Posts from September, 2015
No comment yet
September 27th, 2015

离上次预测未来2016已经过去了四年。本打算每两年一次预测,每次预测四年之后的情况。因盖茨曾说过,我们总是高估两年之后的科技发展,而低估十年的科技发展。而十年之前,将奔腾四芯片的运算能力放到手掌大小的运算设备上几乎是不可能的事情。哪怕是八年之前,当我们要订制有当时高端台式机处理能力的一体化电视也是困难重重。

回顾过去四年的预测

虽然四年前的预测现在看来相当精准。但是经济政治稳定的大前提在各国政府的量化宽松政策主导下得以保持,否则这样的预测难言准确。回顾四年前的预测,互联网接入、3D电视悲观市场前景、电视剧点播、计算机性能、辅助驾驶、自拍技术均和现实没有太大出入。然而,对于无线电源的普及、Pad和笔记本融合的趋势、商用超音速客机投产、失业率以及人工智能的发展预测都出现了较大失误,没有预测到无人机的发展。总结起来,一些是过于乐观,而另一些是知识的储备不够。

未来四年经济社会总体概况的预测

对于接下去四年的预测很难以经济社会总体稳定作为前提了。就全球而言,经济增长将会放缓。美国反而会因为向全球输出货币,成为受影响最小的国家之一。四年中,在欧洲,西班牙、希腊等地中海国家的情况很难得到改善。虽然政治情况的发展总比预测的缓慢,一个或多个国家退出欧元区的前景将会变得非常确定。在这样困难重重的国际环境中,日本的情况却在数次有计划的税率提升后有些许的改善。然而,最难以预测的情况是中国。在未来四年,中国可能会出现如下情况中的任意一种:

1). 在货币和经济政策调控下经济成功硬着陆,中国的GDP增长会降到每年4.5%到5.5%。总体上,中国的财政将会更加平衡,作为全球制造工厂,中国所整合的系统生产效率将是其他即使有低成本劳动力的经济体难以企及的。在这种情况下,中国理所应当地成为了人均年GDP产值在九千到一万美金左右的新兴发达国家。

2). 在未来的两年中,中国的GDP增长将降到致命的4.5%或者更低。经济和货币政策在大量的货币对外流失和自2008年以来的资本失控中效果乏乏。社会动乱比想象中的更加容易。地方政府将会疲于应付各种社会暴动。而中央政府可能会和反对党领导人展开对话。而下一步的发展会变得无法预测。

于我而言,选择1作为中国未来四年的背景将是能够做出有意义预测的唯一方法。如果选择2更接近未来四年的现实的话,以下的所有预测将会和现实完全脱钩。

对于印度,我没有任何的系统知识。这也使得我无法预测印度对于全球科技和政治经济的影响。对于俄罗斯和中东产油国,油价将会徘徊在每桶40到100美金。因为油价会在未来四年大幅波动,而俄罗斯的经济也将会受到拖累。

预测的基础

如果想对于任何预测的成功有任何把握的话,我们只有参考过去。在过去的100年中,最重要的事件是人类对于指数级增长的理解。在各种书籍文献中,我们对于指数级增长有着近乎狂热的崇拜。然而,如果只是将指数增长的图表画出来,而不了解其后的科技规律的话,我们也许会遭遇到物理学的根本极限而不自知(但是,过于聪明的理解物理学的极限也会让我们过早放弃对于指数级增长的发掘)。

过去100年人类的指数级增长成为可能只是因为这两个关键字:标准化和大规模生产的效率。现代集这两者为一体的就是iPhone了。如果没有iPhone的生产规模,现代的高分辨率电容触控显示屏将会花费数千美金每平方英寸。但是现在,人人都可以仅花十几美金获得这样的一块屏幕。

标准化和大规模生产,在未来的四年中,将会以各种形式,继续展现其惊人的有效性。

预测

智能硬件的概念提出已经有十多年了。究竟智能硬件应该是什么样的呢?

  • 智能硬件上,它原有的功能应该变得极其傻瓜化。轻松、一次完成、不用动脑的操作;

  • 除去基本的功能以外,智能硬件能够用有限的方法解决一些之前使用中的“痛点”(一个好的例子是可以自动下载云端内容的路由,而能自动预订食物的冰箱就是一个坏案例);

  • 在未来四年,家中的智能硬件不大可能是一种全新的东西。

接下来的四年将是非PC的时代。大多数家庭将不会拥有桌面电脑,虽然在大部分家庭中,所有器件的总计算能力将超过10Tflops。交互式界面也将发生变化。人们将主要通过触控和语音与设备交互。图形化界面将会被对话式界面改造或替代。

虽然有地方冲突和不稳定因素,总体上来说,交通的成本效率更高了。地面交通而言,自动驾驶或者驾驶辅助技术成为新汽车的标配,但离强制标准仍然差很远。曾经Abu Dhabi的PRT虽然不会在中东成功,但类似的交通工具将在一些城市商业化。长距离的交通工具革新在美国仍然没有摆脱实验阶段。不仅如此,一些长距离的不间断航班由于成本上升而被取消。商业交通将会更慢,也更贵(虽然成本效率更高)。

