A big chunk of how the Polylane team builds software now happens in Slack, with agents fixing, reviewing, and arguing over PRs. @boristane on why the next step is closing the loop on what happens after code hits production.
Hours debugging a failing CI run, only to find out GitHub is down again. @boristane on teaching Polylane to tell the difference between your app breaking and the platform underneath it.
An agent on call needs more than one kind of memory. @boristane breaks down Polylane's layers: a map of your system, what it learns from your team's feedback, and lessons from past investigations that keep it from chasing false positives.
Developers are lazy, so an onboarding flow that starts with "tell us everything about your stack" is dead on arrival. @boristane on how Polylane starts from one connection and discovers the rest of your infrastructure on its own.