Outlook In Robert Harris's The Fear Index, a socially hopeless physicist with a brain too large for his own good devises an algorithmic trading machine that is remarkably adept at predicting moves in financial markets. As a high-frequency trader, the computer doesn't much care what is going to happen tomorrow, it intends to make a turn in the next hour, minute or millisecond.
This being a novel, you know the computer is going to go wrong. It's so advanced that it gets smarter, keeps learning how to better make such astonishing profits that it begins to threaten the existence of global markets.
Moreover, as it learns how to ensure that it survives even if not much else does, it takes on (seemingly) human qualities.
Whether artificial intelligence plausibly could take over the world (don't say: it already has) we'll leave to theoretical physicists and Steven Spielberg.
What seems a matter of fact is that machines very like Harris's VIXAL-4 now drive most markets.
Human traders simply can't keep up, which raises the question of whether they have a future, of what they are for.
Perhaps City of London trading floors in 10 years' time will consist of nothing but banks of computers and one man whose job it is to turn them on and off.
Stockbrokers or fund mangers looking to take a longer-term view, to prepare a client for retirement or meet a pension fund obligation in 30 years' time, would still seem to have a role.
The thing is, even these people are supposed to have a view on what markets might do in the short term, or at least explain why they did what they just did.
If such moves are being driven by algorithms, what's the use of any human financial or economic analysis unless it is an impenetrable series of maths equations?
At the moment, global events are so skittish, even before computers start doing mysterious things at the speed of light, that even old hands admit they are at a loss.
I asked an experienced market-maker yesterday morning what was happening that day. He repeated what he said he had just told his best client. "We are rangebound with a slight bias to the downside."
And that means? "I've no idea what's happening."
What's the market doing in the next few days?
"Searching for direction," say the supposed experts about the hiatus in the FTSE. "Search me," they say in private.
Even the human trend-spotters, the computer-mimickers, admit that the patterns they are paid to spot are all over the place.
FxPro offered as commentary yesterday: "We are seeing a plethora of moves that are out of kilter with recent norms and correlations."
It was always the case that market participants would be lucky as often as they were foresighted, but it seems ever harder for them to make the right call based on an analysis of the evidence.
Even the most thoughtful calculations become guesswork in the face of a eurozone crisis being traded by 50,000 supercomputers.
Ten years ago, if you asked a City market-maker why a stock had finished sharply up or down on the day, they could probably give the right reason, the one that reflected how actual people had acted.
Lately, you can sense them trying to come up with something plausible rather than something they know to be true.
Where is all of this heading? Nowhere good. Ask Robert Harris.