Korva is a scroll of the internet’s best ideas.
Startups turn on all the water
Turn on all the water
Picture a house with a number of cracked pipes and no way to know which ones. A big company person checks every pipe one by one, and it takes two years. A startup person turns on all the water, sees where it leaks, and fixes that pipe.
Big companies plan to be error free. At a company like Google, that is why anything takes forever. A startup's path is full of errors, and the errors are the useful part, since each one points at what to fix next.
When demand crashes you
If a product is so popular that it crashes, the motivation to fix the bugs goes through the roof, and they get fixed quickly. Startups dying because they crashed under demand is something nobody in this group has seen. The problem is solvable because the incentive is so high, and a company with that problem is on the path to becoming a large one.
One test for the startup playbook: whatever would have gotten you fired at a big company is what to do.
How do engineers break systems on purpose?
Netflix built a tool called Chaos Monkey in the early 2010s that randomly shuts down servers in its production systems. The goal is to expose weak points while engineers are watching rather than during a real outage. The practice grew into a field called chaos engineering, which other large companies have adopted.
Why do big companies try so hard to avoid errors?
A company with millions of customers, regulators, and public shareholders pays a high price for any failure. Large organizations respond with review stages, sign-offs, and testing cycles meant to catch problems before release. Each added step slows the pace of change, a pattern described in studies of organizational bureaucracy.
What was Twitter's Fail Whale?
The Fail Whale was an illustration that Twitter showed users during outages, most famously in the late 2000s. The outages came as usage grew faster than the company's infrastructure. Twitter rebuilt much of its backend over the following years, and the image was retired.
More from Y Combinator
