Hacker News Reader

I remember that too, the ambiguity for me was how their fuzzing was good enough to explore an arbitrary state space efficiently enough.

The deterministic hypervisor is 'simple' enough albeit a pretty heavy engineering lift.

View on HN

One of the cool tricks we can use is that since the testing is all fully deterministic, once we find an interesting point in a test run - even if it is “deep” into the run time wise - our system can start many new branches of test runs off of that moment or moments just prior. So it is much more efficient than having to re-do the work to get to that rare interesting moment for each new branch.

I’m curious if you’re willing and able to share: Are you using FoundationDB as the data store for Antithesis?

We’ll be writing a lot in the near future about how Antithesis works, stay tuned :)

Can’t wait!

Reply on news.ycombinator.com