Occupied alphabet Google doing? -Google, AI-IT information

Google may be the world's most amazing company, Facebook with a Pegasus stars do, but on the degree the degree of madness is actually far weaker than Google. Google brings together the smartest people in the world, simply say, crazy is certainly not appropriate, they must be a kind of intrinsic quality supporting Google do this thing today, and tomorrow to do the thing.

Google doing?

Google the most radical technology company on the planet has recently reorganized the own, the company changed the name of the Alphabet (alphabet), alphabets of various companies varied, not hard thinking does not understand that this company is doing, only back may have a huge impact on the future of artificial intelligence in order to better understand the alphabet on the company in the first place.

Alphabet on the company and its products now are as follows:

A–Android / AdSense / Analytics / Ara / AdMob / Alerts

B–Blogger / Boston Dynamics / Books

C–Calico / Cardboard / Capital / Contact lenses

D–Drive / DeepMind / Design / DoubleClick

E–Earth / Express

F–Fiber / Fi / Flights / FeedBurner / Firebase / Finance

G–Google (duh) / Gmail / Glass / Groups

H–Hangouts

I–Images / Ingress / Inbox / Invite Media

J–Jump

K–Keep

L–Local / Loon

M–Maps / My Business / Makani

N–Nexus / News / Now / Nest

O–Offers

P–Plus / Play / Photos / Picasa / Pixate / Patents

Q–[Nexus] Q

R–Refine / reCaptcha

S–Search / Self-driving car / Shopping / SageTV / Stackdriver / Skybox / Skia / Scholar

T–Translate / Tango

U–URL shortener

V–Voice / Ventures / VirusTotal / Video

W–Wear / Wallet / Web Toolkit / Wing

X–Google X labs

Y–YouTube

Z–Project Zero / Zagat

This addition to it is easy to understand the various software products such as Android, Gmail, Youtube, and Blogger, Search etc in addition, special surprises the Calico (medical), Nest (smart home), Self-driving car (automatic), Boston Dynamics (robot), the acquisition of these companies basically put on very large, tens of hundreds of millions of dollars, But normal can not see their relevance, robotics and medical companion, many companies like Boston Dynamics that are not able to see in the short term commercialization possibilities. Google the company has gathered almost is one of the most intelligent people on Earth, individual companies may be spur of the moment, but no methodical to do things that it hardly, then what exactly is it you want to do?

To switch to the perspective of artificial intelligence. We to imagine Xia, if we really of build has a super brain, this super brain in application Shang will has difference, may reflected for AlphaGo and human Xia go, also may reflected for automatically driving car,, but in algorithm, and data, and based facilities Shang actually rely on Yu is similar of things, and this super brain of intelligent degree rendering out apparently of data drive features---data more more, intelligent degree more badly. This time assuming the company's leaders are a group of ambitious people, in the past had this ambition was a great success, that the group might do?

They use their data and infrastructure advantages for smart reach and there's no degree of, what's going to make this smart embodied in consumer products as much as possible, such that they can continue to obtain first-hand data, these data, in turn, to improve intelligence level. This is a positive feedback, if only one company on the planet, that is likely to be the company. For as long as the above positive feedback continues, the business almost across all areas of the company. And because the company's business is Super Intelligent drive, often can obtain very high profit margins, it can do so many things with investments and acquisitions.

From this perspective to understand Google, that many of its actions will become more logical. Letter of many companies and the products can find their place. Like it hot air balloons in the sky, that is to get more people online, good access to more devices; $ 400 million acquisition of DeepMind because I want to buy talented team to enhance brain power; $ 3.2 billion acquisition of Nest because the family needed a center to understand various data; unmanned because only a few typical high-frequency terminals used by mobile phone is outside the car ; Pharmaceutical companies can be seen as the Super brains to this an attempt to land the lucrative industry.

March 2016 broke news of the Google Alphabet of the parent company to sell Boston Dynamics, networks for multiple reasons such as: Boston Dynamics merged with other team members do not wish to receive the company's arrangements, public relations team think humanoid robots is a bit scary. More rational analysis argues that the main reason is the Alphabet that humanoid robots it's a little early, at least until 10 business. Many people think just after under dash by virtue of AlphaGo won the Li Shishi DeepMind chances are also dangerous, because this company has no output. But if you understand the ideas, you will find, more likely to be on the Alphabet will be adjusted according to the end of the maturity to end business layout, such as Nest, robots, medical items, and even a driverless car, but is unlikely to do that part of the hands and feet of the brain. Because the parts of the brain are roots, Foundation is not solid the whole building would shake.

