem in today’s rapidly changing technology. Wi Fi audio SONOS provides a set of answers, the core of which is software.
In the field of Wi Fi sound, SONOS has a great influence. From the perspective of product iteration speed, SONOS is a typical Internet of things equipment manufacturer. In other words, what they want to do is not the FMCG that users change every year, but the “XXX” products that they want to stay at home. This is in line with the user’s positioning of home Internet of things devices, and it is also powerful and even iterative.
Founded in 2002, SONOS has the software gene of Silicon Valley since it came out. When the iPhone started the smartphone revolution, SONOS gave up its audio control terminal and developed software applications supporting mobile terminals. In a sense, SONOS benefited from Apple’s app store. Relying on the application running on the mobile phone to control, not only for SONOS, but also for users, bring more imagination.
At the end of the year, SONOS launched a function called “tureplay automatic tuning”. In short, “tureplay tuning function” is to let the audio play the specially made test sound, measure the room structure through the microphone of iPhone, such as the reflection of walls and furniture on the sound, and then automatically adjust the sound effect settings of SONOS audio, so as to create a room most suitable for today’s room The atmosphere of listening to music.
Strictly speaking, it’s not new to tune the sound through a microphone. Traditional high-end home hi fi audio is usually equipped with a microphone specially used for tuning. However, SONOS skillfully uses the voice of mobile phone microphone, and then calculates and matches it through cloud based algorithm, and finally presents a unique sound effect that belongs to the current environment. This kind of problem-solving concept is different, and the way of software defining hardware enables users to save the cost of purchasing hardware. This function not only supports the new generation that will be released soon PLAY:5 Audio is also suitable for several products in the past, including PLAY:1 、 PLAY:3 And the previous generation PLAY:5 .
On the other hand, domestic audio manufacturers generally use “intelligence” and “Networking” as gimmicks, or cooperate with e-commerce platforms to create the so-called Internet of things entrance, or rely on voice interaction to achieve some kind of home Internet of things “transit station”, but the essence of audio is music experience, including both content and sound quality.
According to the latest data of RIAA, the revenue of streaming music such as spotify and Apple Music in the U.S. market exceeded $2.4 billion in 2015, surpassing the paid download of digital music represented by iTunes for the first time. When music enters the era of streaming media, the competition for content has become a part of the competition for user experience. At the same time, the competition for sound quality will become more and more fierce with the emergence of more and more intelligent sound. Considering the price, there is no debate about the hardware of smart audio, and we still need to refer to SONOS software defined audio solutions.
BOT: a new software in the age of artificial intelligence
At the just concluded Microsoft Developer Conference, Satya NADELLA, CEO of Microsoft, defined the keynote of her keynote speech as “software in the age of artificial intelligence — bot”, which is also the abbreviation of “robot”, a tool that can run autonomously and manage automatically. To some extent, the Internet can also be called BOT net, because more than half of the Internet traffic is brought by BOT.
In the context of the imbalance of information supply and demand of the Internet and the rise of platform economy, the outbreak of BOT seems to have no suspense. However, from a deeper perspective, the rapid development of BOT in recent years has an inseparable relationship with the fact that artificial intelligence has increasingly become a new technology infrastructure of the whole Internet. In a sense, BOT is a new generation of software shaped in the era of artificial intelligence, and it is developing rapidly “Eat the world.”.
First of all, the progress of speech recognition, image recognition and natural language processing provides technical soil for “new software” – bot. Take image recognition as an example. In the 2012 image classification competition Imagenet, the research team supported by deep neural network won the first place and reduced the error rate to less than 20%, which shocked the giants including Google and Facebook. Then, Google bought the team and let Geoff Hinton, the godfather of deep neural network, work in Google. With the help of deep neural network, Google’s image recognition level has been greatly improved, and the error rate has been reduced to less than 10%.
In 2015, based on the computer vision system of deep convolution neural network (CNN), Microsoft engineers even reduced the image recognition error rate to 4.94%, even exceeding the human recognition performance. At the same time with image recognition, speech recognition and natural language processing are rapidly improved. These technologies also make the presentation of BOT more intelligent, such as entertainment bot that infers the age of a person by image recognition, or home bot that creates voice interaction by speech recognition, etc.
Secondly, deep learning and machine learning are more and more concentrated in the cloud. As cloud service providers including Microsoft, Amazon and Google open source machine learning architecture or suite one after another, the cost for developers to develop a bot is undoubtedly reduced a lot. In addition, a large number of cognitive service tools based on the cloud can be called at any time, BOT is becoming more and more intelligent.
Thirdly, the popularity of message application provides a new scenario for BOT distribution. Benedict Evans, a silicon valley investor, has repeatedly pointed out that message applications based on mobile platforms will become the main variable in the evolution of mobile terminals. As mentioned above, thanks to the development of natural language processing technology, BOT in message applications can identify users’ intentions and emotions, which makes Facebook and wechat see the possibility of subverting IOS and Android.
Facebook’s M is a bot built into messenger. You can chat with him, search for questions, book flowers, buy coffee and so on. Although it is not yet open to the public, in the evaluation of many internal test users, the user’s evaluation of this AI driven messaging application is very positive. On the one hand, this positive emotion will increase the user’s stickiness to Facebook, and on the other hand, it will further train m to make it more intelligent.
How can software eat the world step by step?
Comparatively speaking, wechat is more ambitious. It wants to seize all the scenes of users. Recently, wechat released the weui library, and these controls of the official UI library have shown the ambition of wechat to end the mobile app, which is also consistent with the rumors of wechat application number. Nowadays, the vast majority of life needs of an ordinary individual can be met through wechat, and a large number of work needs increase the stickiness of wechat. In the future, when there are more and more “apps” – BOT based on wechat, how many people will download apps again?
In 2014, Gartner pointed out in the “top 10 strategic technologies of 2014” that software is defining everything in the world, which coincides with Anderson’s slogan. According to Gartner’s definition, the so-called strategic technology refers to the technology that will have a significant impact on enterprises in the next three years. From the time node of 2016, software has subverted the industry rules of smart phones, just like the re innovation of iPhone on mobile phones, software has also given a new development direction to the future of the Internet of things, just as SONOS redefined sound. At the same time, software is evolving towards BOT and will continue to define the world.
Editor in charge: PJ