Resource Constraint and Innovation

The no free lunch theorem applies almost everywhere. True innovation comes through iteration, and iteration must be paired with experimentation to drive convergence. But this process is costly—sometimes in terms of tools and platforms, other times in terms of people and their time. While scientists need persistence, investors must show patience. After all, scientists are artists too; they need academic freedom to explore, and that freedom must be backed by ample resources to truly thrive.

That said, scientists also need to remain mindful that everything has a cost, and sooner or later, businesses will seek a return on investment. For example, running multinode clusters just to execute single-node programs is hardly a smart solution. Similarly, data extraction could often be done on lighter compute resources, with results cached for quick reference, instead of repeatedly running heavy jobs on the fly.

On a related note, I rarely see people considering the environmental impact of their computes, documents saved online, or social media post made. These actions are the digital equivalents of leaving a small LED light on a power point or on a Wi-Fi router running 24/7. Now, multiply that energy use—or the “cloud” compute by the number of active users, and the total power consumption grows staggeringly large, very quickly. I’ll stop here on this rather somber note, not to mention about the waste of paper printing within companies—a topic equally deserving of attention.