Hana Sorensen · 2026-03-02
Designing lab resets that do not erase learner confidence
labs · operations · teaching
We treat every reset as a teaching moment, not a punishment. When a learner breaks a subnet on purpose, the environment should return to a known baseline within minutes so curiosity stays high.
Before we automate anything, we script the human steps a facilitator follows: who gets notified, what gets logged, and how we explain the next attempt. That sequence is more valuable than a fancy portal button.
In practice, teams that document resets alongside the learner’s own notes retain skills longer because they can replay the mental model, not just the clicks. We publish those paired notes inside the cohort space with consent.
Finally, we avoid stacking multiple unknown variables in the same session. If identity policies are new, we keep networking stable. If networking is new, we freeze identity changes. That discipline keeps support queues calm.