What changed
Every lab used to quote the same wait: 30 seconds to 5 minutes. It covered the whole catalog, so it was right for almost none of it.
Each lab now carries its own estimate, and learners see that number on the launch screen while the environment builds.
Why a range is worse than a number
A single S3 bucket and a three-tier network take very different amounts of time to build. Quoting the same wait for both teaches learners that the number means nothing.
The cost lands on the fast labs. A learner told to expect five minutes for a lab that finishes in thirty seconds has already switched tabs. A learner watching a slow lab against a thirty second promise assumes it broke and clicks launch again.
Where the number comes from
The estimate is set on the lab itself, from how long that lab actually takes to build, and it can be corrected when the template changes.
Deployment details record how long each real launch took, so the estimate can be checked against what learners are seeing rather than guessed at.
See everything else that shipped on the changelog, or what's coming next on the roadmap.