Technology is ubiquitous and has always brought with it benefits and risks. “Sure fire is great – it keeps us warm and scares away bears but did you hear about Og? Burned out of cave and home.”
But the real problem with technology is its limitations. For a lot of people, Arthur C. Clarke’s dictum rings true. Any sufficiently advanced technology is indistinguishable from magic. And magic – like all supernatural things – is infallible. It always works.
This explains people who follow their GPS right off the end of wharves or who take a nap while their ‘self-driving’ car runs into a transport truck. Technology can do wonderful things but in the hands of idiots? Well, Og shouldn’t have poured mastodon grease on the fire.
Technology, to again paraphrase Clarke, is a very powerful, very fast idiot. Machines don’t really think – at least not yet. They rely on programming to do their work which means they rely on programmers. And there is the rub.
Programmers are exceptional at what they do – which is write code. However, their expertise doesn’t necessarily extend to the things they write code about. Anyone who has ever used the grammar function of word processors will know what I’m talking about.
In Ottawa right now, the new IBM developed pay system is failing to deliver pay and benefits to nearly 80,000 people. My wife is one of them. She fortunately is being overpaid and has been for nearly 3 months. Being a rational person, she hasn’t spent the surplus but has stuck it in the bank. She hopes she has put enough aside so that she can pay it back when they finally get things straight. But it is worrisome because, like the programmers who developed Phoenix, she isn’t a compensation specialist. It is notable that to fix the problem they are not only calling on programmers to write new code (at least I trust they are doing that) but are mostly relying on human experts to identify and fix the problems one by one.
That’s what happens when decision makers think they are smarter than experts and buy the bill of goods that proclaims that technology can do anything and do it cheaper and faster.
Part of this problem lies not with programmers but in the nature of expertise. When you are good at something, you generally don’t think through every step in a process. You have internalized best practices and have a hard time explaining it in clear tiny steps. Which is exactly what a programmer requires when they are writing code. Think of it this way: Wayne Gretzky was a great hockey player but when it came to coaching he struggled to impart that greatness to other players.
Most of the problems caused by inadequate technology can be resolved by the application of human expertise and hard work. Eventually the program ‘evolves’ (that is, is changed by human beings) and the initial bugs are resolved – only for new ones to be discovered.
Not a problem when all that is involved is money but I have to wonder – how far should we trust automated medical technology?
And that’s ten minutes.