Buy Artwork
Krister Olsson, 3-Part Pscyhobuilding, 2015
6 May 2019
Interview • Issue One • Chapter Four
Space and Time:
Learnings from China

I write this short introduction from a quaint, yet modern boutique hotel in a pristine UNESCO World Heritage Site. Nestled in the hills of China’s southerly Yunnan Province, its name is Lijiang. In the labyrinthine, cobblestoned old city, vendors hawking yak meat jerky, 400-year-old pu-er tea leaves, and traditional handmade paper notebooks engraved with the indigenous Dongba hieroglyphs [1] all look at you oddly if you try to pay with cash. They prefer WeChat Pay, the P.R.C.’s now nation-wide digital payment protocol [2].

The exhilarating modernisation that has occurred in China in the past four decades since it opened its great walls to the western economy has astounded the world. Even now, despite its struggles with the polluting effects of rampant industrialisation, the nation leads the charge towards solar and alternative energy sources. It is miles ahead of its neighbours and even the US in technological development, especially in the fields of AI, robotics, driverless cars and identity recognition.

Late last year, we had the opportunity to interview the National Technology Officer for Microsoft China, Qing Wei on his consultation with smart cities and AI. As Qing approached our table in the crowded cafe, he was taller than most, composed and elegant in a kind of archaic way in his traditional linen Tang shirt. He looked in every way a sage, as I imagined the literati of old to be as they meditated on the balance and poise of the scholar’s rocks, comparing them to the fundaments of the dynasty, or even the civilisation. His words and reasoning were a combination of pragmatism and poetic anecdotes, as he gestured towards the way forward for his homeland and also, the rest of the world.

Krister Olsson, Motivated Sundial, 2016
The key is to tell me how to go from today, step-by-step, into the future.

Adeline Setiawan: In your experience with Microsoft, how do you start to advise your clients — governments — how to implement smart cities?

Qing Wei: Well, if you really look deeply, I don’t think there are any real smart city projects happening in the world yet. Let me share with you why. It’s also our debate with our customers. Many of our customers, when they approach us, they will always come and say, “Microsoft, help us to build a smart city project. Start from your high-level planning to real implementation.” A few years ago, I would always take the job and try to do it, but none of the projects were successful. We did achieve something, but it had nothing to do with the smart city at all. Now, I’ve learned from it. When any customers or governments come back to us with the same request, I ask them this, “Can you share with me any example where you have done a smart project successfully?” And then, they say, “Oh... No.” Then I’ll say, “Can you show me if you’ve ever had a smart building?” Most likely, they’ll say, “Uh... No.” Then I say, “Can you show me, maybe, a smart floor?” And they say, “Uh... No.” And I’ll say, “Can you show me a smart office?” They say, “Nope.” Okay, that’s it — that’s the problem.

Maybe I’ll use a Chinese analogy. We have been trained to say, “The future is communism,” right? But no one tells us how to reach that future because we’re still at whichever stage of capitalism, or socialism, or secularism. The key is to tell me how to go from today, step-by-step, into the future. I think that’s an almost fundamental problem nowadays, with all the buzzwords, be it “AI”, “smart city”, “smart nation”.

So for us, we have learned from very hard lessons. And now, when we do projects, we always begin with what we call the Digital Maturity Model Analysis to evaluate the customer’s digital maturity. Because, let’s face it, “smart nation” or “smart city” is based on the assumption that you have some algorithms to help you to manage the city or the nation. And based on the current level of technology, we know the algorithms are dependent on data. Data is coming from the digitisation of everything. You need to digitise your work, digitise your process, your identity, your building, home, and country. Then, you can have data. So, if you don’t even have a digital maturity level to a certain extent, how can you even convince yourself you have data? And if you don’t have data, how am I going to help you to come up with algorithms to have a smart city or a smart nation? So that’s the whole concept when we do a project for anything smart, like smart transportation, smart classrooms, smart education, smart cities, smart nations. We always start from digital maturity modelling by analysing the customer’s existing processes, and how many of them are able to generate data in a manageable, consistent, secure way, continuously. If they are not able to achieve this level of digital maturity, then we will say, “Can we start by digitising your workflow, your management flow or governing flow first?” I think that takes a fundamental change of mindset.

You need to digitise your work, digitise your process, your identity, your building, home, and country. Then, you can have data.

