Oddbean new post about login | logout A lot of what you think of as essential software developments are just indifference and lack of ingenuity. They don't need to do anything groundbreaking, completely reorganize the workflows, rethink their data transmission concept, fix bugs, or refactor. They can just say: You need a bigger battery.
π
I notice this because I worked in automotive, for so long. The customers' computer hardware is embedded in the car and covered by the extended guarantee. Anytime we wanted to upgrade or change the software, we had to "make room" on the current system or buy new hardware at our own cost and exchange them in the shop. And we had a financial incentive to choose the least-powerful processor and simplest sensors possible, and then tune them and program them to maximum efficiency, because that raised the margin on each automobile, without lowering the customer experience.
When you align consumer and producer incentives, in this way, the engineers have to actually... think. How can we do the most, with the least? The cost of the thinking is high, but it only has to be done once, whereas the costs of inefficient design are long-lasting. And each inefficient version is packed onto the last one, making it worse and worse.
I am realising that "Aligning incentives" is a solution to many problems around me. At work At home with the kids Governments and citizens When resolving an argument with a neighbour ... Everywhere!
Wait a second you were in auto too??? My background was ECU firmware! We had 256k of space to work with, and only 108k of text segment.
π I wasn't developing, but I did QA for powertrain.