A very large number of articles on technology at the moment begin by saying, “[x] technology is one of the industry's hottest topics right now” or “changes in [y] technology are accelerating ever faster”. This is as a result of living through a ...
A very large number of articles on technology at the moment begin by saying, “[x] technology is one of the industry’s hottest topics right now” or “changes in [y] technology are accelerating ever faster”. This is as a result of living through a ‘technology purple patch’ where there is a step change not only in what the technology can do, but also crucially the skills that are required to do it. You are not alone if you have an uncomfortable feeling that there is a ship leaving port and you’re not on it!
Minimal data artificial intelligence (AI)
As an example, take Artificial Intelligence (AI). As I have discussed before, the topic is very hot right now, and required Data Science skills are harder to come by, than say, vanilla coding – so far so normal. However, just last week the Deep Mind team at Google announced that their already supreme AI-based AlphaGo algorithm had been surpassed by their AphaGo Zero algorithm, using a more sophisticated reinforcement learning approach.
The key point here being that the new algorithm required only the rules of the game to win. No historical moves, no training set, no data lake. Now I appreciate that this is in deep research, but could it be that for some use cases a large data engineering project beforehand could no longer be required? Of added interest is that the paper detailing the results was submitted to the magazine Nature on April 7, i.e. 6 months ago. This seems like an eternity when the landscape is moving so fast - just imagine what they might have up their sleeves right now!
For me, this situation has gotten even more acute in another of my areas of interest; Blockchain and other Distributed Ledger Technologies (DLT). There are many similarities between how the two areas are applied in financial services. Both are technologies looking for a solution, both have the potential to literally transform the industry, and both have similar resource issues; just substitute Cryptography for Data Science.
The blink-and-you-missed-it initial coin offering (ICO) bubble
When the elusive Satoshi Nakamoto created Bitcoin ‘he’ naturally kept some for himself - some estimates put this at about 1 million BTC, which would be roughly $5.6 billion at today’s rate. A start-up today can raise funds by creating their own currency or token and issuing it in a pseudo crowd sourcing model, crucially without giving up any ownership of their business at all. (Un)fortunately, depending on your point of view, the ICO bubble now seems to have burst with the existence of a profusion of tokens with very little to differentiate them and the number of ICOs raising more than a $1million flat lining.
However, the business drivers, especially from the venture capital (VC) world, could be compelling as investing in tokens rather than ownership turns the most illiquid asset (i.e. an investment tied up for years) into one of the more liquid ones (i.e. an investment that could be traded out tomorrow).
It is ok if both these examples leave you feeling as though you have been left behind. The bright side is that the application of these technologies will ultimately still require large numbers of people to do all of the required integration work, distilling down into three main areas:
1) Bringing the technology to the masses. The complexity of these technologies is crying out for layers of abstraction that allows non-specialists to participate. Witness the public AI application programming interfaces (APIs) available from Google and Microsoft, or the recent announcements about Gluon from Amazon and Microsoft. In the DLT space, expect to see new platforms with simpler ways of writing smart contracts.
2) Change management. Assuming that the firm is still viable in markets that are disrupted by these new technologies, organisational change is still one the hardest, and least appreciated, challenges to accomplish. However, even for new market entrants, laser-like focus on achieving the quantifiable benefits of these technologies is key.
3) System integration. Implanting this disruptive technology into existing IT stacks requires an understanding of the data flows and how the architectures will need to change over time. How the technology change interplays with the need for business agility is, again, key.
In some ways, the current situation brings to mind the stories of the Californian gold rush, whereby most real money was actually made by the entrepreneurial store owners selling pick axes and work boots, rather than the miners who went out and found the gold. They came out on top because they understood the timing of what was happening, their core skill set and their relevance to the new world dynamic.
Fast forward to today, and the real winners who are set to capitalise on these exciting new technologies are those who can take the wider view, and apply their core skills in order to move with the evolving landscape and capitalise on the opportunities that the new technologies will surely bring.