Setiawan: In your experience, which are the cities or nations that are nearest to digital maturity?

Qing: Maybe I’m too exposed in the field, but really, almost zero... Actually, I think Singapore should be considered the most developed because it’s small. Generally, countries like Israel and Singapore have a higher chance to reach this kind of a maturity level. And there are a few cities such as San Francisco, Los Angeles, Seattle, Toronto, London and Cambridge that I know are working towards that goal.

Roland Turner: How about anywhere in China?

Qing: China is investing in a huge way into smart city infrastructure. But China is huge — it’s too big. Say, even in a single city, we’ve got a population of 10 million. So it’s very hard. That’s why, especially in China, it’s even more critical to understand the digital maturity level first. It’s almost a nightmare to be able to make a 10-million population city become digitally mature.

Setiawan: We heard about how Google is building a smart city — or at least a smart waterfront — from the ground up in Toronto. Do you think there are different kinds of goals in which, whether a company or a city wants to build smartness from the ground up, versus coming in half-way through and saying, “We’re going to help you analyse your existing processes and then try to digitise those”?

Qing: Yes. Ground up from a new city and ground up from remodelling are two different things. But it’s the same level of difficulty. There’s a very good case study that Microsoft did called 88 Acres [3]. Darrell Smith and his team of engineers had a strong belief that they would turn our headquarters into a smart campus. But I think it took him over ten years to make it happen. The reason was, in these buildings, he would have at least two to three different vendors for different controlling units — from Siemens, from Honeywell, GE, Rockwell. Even a simple thing like having these controllers talk to each other — it was a nightmare.

Setiawan: Even with Microsoft’s existing technology, it took ten years! What is it going to take going forward?

Krister Olsson, Vampire Vessel, 2016
If you believe a smart city that is built in five or ten years is too long, what’s actually more important is whether you have the patience to wait for four and a half years without seeing anything. That’s the mentality that you should have.

Qing: Actually, I really believe it has to do with having patience. Ten years is a full journey, but we started to see progress even after a year. It’s hard because you need to do a lot of ground work. For example, I spent the last three years to insert censors in my home, under every window, so I could monitor my home temperature, air pressure, and humidity in real time. And I hacked all the commercially-available fridges and reprogrammed them to be be able to connect to my cloud in Asia, so I could remodel my home. What I learned from it is this: you have to solve a lot of hidden problems.

Say I have a paper cup, I fill it with dry sand, and drip water into it slowly. Only one drop of sand is wet. The rest is all dry, so it makes me frustrated. I continue to drip a lot of water, and the cup is slowly filling, but still, I see no results coming in. But in a single moment, once it reaches the level — zoom. Within seconds, it feels like the cup is full with water immediately. This is the analogy for smart cities, smart buildings, smart offices. It doesn’t mean there’s no progress, it’s just that you cannot see the progress. But once you see it, it’s already done. If you believe a smart city that is built in five or ten years is too long, what’s actually more important is whether you have the patience to wait for four and a half years without seeing anything. That’s the mentality that you should have.

Setiawan: That’s the feeling living in Singapore, it’s as if Smart Nation is creeping up on upon us, but we don’t really know exactly what it is.

Turner: You can’t make any decisions until all the intelligence has been done, and then it’s as if you switch a button, and —

Qing: And everything is happening!

Setiawan: Roughly how long does it take before you start to see the effects of that happening?

Qing: Well, I always tell my customers that they will take very long to finish with smart city projects. I find a lot of projects being killed at the orphan stage because they overpromise without setting the customer’s expectations to deliver phase-by-phase. So I can assure them that I will give them a smart floor in three months to half a year. That eases their anxiety. I also pick newly-built buildings to work with, so I have the advantage of joining the planning stage. When I’ve installed the protocol successfully, they already trust me. Then they’ll say, “Oh, now I can wait another three years for the campus, and five or ten more years for the city.”

One of most critical decisions to be made is how you build the data structure for smart cities at the beginning.

Setiawan: Do you see these customers working with the same type of data collection points or wanting the same type of system on the back-end?

Qing: Well, you’ve already addressed the basic issue that they are not. So they will never be able to build smart cities. There are some fundamental illusions that we need to clarify, especially about unstructured data. The newspapers are always saying, “What’s big data and AI good for?” They claim how AI can extract patterns from an unstructured big data set, which is totally misinformed. Why? Because there are no algorithms that can draw patterns from unstructured data. This has misled a lot of decision-makers. One of most critical decisions to be made is how you build the data structure for smart cities at the beginning.

