Smart Tech Needs Smart People!

We see many systems these days where all of the intelligence in the system is embedded in the technology.

Some examples:

  • Driverless vehicles, in mines and on the streets.
  • Highly sophisticated prosumer cameras.
  • Most tablets – they can’t be programmed at all, really.
  • High speed stock trading.

When I talked about this recently, I made this table to outline the issue:

All of these highly tech-enabled examples are in Stage 3 – where all of the intelligence sits in the technology, and none is assumed to reside in the users – the tech is smart so that the people don’t have to be.

This often seems like a logical endpoint, but smart tech/dumb people is an extremely unstable system. For the latest proof, look at the latest news from high-speed trading – the blowup of Knight Capital.

Dominic Basulto has a terrific write-up on this. He says:

Have we given too much power to the machines?

To give you an idea of the scale of the problem: in the trading of a single stock, the algorithm was literally losing 15 cents on every trade, 2,400 times a minute, for 30 minutes straight. Knight, which saw millions of dollars hacked off its stock price in the course of days, was begging for a $440 million rescue package from creditors over the weekend. Rather than admit the scope of the problem – which affected trading in 148 different stocks – Knight refers to this as “a technology issue” – as if it were something that the office IT guy could fix.

Quite simply, a rogue algorithm could take down Wall Street because we no longer know exactly what’s inside all of these marvelous black boxes owned by companies like Goldman Sachs. Trades are executed in the blink of an eye, with computers zipping out of stocks multiple times per minute.

The major financial participants are literally more worried about the speed of their algorithmic computers than they are the intelligence of the humans programming those machines. But isn’t there a hubris in assuming that we are able to reduce the financial markets to a series of blinking 1’s and 0’s, and that whoever has the fastest supercomputer wins?

That’s a pretty much perfect description of smart tech/dumb people system. And they always blow up. Well, they always have. Maybe some day we’ll be good enough at programming that these systems will be stable.

Stage 3 leads to a false economy – systems here are operated in the belief that we can save costs on people by embedding the intelligence into the technology. This might work over the short term, but the costs of the blow-ups more than make up for these savings.

The innovation opportunity here is huge – figure out how to move to Stage 4 – smart tech operated by smart people.

That’s where a lot of modern medicine is these days. And weapon systems. A handful of people with their Nikon D90s have gotten there too, through learning.

Learning is the key. Stage 3 technology is psychopathic. To be genuinely smart, technology needs to interact with and be directed by smart people. Getting to that point is where the real opportunity lies for trading companies.

And probably for yours too.

Student and teacher of innovation - University of Queensland Business School - links to academic papers, twitter, and so on can be found here.

Please note: I reserve the right to delete comments that are offensive or off-topic.

Leave a Reply

3 thoughts on “Smart Tech Needs Smart People!

  1. Speaking from personal opinion as opposed to from any particular established school of thought; I feel that if anything, technology should be used to streamline complex people-centric processes, not to remove the human contribution. I say this because as much as technology can be quite capable, an individual piece of technology must be focused on and designed for a particular non-complex task for it to be truly effective.
    Where technology is good at mimicry and solving immense, tedious problems, it is no good at solving complex or ambiguous problems. Fortunately, people are good at addressing such problems. In a well designed process, people and technology are entirely capable of complementing and mitigating each other’s strengths and weaknesses. The first prototypical good example that comes to mind is the assembly line. The first poor example that comes to mind (for the reasons you’ve covered above) is automated trading.

  2. Definitely good points Mathieu. And I think you’re correct in framing it as a division of labour issue – that’s probably a reasonably good way to put it.