Making AI/ML Real (because it's often not yet)

Where is AI/ML going in the 2020s? It's going towards making it real - beyond consumer products. Hear me out...

AI/ML in the 2010s was marked by tremendous hype and investment. Academics, and the Deep Minds' of the world, shared exciting demonstrations of these technologies. Public and private investment followed suit, from multi-billion dollar government initiatives to DS/AI/ML teams burgeoning in even the smallest of companies. But operationalization hasn't followed suit, and I see this across the board. From my government and commercial customers at Draper, the growing emphasis isn't on 'cool demos', but creating no-kidding working end-end pipelines from data, to software, to ML, to UX.

But one place where the emperor has the least clothes is on Wall Street...

AI/ML investments in finance have often been committed at levels that would even make government PMs blush. But operations like the RenTech Medallion fund (who've been doing this for decades) continue to generate 'impossible' alpha over everybody else. And I've seen the same with projects my Draper team has led, as well with my personal projects/investments. Relatively straightforward algorithms, implemented well, can generate significant investment revenue, at this very moment - proof that even easy AI/ML alpha hasn't evaporated.

AI/ML derived investment alpha is bountiful, and the most concrete example of the tremendous benefit still to be gained from implementing these technologies.