Setiawan: It’s almost like the foundation of the city.

Qing: Yes. There’s something new happening in the data structure world called Spatial-Temporal Data Structure. In Chinese, when we say “world”, it’s “世界” (shì jiè). “世” means time. “界” means space. “Universe” is called “宇宙” (yǔ zhòu). It’s also time and space. When humans describe the world or the universe, indeed, we are talking about the two dimensions of space and time.

With every other attribute in your database, you need to have a Universal Unique Identifier (UUID) that has both a common timeline and common spatial data. The UUID is a string that links everything together. Your car license plate — that’s a UUID. Your personal ID or passport number — that’s a human UUID. Unstructured databases don’t have this UUID. Without the UUID, you cannot link data to cross-tabulate it and make it meaningful.

Krister Olsson, 3-Part Psychobuilding, 2015
Only with a universal Spatial-Temporal Data Structure and domain-specified UUIDs are you able to analyse the data of a city.

Every government or enterprise, or even for me as an individual when I work on my smart home, needs to decide how to plan their data structure. I, or the company, or the city need to build a Spatial-Temporal Data Structure. Then, you can pull all the strings back taking your education data, traffic data, healthcare data, and they will be tied to the same space and time. Only with a universal Spatial-Temporal Data Structure and domain-specified UUIDs are you able to analyse the data of a city. Then, you can have a smart city. So, to answer your question, I don’t think they can do anything because they don’t have this data structure.

Turner: So the intention to build smart city infrastructure to solve smart nation problems affects what’s built, how it’s built, integrated, labelled, how these sets are restored, how they are restructured. As a counter-example, there was the case of the incidental use of surveillance footage in London. It took MI5 three months — an unbelievable amount of time — because they had to take 40,000 hours of surveillance footage from 100 different systems that were all structured differently and unlabelled, and painstakingly put it all back together to work out what to do it in — a task that in Singapore, would have been possible in much shorter time.

Setiawan: So what you’re both saying is that the first step to a successful smart city is to lay out the infrastructure to collect all the data. The second part is to collect these data according to specific, unique UUIDs that are universal across different systems. And the third part is then to integrate, then the fourth part is to analyse and structure that data.

Qing: Precisely.

Setiawan: I think there’s this pervasive suspicion about smart cities, how artificial intelligence is going to govern our lives, a sense like, “Will the machine or the algorithm take care of me?”

I believe that today, the Chinese government has learned this same lesson when it comes to AI. They say, “Let’s put it aside until we are able to master it, and then we will decide how to use it.”

Qing: Well, I have to give some credit to the Chinese government. There was an incident that changed my beliefs about the system. I used to live in Chengdu, in the southwestern part of China. During the weekends, I would drive to the mountains. I liked to visit the old temples, the ruins. In front of one of the temples, there was an inscription on the gate. It was engraved with the emperor’s command, saying, “We know there is black gold in the core of the mountain. But as emperor, I ask you not to dig into its core — its energy. Because this fengshui [4] line proceeds from the Himalayas, from Mount Everest, all the way to Sichuan, and it will determine the future of our nation.” Even a few hundred years ago, the emperor already knew about the usage of petroleum and crude oil, and made a deliberate decision not to use it. I believe that today, the Chinese government has learned this same lesson when it comes to AI. They say, “Let’s put it aside until we are able to master it, and then we will decide how to use it.”

Turner: It’s sort of like The Sorcerer’s Apprentice in the western fable. You don’t much around with the stuff that you don’t understand.

Qing: Another good reference is Elon Musk. I think that he sees the same thing happening, and this is why he is such a strong advocate of OpenAI [5], and I share the same belief. If something important is going to be invented by humans, we must have absolute confidence that it’s not going to be misused by someone. So Elon Musk’s solution is, “If I cannot stop it, I’d rather make it such that everyone can own it.” If such a powerful weapon could be owned by everyone, then no one is going to utilise it against anyone, because everyone would have the capability to fight back using the same power. This way, this power would be neutralised. It’s the same way that we deal with nuclear power today. Now we have about six, seven nations that have nuclear capability, right?

