As with many presentations on the future of technology, there were fun examples and predictions related to how technology is changing things.
- There is a near fully-automated hotel in Japan where there is a robot dinosaur at reception.
- There is a speculation that 8 or 9 year olds alive now will never have a driver’s license as driverless cars become standard.
- There are a handful of companies who are adding algorithms to their Board of Directors.
- I learned a new term: Brontobytes. Spoiler alert: it is a really, really, really, large amount of data.
If they were still making Flintstones episodes, there could be some good jokes here as Fred, Wilma, and Bam Bam pull up to the Bronto Burger sitting in the back seat of a car peddled by a saber tooth tiger with a small pterodactyl on the roof giving directions. The next day, Fred can visit the unemployment office because his Slate Rock job was eliminated by 3-D printers that print rocks rather than having to dig them out of a quarry.
The Algorithm Made Me Do It
Out of the many technologies mentioned, the trend around using learning algorithms to make decisions seemed to resonate the most for the audience. A specific example relevant to Houston highlighted an oil and gas company that developed an algorithm with the intent of using it to improve the performance of their poorest performing drilling teams. It turned out that the algorithm began outperforming the best teams. The logical conclusion was to have all the drilling teams “report to” the algorithm. It may be that Pete Townshend was wrong in “Won’t Get Fooled Again”: Meet the new boss, NOT the same as the old boss.
There is an obvious concern that algorithms will replace much human activity. This will certainly be true in some areas. However, I think the bigger idea is that learning algorithms will actually be used to solve problems that no human can solve.
Most automation up to now begins from a premise that a human being can perform the steps in a process, and requirements are gathered to replicate these steps in a faster, more accurate, more consistent way. Now and going forward, the roles can reverse. Human beings are doing things and taking actions that generate enormous amounts of data and only an algorithm can learn from the data to solve problems that the humans are experiencing. From this perspective, the algorithms can be viewed as no different than other types of tools that have enhanced human strength, speed, or endurance.
And, if you buy into an argument that artificial intelligence will not solve the truly hard problems around consciousness or even truly understanding what language means (versus just processing language), then an optimistic view is that humans will have plenty to do working with each other in only ways that humans can.
Or, the pessimists among us can welcome our new algorithm overlords.
But... I Still Can’t Measure My Profitability
An irony I experience when attending these “future of IT” presentations is they highlight a huge disconnect between what could be and what is. If you walk up to many companies and ask them how an algorithm could help their Board of Directors, you could hear things like:
- It might be great… if I could get useful information into and out of my CRM.
- It might be great… if my ERP master data was not such a mess.
- It might be great… if we could accurately track our margins.
- And finally, "Here are 200 Excel spreadsheets, have an algorithm help me with that. Oh sorry, here are 50 more spreadsheets my team gave me after finding out your algorithm was stopping by.”
Struggling with basic problems that have been around a while means the pace of the transition to the new technologies will be highly varied. And, unfortunately, the way the filter will probably work is that companies stuck with the old problems will be at risk of not making it. New companies who are able to start fresh can design themselves to avoid the old problems (and of course, create new ones!). Mark touched on this by asking the group a challenging question: If you knew 20 years ago about the coming digital technologies, how would you have designed your company?
But, for the pessimists fearful of the algorithm overlords, there can be optimism from the struggles with the old problems: A human rebellion can begin by just making a mess of the data. We are good at doing that!