在经济萎缩的时候,娱乐业反而会迎来一次增长。人们仍然会花大部分的时间在电视上,但是选择不订阅任何有线电视的人群将会大大增加。在美国,15到35岁的有线电视观众将以年均10%到20%的速度加速减少。现在收视率最高的电视(首映时五百万左右的观众)在四年之后仍然会保持大致这样的收视率。但是大部分收视率在两百到三百万左右的剧集将会降到一百万以下。在未来四年,某一家在线播放服务商会和一家主要的体育赛事签订在美国独家直播的协议。总体上,从电视和移动终端上,人们每天将会花超过三个小时观看在线视频。

共享经济的未来或许不会是像大家所想的那样。共享经济的核心是将重资产从公司的资产负债表中剔除,从而提高公司的盈利能力。在经济增长期,轻资产使得公司能够快速发展,将盈利能力不佳的项目轻松砍掉。但是在经济萎缩的时候,这些公司将会尝试在资产价格划算的时候收购一些资产。但是最流行的方式或许不是直接的购买。这些公司将会提供给共享经济工作者们各种融资项目让他们能够有资本去购买资产,将资产的贬值风险留给这些共享经济工作者。

移动消息服务将会集中在少数的几家公司手中。这些公司每家将会拥有至少三亿的日活跃用户,每天的消息发送将会超过二十亿条。现在的移动消息服务商中不能达到这些数字的公司将会死掉。全球主要的移动消息服务商只会剩下三到四家或更少。所有的这些移动消息服务商都会提供语音和视频通话服务。这将进一步使得传统电信服务商的语音通信服务变得无关紧要。在美国,至少有一家大型的在线互联网服务商将会进入互联网通信服务市场。互联网接入速度将会进一步提高。全球而言,家庭有线互联网服务接入平均速度将在100Mbps左右。全球移动互联网接入平均速度将在10Mbps左右。在中南非洲,互联网接入速度将会到500Kbps。换句话说,只要有钱支付,在全球除了南极洲以外的任何陆地上,你都能够通过手机进行稳定的视频通话。

医疗仪器也会变得更加便宜。低价的计算能力和机器学习在信号处理中的广泛应用使得必须的医疗设备变得更加便宜耐用。它们能够在地球上最偏远的地方正常工作。这一变化将会对全人类的平均寿命产生深远的影响。

在其他方面,虚拟现实设备将会出现在更多的家庭中。它们仍然在寻找一个杀手级的应用。即使这样,在2019年底,全球将会每年售出三千万台左右的虚拟现实设备。工业机器人将会在更多的工厂取代人类,而这对于中国而言是一个利好。航天技术的私有化仍在继续,私人企业将会成功完成至少一次载人航天任务。

就这样,2020年还没有到人类历史上最坏的时代呢!

It has been 4 years since the last prediction for the year 2016. My original plan is to draft a prediction every 2 years, and scope for the next 4 years. Gates once said, we always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten. A decade ago, having computing devices as small as a palm with Pentium 4 computational power was unimaginable. Even 8 years ago, it was a difficult fate for us to build an all-in-one TV with high-end PC capability.

Review the Prediction of the Past 4 Years

The prediction of the past 4 years has been accurate. The biggest promise of economic stability has been kept with all the unusual fiscal policies, otherwise such predictions can hardly be any believable if at all. Reviewing the prediction I made 4 years ago, Internet connection speed, the unfortunate market share of 3D TV, Television on demand, computational power, driving assistance (self-driving), and photography technology have matched the reality pretty well. However, for wireless power source, Pads and ultrabook merging, commercial supersonic flight, unemployment rate, and artificial intelligence has been off quite a bit. No predictions on unmanned aerial vehicles. Overall, some of these predictions are too optimistic, and some of these are simply ignorant.

The Economic / Social Outlook for the Next 4 Years

However, it is harder to predict the next 4 years on the same social / economic stability promise. Globally, the economy growth slowdown will be a given. On the contrary, the United States will be least affected due to the dominance of Dollar in Global economy. In Europe, it is unlikely the economic situation in Spain, Greece and other Mediterranean countries will get any better. As slow as politics go, the possibility of one or several countries exiting euro-zone becomes ever more real. However, under the gloomy environment, Japan's outlook improved marginally after several scheduled tax hikes. The tricky bits, is China. China would likely to take either two paths:

1). Its GDP will land at around 4.5% to 5.5% yoy growth in the next 4 years. This is after a controlled turbulence landing, with some finesse mix of fiscal / monetary stimulus. Overall, the fiscal sheet is more balanced, and as the world manufacturer, China integrates more efficiency in its system, and it is harder to compete on efficiency front even with much lower labor costs. This is a China as a newly-minted developed nation, seating comfortably among the rest of developed nations with GDP per capita between $9,000 and $10,000.

2). Its GDP will land at 4.5% or even below in the next 2 years and will be considered as fatal. Fiscal and monetary tools seem ineffective due to large amounts of capital outflow, as well as loosen control over capital in general after 2008. The social uprising turns out to be much easier than expected. The regional government would be hard to contain the unrest, and the central government would likely to have several rounds of negotiations with opposition leaders, it becomes impossible to predict what would happen afterwards.