Minority game

After understanding the Alphabet story, we sadly realize that intelligence-driven product few people play the game, and once the software development and the Internet are completely different.

Up to now, China's IT industry after three transformations: once for the popularization of computers, software, and once for PC the rise of the Internet; the last one was because the rise of mobile Internet. Each protagonist is different, appearances are not the same.

First generation software heroes are mostly solo, that today, most people have now ceded the central location of the stage. Qiu bojun, can be said to be representative of the programmers of the time, he was almost a completed WPS under DOS.

Mr Qiu bojun programming level is extremely high, this m CEO Lei on the memories of it from childhood:

In early 1990, I had a friend that WPS Hancock, was shocked at that time. Interface to use beautiful, stronger is the simulation print results can be displayed. Signature is entertained in Hong Kong Qiu bojun, think of the "Hong Kong" software was really well written.

At that time developed a widely recognized products will be able to start a career, at that time, too many people are repeating the story, Qiu bojun, WPS, Wang Jiangmin KV300, Bao Yueqiao UCDOS and so on.

Internet mobile Internet to has Hou story slightly has different, this work began Yu portal, but end Yu BAT, certainly this industry brings of opportunities in the has a huge of long tail, so many venture type of small project as can get good of success, like weather, and years calendar, and novel, and sent meal, and reviews, and taxi, and beauty methyl, and map, and domestic, and sent fruit and so on.

In the past two in the wave, developed a product and technical threshold, but this entry is not too high, especially in the various platforms and tools are mature, so there are so many teams have developed a much more successful App. So basically you can view software era, the age of the Internet and mobile Internet are creative, mode of operation, these things are driven rather than technology-driven. Does this product also will test technical skills, such as how to handle multiuser concurrency issues, but the threshold is not high to the daunting, at least in the number of users by more or less has a chance to grow. Also for this reason, it requires less capital, not too much money you can start an App Development, as long as the number of users that can access continuous financing. Can tell that this is basically belongs to the wave of mass.

But intelligent software in trouble a lot, I really want to own up to algorithms, data, GPU. Even Google open source it, but who wants to carry these frameworks also need to understand the machine learning, otherwise you are not optimized for vertical areas, while data and GPU will be a waste of money. A direct consequence is that many traditional programmers get the hang of this business, which would have set a very high threshold, but more annoying is the smart itself cannot directly create added value, to find a floor. Landing on a purely artificial intelligence software, even Google Photos cannot be regarded as very successful and this means smart floor often require integration of hardware and software. It is not difficult to understand, simply do a chat program such as ice is not integrated into the system of the Siri or Google Now user value, while the former user will have certain advantages.

When integrated into intelligent hardware, things become complicated, someone needs to make operating system, industrial design, may also need to get the acoustics, also need to get the supply chain. Even if all these are done, also may not be able to sell.

Robot running around if you really want to do that, it would be even more trouble, because you need to clear perceptions of the environment, and analysis, also with mechanical parts, joints can be a very good speed and can be used many times.

Smart product base step by step threshold so high, every step of the elevation, it filters out some of the people. We can't deny the genius of little a person can fix a lot of things, but this does not change the smart times only for the few opportunities a reality. Domestic growth better company over the years, mainly by not the technology but the timing sure, operations the pursuit of strong, product details, network effects, and so on, but these estimates adapt to the new rules of the game. It's like the Qing dynasty scholar was good at poetry, when you ask him to do business, so he learned to use the Internet thinking, that he estimates are uncertain.

So smart as a wave of sensation, but it is bound to be a few games. There are few people venture to develop an App product, and there are few people venture to develop a smart car!

In a highly sophisticated artificial intelligence, after the reconstruction of production, individual productivity, a car is possible, but that is a long time from now.

Summary

Things are becoming more and more interesting, because artificial intelligence destroying visible at various positions on the one hand, on the one hand this is a minority game. Meanwhile, if technological development makes more and more people into poverty, it is even more funny. What will our society where?

