Predicting the future is rather tricky, but that hasn’t stopped whole industries being built around divination: science fiction, astrology and industry analysis, for example. While it is a tad churlish to lump analysts and media in with Asimov and Mystic Meg, I’ve always found the prediction game a little “loose”; sometimes driven more by ego than by insight. One analyst memorably introduced himself to me as “the wisest person in the industry”. Turns out he was the sort of guy who happily predicted 10 of the last two recessions. Even a stopped clock tells the right time twice a day, right?
When it comes to making predictions, our focus on the short term tends to get in the way. As Shaun Collins, CEO of CCS Insight (more on them later) says: “We overestimate what can happen in two years but under predict the change that will happen in ten.” As the story goes, Bill Gates was the man who put a PC on the desk of every home, but he was also confident that no personal computer would ever need more that 640KB of memory. In the predictions game you win some, you lose some and hindsight rules supreme.
The technology industry is particularly fond of predictions, with the analyst giants making billions of dollars from their future gazing. However, innovation and invention are moving at such a pace that it has become increasingly difficult to predict anything with any certainty. Yet, we still trust the analysts get it right more often than not. In the kingdom of the blind…
I was recently invited to attend a predictions event by CCS Insight in London, and can honestly say I was pleasantly surprised by the clear, compelling and human tone of the speakers as well as the simplicity and boldness of many of their analysts’ predictions. One session, in particular, was incredibly useful in helping to understand the drivers behind the Data-Driven Economy – a subject that tends to get talked about without real definition. Analyst Geoff Blaber’s description was so good that I have summarised his presentation here:
Data ex machina
Blaber began by explaining the shift in connectivity in the last few years. First, he said, single devices connected people to other people. Then machines have become connected to one another - the Internet of Things. This last shift has been one of the biggest in terms of data creation, with a single smart factory estimated to create 1 petabyte of data every day. In old money that is 754m floppy disks of data from a single factory in a single day. The Data-Driven Economy is one where sense can be made of this data, and that requires convergence of technologies and collaboration between companies. And that isn’t easy.
The four horsemen of the Data-Driven Economy
Blaber predicts that four different elements of technology innovation will have to come together to harness the real power of data: Artificial Intelligence (AI), Blockchain, Internet of Things (IoT) and Connectivity (5G).
Cue prediction 1: these four technologies will come together and be fully integrated by 2020.
2020, isn’t really that far away given the multiple players involved and the huge amount of collaboration that will be needed to make it happen. Blaber’s vision is that each element plays an integral and interlinked part in the Data-Driven Economy:
- IoT will create the data
- AI will make sense of the data
- Blockchain will track the data and its ownership to build trust
- Connectivity will move the data as seamlessly and intelligently as possible.
It’s all about connections – if those connections work
Connectivity is the focal point of everything, especially everything we are hearing about 5G. For Blaber, this is not connectivity as we know it, but is a focus on how the network is constructed to cope with billions of diverse machines and devices. Take the differences between a single sensor measuring temperature in a field and the myriad of sensors sending and receiving data in a self-driving car. The first requires very low power, where latency is not an issue; the second is hugely power intensive, where latency is the difference between life and death.
The complexity of connectivity is what makes 5G interesting, according to Blaber, but it is also a bit of a concern as the vision and the marketing promises of the technology are starting to diverge.
Prediction 2: there will be huge variations in users’ experiences of 5G in 2019, and 2020 become a marketing headache for the technology
On top of the 5G issue, is a major problem with “the cloud”. If you are generating a million gigabytes of data a day then there is no way all of that can go back and forth to the cloud in any meaningful or timely way. Mobile Edge Computing (MEC), where AI and processing power resides in the device itself, allows you to make sense of the data at source, rather than in the cloud. Autonomous vehicles, for example, must be able to make decisions on board in real time. Cloud connectivity will be crucial to harness data from third parties, such as CCTV, traffic data etc, but you wouldn’t want your car relying on information from the cloud as it navigates the Hemel Hempstead roundabout system.
MEC offers convenience, speed, security and privacy, but it also changes the notion and definition of the cloud as we know it. We will start to see servers placed at base stations by mobile carriers, so computing can be done locally. The cloud will come closer and closer. More of a “fog”.
MEC is a big part of the monetisation plans for 5G at the carrier level, but it also means manufacturers can create thinner and lighter devices with less horsepower because all the work is be done at the network edge.
For this to become a reality, it requires huge collaboration between carriers and cloud companies. As Blaber says: “That will be a big headache.” But it is inevitable. Probably.
Blockchain is the backbone
Machines create the data, AI makes sense of the data and 5G will connect the data. But what about monitoring the data? Transactions work today in linear ways, with numerous third parties involved across complicated supply chains. Data is created and communicated at each point, through multiple systems and multiple records. This creates an overreliance on the strength of each system. It also lacks visibility and standardisation, requiring a huge amount of human intelligence to audit and interpret data. In short, there is huge scope for confusion, errors and fraud when people get involved. You are the weakest link.
