I just started an MIT course — Artificial Intelligence:
Implications for Business Strategy & 6 more weeks to go. Hoping, an old dog can learn new tricks. Before that, let me share a few simplistic reasons, that this time can be different. Which may make Machine Learning tick better.
The planets are lining up…
Recently, Google co-founder Sergey Brin opened Alphabet’s 2017 Founders Letter with a famous quote: “It was the best of times, it was the worst of times”. He outlined exponential advances in Information Technologies are revolutionary. He knows better. Let’s be open-minded as Google’s AI-First strategy is not AI-Only. Naturally, it’s easy to realize we are at the brink of a paradigm shift. It’s starting with computing and will end up changing every life fundamentally sooner than we think.
What’s new this time?
Big Data, like really, really big. At this iteration, ML is powered by data, aka data-driven. Without a huge amount of “right” data, ML won’t be able to much. Human programming is still a better alternative, and the growth is limited by diligent & bright people. So, if you are not yet “Big Data Ready”, go back to do the 101 first.
Commercialization and migration to Cloud Computing models is an essential step too. By leveraging the cloud computing platform as a service, you focus on your problems, instead of building private IT capability & operation. It also provides a more flexible way to scale on demand. Unless your business can wait years & spend billions to build a data center. Oh, one data center has physical location & connection limitations too. Most users prefer something that just works in The World, which is Flat, I suppose. Furthermore, the data center has become very specialized. There will only be a few giants that can compete and sustain effectively now & forward because of economies of scale. Take IC manufacturing as an analog: TSCM, making IC for others as its main business(32B) is bigger than SK Hynix, No3(26B) IC supplier in 2017 sales. Fundamental, we have to rethink data ownership & embrace Open Data one way or the other. ML can be “smarter” when connecting various data sets. Of cause, privacy is a big challenge, or opportunity if you like.
With the mass distribution of smartphones, 4.4B users are consuming digital content & participating in e-commerce today. It also uploads user-generated content & signals to an unprecedented level. “Mobile-First” may not be hot anymore. It will continue to be the computing device for major users. With 3G/4G mobile connection, mobile pushes computing to anywhere/anytime for users & triggers data explosion for Internet service. What is data used for? It is only limited by your imagination. May the force(ML) be with you. The progress in mobile may accelerate IT development in other industries when diversification is done right.
Even the market is less excited about IoT now, it is a key building block of datafication(taking inputs from) and extension of control to the real world. For consumers, it enables further automation and personalization. For industry & commerce value chains, it is the scalable measure to automatically harvest data, complete tracking, and impose control. With 5G mobile technology, let’s get every moving thing online.
Blockchain is obvious hype when changing the name to Long Blockchain can push the share price to 289% higher. Nevertheless, blockchain brings decentralized consensus bookkeeping to the digital world. It’s possible to build value exchange, record integrity tracking &, etc. network applications. Even Blockchain technologies still have many limitations, and cryptocurrency bubbles will burst soon or later. But some of the investment may improve the technologies for the next big things. Ethereum is especially interesting. It incorporates business logic/code into a block to make a digital Smart Contract. It could be used to build trust & distribute computing systems. In short, it may make IoT more secure & smart, and authenticate data points.
VR/AR augments the Human-Computer interface. Increasing commercial adaption further blurs the line between virtual & real worlds. It’s driven by the gaming & entertainment sector. Which is much better than driven by the military. Game engines & simulation technologies can also be used for ML training & validation. In a virtual world, it’s a function of computing powers instead of physical resources.
Emergence is another promising direction in AI. Conway’s Game of Life in 70th shows complex behaviors can be evolved & emerged from simple rules. Crazy talk right? Check out Intel’s Drones at The Olympic Games & think again. Now, if stupid things look smart, behave like smart, and feel like smart, then they are probably smart according to the Duck test.
What’s next? You decide!
I don’t know about you, but I’m very excited, and maybe too optimistic about such a once-in-a-lifetime opportunity. Yes, AI will replace jobs and it has become very efficient doing so. But for me, I don’t mind machines taking over boring works. If so, it free people to do something else better, or just enjoy life.
Furthermore, our predecessors made through the first 3 industrial revolutions fine. There is no reason the 4th can’t work for us. At least, we should die trying. How about 4 Day Work Week. Be creative while you can ;p
My dear Taiwan, there is nothing wrong with an individual pursuing a small bliss in life for whoever chose so. But please don’t let it become the only choice for the next generation because they can not afford to miss this train.
Full Disclosure
The opinions stated here are my own, not those of my company. They are mostly extrapolations from public information. I don’t have insider knowledge of those companies, nor a whatever expert.
