Wired profiles the rise of artificial intelligence (AI) in investment management. Ben Goertzel’s company, Aidyia, recently began using AI to make real trades. Another company, Sentient, has been making trades according to AI recommendations since last year, according to CEO Antoine Blondeau. AI might be seen as an evolution from the use of complex statistical computer models to inform trades, but the two are different in that quant models tend to be static and AI is designed for machine learning. Two techniques that may be built into AI illustrate the point. “Evolutionary computation,” as explained by Blondeau, involves creating digital traders and tests their performance using historical data (not actual trades) so that “Over thousands of generations, trillions and trillions of ‘beings’ compete and thrive or die . . . eventually, you get a population of smart traders you can actually deploy.” While evolutionary computation appears to be the state of the art, there is debate about whether “deep learning” algorithms may provide another valuable method of AI investing. Sentient’s chief science officer, Babak Hodjat, sees potential in such algorithms, which are perhaps best known for learning to identify images, spoken words, and the natural patterns of human speech. Goertzel of Aidyia disagrees, however, suggesting that the patterns in the market are very different than the kinds of patterns that deep learning has been useful for. He also notes that deep learning may lose any advantage it creates by becoming widespread: “if everyone is using something, its predictions will be priced into the market.” Hodjat’s response to these concerns is that evolutionary computation and deep learning may be combined: “You can evolve the weights that operate on the deep learner . . . but you can also evolve the architecture of the deep learner itself.” Author and finance manager Ben Carlson challenges the potential of AI overall: “If someone finds a trick that works, not only will other funds latch on to it but other investors will pour money in too. It’s really hard to envision a situation where it doesn’t just get arbitraged away.” Goertzel, as his concerns about deep learning suggest, is sensitive to this risk and seems determined to stay on the cutting edge for that precise reason: “Finance is a domain where you benefit not just from being smart . . . but from being smart in a different way from others.”