Blaber talked about WalMart, who conducted a study to see how long it would take their systems to track a shipment of sliced mangoes back from the point of sale to the original producer. It took seven days. When that week is scaled out across their whole product line, that is a huge problem for security and trust. How do you deal with health and safety issues and product recalls. The cost and reputational damage would be astronomical.
Blockchain promises to solve the problems by providing single shared record for every transaction that everyone can see in real time. When WalMart implemented blockchain they reduced the time to track a single shipment down from one week to 2.2 seconds. It took WalMart 30 days to implement blockchain. That is huge return on investment by anyone’s standards.
The problem, Blaber says (and there is always a problem to caveat any prediction) is that blockchain has reached epidemic proportions of hype. ”The fact it is being touted as the solution to the Irish border problem shows that,” he says.
But there is no doubt that blockchain is hugely important for the Data-Driven Economy as it has countless applications to simplify processes, create trust and cut out the bureaucracy created by human beings and their love of paper. Just think of all the people involved in a house sale: buyers, sellers, lawyers, The Land Registry, estate agents, etc. Blockchain allows contracts to be automated, and those “smart” contracts mean parties can be paid almost instantly and automatically. No invoices, no delays, but, of course, no-one can promise that estate agents still won’t be dicks. Technology can’t solve everything.
Blockchain has so much potential; it has the scope to be the contractual and transactional backbone of the web. And, just as the infrastructure of the web has to scale in parallel to the rise in data, blockchain has to be there to remove the friction and create new business models.
Blockchain is still rather brittle and slow; a game called CryptoKitties brought the whole Ethereum blockchain to a standstill in 2017, for example. Blockchain is also incredibly costly to maintain. Blaber quoted one of blockchain’s founders as saying: “At the moment, blockchain is just a very expensive, inefficient database.”
However blockchain is scaling rapidly and lots of progress is being made. Cue prediction 3:
Prediction 3: All major cloud service providers will deploy blockchain commercially by the end of 2019.
For Blaber, the big driver here is ‘Decentralised Computing’, which he describes as “Edge Computing on steroids”. Rather than data being stored on one company’s infrastructure, with DC it is distributed across many nodes across the web. The big philosophical principle here is currently too much power is concentrated in far too few players. It is better for everyone and better for security and privacy if blockchain is used to decentralise all computing power. A related argument in this debate is we also need stronger laws to curb the power of the internet giants, but it is perhaps more likely that technology will provide the cure as well as being the origin of the problem.
While Blaber doesn’t see a rapid shift to this Utopia any time soon, he does think the Data-Driven Economy will move us to more of a hybrid of traditional cloud and DC. He also says that quantum computing is turbo-charging this move – but that’s a topic I can’t even pretend to understand. Perhaps for another day.
Decentralised Computing, enabled by blockchain also enables completely new business models. The machines that collect data can also sell that data automatically - so long as it is non-sensitive, of course. For example, a sensor on a shipping container that collects weather data can then automatically sell that data to meteorological agencies for their forecasts. Sensors on trains primarily used to monitor the transport of good, can also sell data to the rail network owner to monitor the condition of tracks.
This is well under way, according to Blaber, and he expects to see the first commercial products built on blockchain to be available by the end of 2019. This is potentially hugely disruptive for cloud companies, so they need to embrace blockchain or risk being disrupted themselves. The only way to understand blockchain is to start to use it, and that’s why it will become mainstream very quickly.
Reputation, risk and the balance of power
We live in a world where compliance, regulation and trust dictate the way we do business. As Warren Buffett once said: “It takes 20 years to build a reputation and five minutes to ruin it. If you think about that, you’ll do things differently.” In wake of the Cambridge Analytica scandal, Facebook learnt some tough lessons about its responsibility with regards to data. It needs a clear audit trail of who historically has had access to data and what it is used for. For Blaber, the most shocking thing was that Facebook simply assumed all the data had been deleted. It had not.
Prediction 4: Facebook deploys blockchain technology to track social networking data by the end of 2020
As well as Facebook launching its own payment mechanism based on blockchain it will also develop data tracking and authentication in the coming two years. We can expect the other social media and internet giants to follow suit as they all run in fear of litigation and the next big breach. After all, compliance is ticking a box, but *proof* of compliance is what keeps you in business.
While the US is burying itself in trust issues, other parts of the world are still scaling their technologies rapidly. The next big macro trend that CCS sees is that Silicon Valley’s dominance will wane over the next five years. The region will always be an important contributor to technology innovation, but the Data-Driven Economy requires rapid scale and low barriers to entry. Silicon Valley is simply too congested and too expensive. Real estate prices are out of control and the talent pool is arguable too small and too overpaid. Instead, China is a rising tide, especially in AI.
But no one region will win. The Data-Driven Economy will need a high level of collaboration across the nations and across industries. If, but more likely *when* this happens, Blaber confidently left us with his final prediction:
Prediction 5: The tech advances of the last decade will look like a speed bump when compared to the coming ten years