The next logical stage for Autonomous Driving and Computer Vision systems.
Like in the previous piece, this brief text was a quick analysis created as part of a job application. It ended up being a pretty good biological heuristic that can predict new technological niches and a given tech’s developmental path.
As 2022 approaches, it is convenient to take a moment to reflect on the general landscape of Autonomy, and where will the compounding power of the combination of advanced visual sensors and super-efficient computation might make its next nest on. This piece explores the next logical developmental stage of Machine Learning based Computer Vision and Autonomous Driving Systems.
While the race for the understanding and domination of self-driving automobiles has taken off, most of us understand how and why Tesla has a substantial lead in training data, simulation, computation and actual coding… but the efforts and results of other companies are also worthy of consideration for this analysis (ex. comma.ai with its super-contained and modular smartphone-based solution).
Like in biological organisms, it is usually a good heuristic, that of analyzing a given emerging technological tendency by looking at the physical scenario or niche where it’s taking place, given that (as with most complex systems) the internal characteristics of the phenomena are highly context-dependent. Let it be chemical molecules shaping the right RNA combination for unicellular organisms in the mythical primordial soup, or a cloud-based ML computervision system feeding Millisecond-level multi-sensor data from hundreds of thousands of agents in semi-realtime… It is all about information exchange between Agent/Arena in pursuit of an accurate internal representation of the environment that allows agents to maximize Return of Energy Investment and minimize Risk.
Now, the common factor between all the competing participants, is that the vast majority of the attention has been put in making vehicles capable of finding their ways in urban environments… but as of today 44% of the world’s population lives outside or far from the geometrical comfiness of the cities.
Without going too deep on the particularities of rural contexts, their increased dynamism and ever-changing nature, the elevated difficulty of parametrization, the abundance of borderline cases and substantial general increase in the complexity of visual data and respective ML models; we can say that the next competition niche for Autonomous Systems will be in the Rural Transportation and Agricultural sectors.
So… the first companies to deploy their capabilities in these niches will dominate the highly strategic first few hundred miles of Food Production in their respective supply chains.