Experts are overfitting to the territory. It is the territory which changes (or expands), not their knowledge of it. There is a necessary reason to the decay in pheromone trails: adaptivity.
Compare an army scout on a Starcraft map. First they make visible the entire map. Then they start tracking details. They detect patterns: if it rained last week, this area will be flooded, and other routes are faster.
The army scout operationalized their expertise, made it valuable to others, and hence has something of value to exchange. Anything which attacks their operationalized expertise thus attacks their livelihood. They invested all this energy into materialization, and are prone to sunk-cost fallacy and economically motivated thinking. Don't patch their exploit!
If the hive has enough energy for another scout, then too much communication and history of failure of the current expert may negatively effect their own diverse map constructions. There are multiple ways to victory and it is better to know all of these in case one way is flooded. But it is very risky for one agent to explore all these roads.
Imagine any time the new scout wants to veer of the map, or when it has rained, the expert loudly exclaims: I already checked that road last year, you can not veer of the map there. Also, take these roads, they are always shorter when it rains. Now last month, the roads stopped flooding after rain, but the expert did not waste a failed exploration on that, and the new scout never gets to try (they also do not want to waste energy on failures, but now contribute to enforcing the wasted/stale energy materialization of the outdated expert.
For the expert to admit their ideas are outdated, is to admit their own loss of value. Retiring a fighter who practiced a defense against a kick people stopped using in the 90s. Hardly, if ever, they themselves step down. They want to keep "continuous learning" and adding even more detail to their expert map. When the hive veers off the map, the expert knows they are at a disadvantage, their value no more than a newby scout. Ossification is in their best interest (but not in the interest of the hive).
https://www.pnas.org/content/118/41/e2021636118
The size of scientific fields may impede the rise of new ideas. Examining 1.8 billion citations among 90 million papers across 241 subjects, we find a deluge of papers does not lead to turnover of central ideas in a field, but rather to ossification of canon. Scholars in fields where many papers are published annually face difficulty getting published, read, and cited unless their work references already widely cited articles. New papers containing potentially important contributions cannot garner field-wide attention through gradual processes of diffusion. These findings suggest fundamental progress may be stymied if quantitative growth of scientific endeavors—in number of scientists, institutes, and papers—is not balanced by structures fostering disruptive scholarship and focusing attention on novel ideas.
A crafts teacher divvied up his class in two parts. One part he taught his expertise at vase making, down to the level of detail, in context of art history. They were to be graded on a single vase, so failure was frustrating and a frightening loss of energy/hurt ego. The other part of the class was to be graded on the number of vases they made. The crafts teacher focused on teaching them to learn from mistakes, cut losses and start over, how to get better at the mechanics instead of the art. In the end, the part of the class graded for most vases made, also created the highest-graded single vases. Failure was a necessary part of the initial exploration phase, allowing them to exploit unclaimed ground, instead of trial-error-mimicking already-existing expertise. There is a lesson there, I think. Deep Learning comes to mind, before and after its hype. Where the statisticians correctly calculated flooded roads of overparametrization, engineers still charged ahead, and some actually came out alive on the other side, establishing a shortcut/conquered obstacle. Some statisticians still don't want to get their feet wet. And this ok too! There may be more elegant ways, which keeps dry boots with similar outcomes. Let them find these.