The idea this lesson is working on — Students hear "positive" as good and "negative" as bad, and so get every feedback question backwards. The book sets the lab against that: the same nudge grows under the positive loop and dies away under the negative one, and body temperature, the most welcome loop there is, is negative feedback. Ask: does the response add to the change or oppose it? The idea to reach is that the words name the sign of the response, not its value.
Positive feedback is the dangerous one — Positive feedback amplifies a change; negative feedback damps it. The lab's model grows geometrically without limit, which real climate feedbacks do not: the ice runs out, and the Planck response (a warmer planet radiates more) damps every loop, so the net effect is amplification to a finite value. Albedo values are typical: fresh snow 0.8–0.9, bare sea ice 0.5–0.7, open ocean about 0.06. The Arctic albedo decline is from Pistone, Eisenman and Ramanathan (PNAS, 2014).
Feedbacks produce thresholds — A system dominated by negative feedback returns to equilibrium. One where positive feedback takes over can shift to a new state and not come back when the push is removed, which is what a tipping point means. It says nothing about speed: some play out over centuries.