Ask HN：你如何防止技能退化？
4 分•作者: xpnsec•4 天前
我对科技行业从业者在当前大语言模型（LLM）时代如何避免技能退化很感兴趣？
我们都听过两种截然不同的观点，一种是“让他们退化吧，LLM才是未来，看看算盘就知道了！”；另一种是“我不使用LLM，它们会犯错，还会碍事”。但对许多人来说，现实情况是，即使LLM会犯错，需要人工干预，它们也能真正提高工作效率，并承担许多任务。
我倾向于谨慎对待技能落后。LLM在中长期内对工作场所的推动作用以及未来会用到哪些技能，都存在太多未知因素。所以，我想知道，当面对“一键生成”的诱惑时，你们是如何防止现有技能退化的？
查看原文
I’m interested in how people working in tech are avoiding skill atrophy in this LLM age?<p>We’ve all seen both sides of the argument spectrum, with “Let them atrophy, LLMs are the future, just look at the abacus!” to “I don’t use LLMs, they make mistakes and get in the way”. But the reality for many is that LLMs give a real performance boost and take on many tasks, even if they do make mistakes and need babysitting.<p>I fall on the side of caution in letting skills fall behind, there are too many unknowns in how much LLMs will drive workplaces in the medium to long term and what skills will be used, so I’m wondering how you keep your existing skills from atrophying when tempted with the “make it so” button?