占领了字母表的Google到底在干什么？ - Google,人工智能 - IT资讯

Google可能是这个世界上最神奇的公司，Facebook虽然也天马星空的做事情，但论程度其疯狂程度其实远逊于Google。Google聚集了这世界上最聪明的人，单纯的说这些人是神经病肯定是不合适的，所以必然有种内在脉络支撑着Google今天做这件事情，明天做那件事情。

Google在干什么？

地球上最为激进的科技公司Google近来对自己进行了重组，总公司的名字现在改成了Alphabet(字母表)，字母表上的各种公司五花八门，不用心思考完全不理解这公司在干什么，唯有回到人工智能可能对未来产生的巨大影响上才能更好的理解字母表公司的初衷。

现在字母表上的公司和产品如下：

A–Android / AdSense / Analytics / Ara / AdMob / Alerts

B–Blogger / Boston Dynamics / Books

C–Calico / Cardboard / Capital / Contact lenses

D–Drive / DeepMind / Design / DoubleClick

E–Earth / Express

F–Fiber / Fi / Flights / FeedBurner / Firebase / Finance

G–Google (duh) / Gmail / Glass / Groups

H–Hangouts

I–Images / Ingress / Inbox / Invite Media

J–Jump

K–Keep

L–Local / Loon

M–Maps / My Business / Makani

N–Nexus / News / Now / Nest

O–Offers

P–Plus / Play / Photos / Picasa / Pixate / Patents

Q–[Nexus] Q

R–Refine / reCaptcha

S–Search / Self-driving car / Shopping / SageTV / Stackdriver / Skybox / Skia / Scholar

T–Translate / Tango

U–URL shortener

V–Voice / Ventures / VirusTotal / Video

W–Wear / Wallet / Web Toolkit / Wing

X–Google X labs

Y–YouTube

Z–Project Zero / Zagat

这里面除了一般很容易理解的各种软件产品项目比如Android、Gmail、Youtube、Blogger、Search等之外，特别让人意外的有Calico(医疗)、Nest（智能家居）、Self-driving car（自动驾驶）、Boston Dynamics（机器人），收购这些公司上基本都投入非常巨大，动辄数亿美元，但正常来讲不太能看到它们的相关性，机器人和医疗确实也八竿子打不着，好多公司如波士顿动力这种更是短期内不太能看到商品化的可能。Google这种公司聚集了地球上几乎是最最聪明的一波人，个别的公司也许是因为一时冲动，但整体上毫无条理的做事情可能性几乎没有，那它到底是想干什么？

这要切换到人工智能的视角。我们来想象下，假如我们真的打造了一个超级大脑，这个超级大脑在应用上会有差别，可能体现为AlphaGo与人类下围棋，也可能体现为自动驾驶汽车等，但在算法、数据、基础设施上其实依赖于很相似的东西，并且这种超级大脑的智能程度呈现出显然的数据驱动特征---数据越多，智能程度越厉害。这个时候再假设这公司的领导者们是一群野心勃勃的人，过去也曾经从这种野心中获得了巨大的成功，那这群人可能会做什么？

他们首先会利用自己的数据和基础设施优势让智能达到别人无法达到的程度，接下来会让这种智能体现在尽可能多的用户产品中，一旦如此，那他们就可以源源不断的获得第一手的数据，这些数据反过来再来提升智能的程度。这是一个正反馈，如果地球上最后只还有一家公司，那很可能是这家公司。因为只要上述正反馈持续下去，那这各公司的业务几乎可以遍及各个领域。又因为这个公司的业务是超级智能驱动，往往就可以获得极高的边际利润，那它确实可以用投资和收购做非常多的事情。

从这个视角来理解Google，那它的很多行为就变的更合乎逻辑一些。字母表中的很多公司和产品也就大多能找到自己的位置。比如它放热气球上天，那是让更多人上网，好能接入更多的终端设备；4亿美元收购DeepMind那是因为想购入天才团队，增强大脑的力量；32亿美元收购Nest那是因为家庭中需要一种中枢来系统了解各种数据；无人驾驶则是因为人类只有这么几个典型高频使用的终端，手机之外就是汽车；医药公司则可以看成是把超级大脑向医疗这个利润丰厚行业进行降落的一种尝试。

2016年3月爆出Google母公司Alphabet要出售波士顿动力的消息，网传这有多重原因比如：波士顿动力的成员不愿意接受公司的安排并入其它团队，公关团队觉得人形机器人有点吓人等。比较理性的分析则认为主要原因是Alphabet认为人形机器人这事有点早，至少还得10年才可能商用化。很多人根据后一点认为刚凭借AlphaGo赢了李世石而大出风头的DeepMind没准也危险了，因为这公司也看不到什么产出。但如果理解上面的思路，那就会发现，更可能的是Alphabet会根据端的成熟期来调整端上的业务布局，比如Nest、机器人、医疗项目、甚至无人驾驶汽车，但不太可能对做大脑那部分动什么手脚。因为大脑那部分是树根，根基不固整个大厦都可能会动摇。