For the sake of making any progress on this prediction, I will pick the China option 1 as the background for the next 4 years. If option 2 turns out to be closer to the reality, it nullifies all the predictions I am going to make below.

India, for the lack of systematic knowledge in that area, it is hard to predict the impact of India to the global technology and economy outlook. For Russian and Middle-East oil-producing countries, the assumption will be that oil per barrel will float around $40 to $100, and Russian's economy will struggle nevertheless due to the more volatility in the oil price.

The Basis of Any Predictions

The success of any prediction, if at all, looks at the past patterns. For the past 100 years or so, it has been the capturing and interpretation of exponential growth. It has been emphasized in enormous books and talks about the fascination of exponential growth. However,by applying exponential growth, without the underlying understanding of technological principles, we risk of hitting some fundamental laws of the physics, and makes no progress at all (and on the other hand, a premature prejudice of "understanding" the fundamental limits of physics, can be fatal too).

The exponential growth is made possible only with two key terms: standardization and the economy of the scale. The modern marvel of this kind, is the iPhone. Without the scale of the iPhone, modern high resolution screen with capacitive touch will cost thousands dollars to manufacture per square inch. But now, everyone gets a modern high resolution touchscreen with a few bucks.

These two key words, will manifest themselves in many forms, and will continue to play wonders in the next 4 years.

The Prediction

The smart hardware has been around for more than 10 years. But what makes sense as a "smart hardware"?

  • It makes the basic functionalities we assumed about that hardware a no-brainer. Smooth, one touch, perfect and care-free integration;

  • It extends beyond the basic functionalities, but operates under well-defined principles (good example, a router that caches cloud content and make the access instantaneous, bad example, a refrigerator that orders food for you);

  • It is unlikely to be something completely new.

Then, there is the un-PC era. In the next 4 years, homes rarely own any desktop computers, even though aggregated processing power in a single-family house can easily reach more than 10Tflops. There is a change of the interface too. People now interact with these devices by either touching or talk. The graphical interfaces now have a meaningful conversational re-touch.

Despite the potential conflicts and regional instability, the transportation will be more cost effective. In terms of the land transportation, self-driving or smarter driving assistant will be standard add-on in newly shipped vehicles. However, it is far from becoming the mandatory standard. The Abu Dhabi PRT was a failure in the Middle-East, but similar transportation services will run commercially in some cities. The next generation of long distance land-transportation is still in experimental phase in the United States. Not only that, some of the longest commercial flights are cancelled due to the cost. Commercial transportation is going to be more expensive, and slower.

Entertainment industry gets a big boost in time of recession. People still spend disproportionate time on big television, The movement of "cutting-the-cord" will happen much faster than expected. The United States 15 to 35 year viewership on cable will drop at the rate of 10% to 20% year over year and accelerating. Today's top TV show numbers (5m viewer at the premiere) will keep steady. But shows with 2m to 3m premiere viewership will see a drop to 1m or less. In the United States, online streaming players will ink deals with major sports and have exclusive rights to stream online. People will spend more than 3 hours a day on streaming services, either on television or on their mobile devices.

Shared economy is not going the way you would expect. At its core, shared economy moves the assets out of the company such as AirBnb or Uber's balance sheet and bumped up its profitability. At boom times, asset-light companies can move fast and quickly get rid of less profitable businesses painlessly. At down times, these companies will try to own more assets as the asset prices are all cheap. However, the most popular way for them to do so will not be out-right purchase. Instead, they will launch finance programs to help its share economy workers to own these assets, and leave the risk of asset depreciation to them.

The mobile messaging service will consolidate. Respectable players on messaging service will reach 300m daily active users, and have at least 2b message sent per day. Any player cannot reach that hallmark will be dead. There will be only 3 to 4 major players in that space, if not less. All the messaging services will have the ability to make audio and video calls, which will continue to marginalize the phone call service business for traditional phone service providers. In the United States at least, more than one online-based business will enter ISP business. The speed of the Internet will continue to improve. Home Internet speed globally will average to 100Mbps. Global mobile Internet speed will average to 10Mbps. Specifically, the mobile Internet service in Middle / South Africa will reach average 500Kbps. In the other word, as long as you can pay, with your cellphone, you can have semi-stable Internet connection and will be able to do video calls anywhere in the world except Antarctic.

Cost-effectiveness is penetrating medical equipments. With lower cost of processing power and general application of machine learning techniques in signal processing, popular and essential medical equipments will reach a point that are cheap and versatile enough to even be delivered to the most remote area on Earth. The profound impact will be a global lift in life expectancy.

Virtual reality gears will have tractions in many more homes. They are still struggling to find its killer applications. But on average, shipped units per year will be around 30m globally at the end of 2019. Industrial robots will replace more human labor, which is a good thing for China. Privatization of space technology continues. One or more private companies will accomplish at least one low-orbit manned mission.

Thus, it is not the worst time of humanity yet for the year of 2020.