I've written about this subject as well:

link

One of my conclusions is that it is easier to build software driven risk managers than it is to build software driven housekeepers. People who are unfamiliar with both computer programming and financial markets have no idea that the world's assets are quickly being grabbed by the people who control the most effective risk management software and that money may not be so easy to come by in the future.

What is money but the ability to control human labor to do the things that machines cannot yet do? As jobs are displaced by technology, money becomes more valuable because there are more people competing for fewer jobs. To look at it from another angle, as money becomes more valuable due to job displacement, debt becomes more of a burden. I suspect most people will have a tough time controlling any human labor in a few decades. Doubly so because they will be competing with very powerful machines who would also like to allocate that labor, rather than just their peers.

Hi Jon, thanks for the comments.
Money may become more valuable - but things are also becoming cheaper. For example, only the rich had cell phones around fifteen years ago and now almost everyone has one. And even the poorest today may live a better life than middle class fifty years ago.
Also, these computers will be competing against themselves - not just humans - which should at least keep things a little interesting.
But as for trading, yes I think things may only get more difficult for humans as each day passes.

Being both a chess player (I still play a couple tournaments a year) and a securities investor not averse to taking an occasional short-term position, in my view trading seems to have more in common with, and can be informed more by, poker, than by chess. Chess is a game of perfect information because the players have access to the entire game state at all times. On the other hand, poker, and stock trading, can be viewed as examples of imperfect information games because the players have to act when they do not have access to all the information about the game state until the showdown or the SEC filing or the fraud trial. Players have to form hypotheses, make deductions and draw inferences about hidden facts about stock values and opponents poker hands. A further similarity between trading and poker is the chance element: the shuffle and deal of cards in poker and the many random occurrences from political upheaval to weather that can affect stock values make both, stochastic games, whereas chess is a deterministic game: chance figures into chess hardly at all; whomever moves their pieces to the right squares at the right time, wins. Programs have been much more successful in modeling deterministic games like chess than stochastic games like poker: chess programs can beat most accomplished human players--e.g. deep blue v. Kasparov whereas good human players demolish the best computer programs, but that is not to say very strong poker and trading programs will not be developed, and may beat humans in the future. They are much harder to program than chess. I see decline of chess ratings to be more due to people leaving the game and taking with them rating points gained at the expense of players who continue to play each other but never have the opportunity to get these lost points. Also, there has been a tremendous immigration of strong Russian and Eastern European chess players to the US recently, especially to Eastern seaboard area who suck up rating points from local players. In large tournaments like the World Open, experienced players usually discount the ratings of red state players, who get wacked by eastern players and then go home and recover their ratings at the expense of their local players. Thus over time ratings everywhere--including on-line go down. Luckily it is much more difficult to replenish lost chess rating points than trading profits!
John B.

Thanks John.
I agree with you that poker is a better anology for trading when talking about computers. And computer programs that play poker aren't as good as the computers that play chess, yet. Still, I think that trading and poker programs will get better with each day that passes and this will make it more difficult for traders (and poker players).