It is well known that the chances of success of the search for extraterrestrial intelligence depend on the longevity of technological civilizations or, more broadly, on the duration of the signs of their existence, or technosignatures. Here, we re-examine this general tenet in more detail, and we show that its broader implications have not been given their proper significance. In particular, an often overlooked aspect is that the duration of a technosignature is in principle almost entirely separable from the age of the civilization that produces it. We propose a classification scheme of technosignatures based on their duration, and we use Monte Carlo simulations to show that, given an initial generic distribution of Galactic technosignatures, only the ones with the longest duration are likely to be detected. This tells us, among other things, that looking for a large number of short-lived technosignatures is a weaker observational strategy than focusing the search on a few long-lived ones. It also suggests abandoning any anthropocentric bias in approaching the question of extraterrestrial intelligence. We finally give some ideas of possible pathways that can lead to the establishment of long-lived technosignatures.

众所周知，寻找外星智慧的成功机会取决于技术文明的寿命，或者更广泛地说，取决于它们存在的迹象或技术签名的持续时间。在这里，我们更详细地重新审视了这一普遍原则，并表明其更广泛的含义并未得到应有的重视。特别是，一个经常被忽视的方面是技术签名的持续时间原则上几乎完全与产生它的文明时代分开。我们提出了一种基于持续时间的技术签名分类方案，并且我们使用蒙特卡罗模拟来表明，给定银河技术签名的初始通用分布，只有持续时间最长的那些才可能被检测到。这告诉我们，除其他外，寻找大量短期技术签名是一种较弱的观察策略，而不是将搜索重点放在几个长期存在的技术签名上。它还建议在处理外星智慧问题时摒弃任何以人类为中心的偏见。我们最后给出了一些可能的途径的想法，这些途径可以导致建立长期存在的技术签名。