Computer programs have been an integral part of the venture capitalist's toolkit for a long time. Analyzing any company involves at least some measure of computational modeling, projection, or examination, though the ultimate decision has, until now, been usually reserved for a human actor. Telefónica's Open Future entrepreneurship network is breaking with that tradition by announcing a new, machine-driven startup competition where both the contestants and the winners are selected automatically by an algorithm.
The first deployment of this algorithm will take place next year as Telefónica kicks off a series of startup battles — to be held in cities across the world, including Valencia, Buenos Aires, and Bogota — looking to identify new companies working on promising innovations in machine learning. That's right, an algorithm produced by machine learning will be put in charge of picking the most promising machine-learning startups.
The overarching goal in eliminating human intervention is "reducing the bias that subjectivity and intuition can add to decisions while improving the chances of predicting the project's success." Telefónica doesn't disclose the sums of money involved, saying only that "the best startups working on state-of-the-art predictive apps [...] will be able to access the funds managed at Telefónica Open Future." Still, even as just a symbolic gesture, this move signals a confidence in the forecasting power of computers that hasn't been around since the time of Long-Term Capital Management. Then again, LTCM is the canonical example of the folly of entrusting all your investment decisions to algorithms, so a small and cautious step would seem to make sense here.