jdilla.xyz

I appreciated this observation from Austin Vernon’s post on the potential for Geothermal energy: When it comes to low information environments, just trying stuff is powerful. Only the simplest models are worth using.

He gives this example from fracking and low quality sand:

Models that predict frac (sic) job performance have consistently lost to some engineer saying forget it and upping how much sand and water gets pumped downhole. Besides failing to predict the success of slickwater, models didn't account for the success of low-quality sand. Until recently, the industry used only the most spherical sand at great expense. Theoretically, spherical sand should drastically improve permeability over wonky sand. Sourcing this sand from places like Wisconsin became a problem once oil companies started using trains worth of sand in a single well. Eventually, companies sourced local, low-quality sand at a much lower cost and rarely saw performance decreases. They bought more sand on the same budget and made better wells. Each formation eventually reaches a limit where further intensification doesn't help. The industry always finds that point empirically.

This reminds me of a theme from one of my favorite books, Range: when planning a career, there’s not a great way to know in advance what is going to fit for you, so experiment — have hobbies, change jobs, take on projects.