Turner: Oh, now it’s almost fourteen [6].

Qing: Fourteen, wow. If it was only one, we would have a problem. Two, we would also have a problem. More, we won’t have a problem. OpenAI is the same. Elon Musk has seen the things that are coming, and he doesn’t trust that people can stop it, so it’s better that we have everyone own the technology. In Microsoft, we are calling this the democratisation of AI. Everyone can own it.

I’ve started to feel the same way as Elon Musk, Jack Ma, Stephen Hawking, Bill Gates. More and more of these people have started to see that something’s coming. But they have realised that many people don’t know what exactly is coming. And they are taking the attitude, that they either fully embrace it without questioning, or they are fully against it without thinking. Both ways are wrong.

Krister Olsson, Mirrored Form with Appendage / Cigarette, 2013
In Microsoft, we are calling this the democratisation of AI. Everyone can own it.

I’ve started to feel the same way as Elon Musk, Jack Ma, Stephen Hawking, Bill Gates. More and more of these people have started to see that something’s coming. But they have realised that many people don’t
know what exactly is coming.

Setiawan: You’ve also spoken about a philosopher that will emerge in the next century or so, who will tell us how to think about AI. There is this urgent need to develop a discourse about the ethics of AI. Democratising it is definitely one aspect. How else is China advancing in the arena of ethics, when it comes to data collection?

Qing:
Well, so far the most advanced is probably the General Data Protection Regulation (GDPR) in Europe. We are working with the policy-makers in China as well, to come up with a Chinese version of the GDPR.

Turner: Which is tricky in China, isn’t it?

Qing: Yes, but in my view, I think people are learning. What they’re saying, maybe, is that the Chinese government is utilising data collection to manage the country in a broader scope than the GDPR allows. I think China is trying to follow the general practice worldwide. Because privacy is money, to a certain extent. There’s BAT: Baidu, Alibaba and Tencent. These three are making a big fortune on privacy. But consumers just freely give it to them.

Turner: Same with Facebook ads. I suppose it’s the same situation.

Qing: Everyone needs to learn because previously, they probably ignored something, and they don’t know that this thing might come back to haunt them. Like The Sorcerer’s Apprentice — it’s the same thing.

Qing Wei is currently the National Technology Officer in Microsoft China which allows him to evangelise Microsoft's latest technology vision and innovation. Qing Wei is an ICT industry veteran who has been working in mobile communication, information technology and smart devices related business for over 25 years in Asia with Microsoft and Motorola. He is also an active public speaker and coach specialized in digital transformation.

Roland Turner is Chief Privacy Officer for TrustSphere where he is responsible for the company's information policy and practices. He is a HackerspaceSG founding member and FOSSASIA organiser, holds a Computer Science degree from UTS, and is an avid dancer, runner, and ham radio operator with a particular interest in space.

Footnotes

1 Dongba symbols have been used by the Naxi people, an ethnic group native to the foothills of the Himalayas in the Yunnan and Sichuan Provinces, as early as the 10th Century.

2 WeChat Pay was an offspring of the popular messaging app WeChat. Its partnership with banks allows users to pay for all sorts of services by simply scanning QR codes with their smart phones.

3 Read more about how Microsoft "quietly built the city of the future" here: https://www.microsoft.com/en-us/stories/88acres/

4 Fengshui is a Chinese study of energies and harmonies in the relationship of individuals to and within the environment.

5 OpenAI is a non-profit artificial intelligence research organisation co-founded by Elon Musk and Sam Altman that aims to develop friendly AI that is beneficial to humanity in the long-run. Read more: https://openai.com/about/

6 Actually, 8 nations have officially declared having nuclear weapons: the United States, Russia, the UK, France, China, India, Pakistan and North Korea. However, nuclear power plants operate in 31 countries. 

PREVIOUS READ
NEXT READ

Collect the art
Krister Olsson
Krister Olsson
3-Part Psychobuilding, 2015
Wood, brass, plaster, clay
View This Artwork


Krister Olsson
Vampire Vessel, 2016
Wood, melamine, drywall, polyurethane foam, plaster, motor
View This Artwork

Krister Olsson
Motivated Sundial, 2016
Wood, melamine, styrofoam, motor
View This Artwork

Krister Olsson
Mirrored Form with Appendage / Cigarette, 2013
Wood
View This Artwork