少数人的游戏

理解了Alphabet的故事后，我们会很悲哀的发现智能驱动的产品是少数人玩的游戏，这和互联网乃至曾经的软件开发完全不同。

到现在为止中国IT行业历经三次变革：一次是因为电脑的普及，软件的兴起；一次是因为PC互联网的兴起；最后一次则是因为移动互联网的兴起。每次的主角不一样，登场的方式也不一样。

第一代软件英雄们大多是单打独斗，现在那一代人大多已经让出了舞台的中央位置。求伯君先生可以说是比较能代表那个时代的程序员，他几乎是一个人完成了DOS下的WPS。

求伯君先生的编程水平应该是极度高超的，这从小米CEO雷军先生的回忆上看出来：

1990年初，我在一个朋友那用了WPS汉卡，当时就被震住了。界面易用美观，更强的是打印结果可以先模拟显示出来。署名是香港金山公司求伯君，觉得这个“香港”软件写得真好。

那时候开发出一款获得广泛认同的产品就可以开创一番事业，在那个时代里，太多的人重复了这个故事，求伯君的WPS，王江民的KV300，鲍岳桥的UCDOS等等。

互联网移动互联网来了后故事略有不同，这工作开始于门户，但终结于BAT，当然这个行业带来的机会中有着一个巨大的长尾，所以很多创业型的小项目一样可以获得不错的成功，比如天气、万年历、小说、送餐、点评、打车、美甲、地图、家政、送水果等等。

在过去这两波浪潮里，开发出一款产品时也有技术门槛，但这门槛并不太高，尤其在各种平台和工具成熟之后，所以才有那么多团队能陆续开发出那么多比较成功的App。所以基本上可以认为软件时代、互联网时代和移动互联网是创意、运营、模式这些东西驱动的，而非技术驱动的。做这种产品也会考验技术能力，比如怎么处理多用户并发的问题，但门槛并不会高到让人望而生畏，至少在用户数由少变多的过程中还有一个成长的机会。也正因此，它需要的资本量也没那么大，不用太多的钱就可以启动一个App的开发，只要把用户数做上去那就可以获得后续不断的融资。可以讲这基本上这是属于大众的浪潮。

但智能型的软件就麻烦很多，真想自己弄上来就要算法、数据、GPU。即使是Google等把这开源了，但想驾驭那些框架也需要理解机器学习的人，否则你没法针对垂直领域优化，而数据和GPU则会很耗钱。一个直接的后果是很多传统的程序员搞不定这个行当，这无疑就树立了一个很高的门槛，但更讨厌的事情是智能本身往往不能直接创造太大价值，还要想办法落地。纯粹的把人工智能在软件上进行落地，即使是Google Photos也还不能算是非常成功，这也就意味着智能的落地往往需要整合软硬件。这并不难理解，单纯的做个聊天的程序如小冰就是没有整合到系统中的Siri或者Google Now用户价值大，虽然前者获取用户上会有一定优势。

一旦需要把智能整合到硬件里事情就变复杂了，需要有人搞操作系统，有人搞工业设计，可能还需要人搞声学，最后还需要搞定供应链。即使所有这些都搞定了，还不一定能卖出去。

如果真想做那种四处乱跑的机器人，那就还要更麻烦，因为你要比较清晰的感知环境，并进行分析处理，还要处理机械部分，让关节能够很好的减速并能用很多次。

智能产品的门槛就这样一步步垒高，每抬高一步，就过滤掉一部分人。我们不能否认极少的天才人物一个人就可以搞定很多事情，但是这不改变智能的时代只给少数人机会这一现实。国内过去这些年成长性比较好的公司，主要靠的并非技术而是对时机的把握、运营的强悍、产品细节的追求、网络效应等，但这些公司估计适应不了这种新游戏规则。这就好比清朝的状元本来擅长的是诗词歌赋，突然你让他经商，让他学会用互联网思维思考，那他估计是搞不定的。

所以说虽然智能的浪潮一样轰轰烈烈，但这注定是场少数人的游戏。有几个人创业可以开发一款App产品，又有几个人创业可以开发一款智能汽车！

在人工智能高度成熟，重构了生产方式后，个人生产一辆汽车其实是可能的，但那是很久以后的事情了。

小结

事情在变得越来越有意思，因为一方面人工智能在摧毁可见的各种岗位，一方面这是一场少数人的游戏。与此同时如果科技的发展反倒让越来越多的人陷入贫穷，那就更加的滑稽。我们的社会究竟会走向哪里？