缓解 AI 疲劳
1 分•作者: hmokiguess•7 个月前
大家好。最近有很多令人兴奋和酷炫的东西，真是个了不起的时代！
尽管如此，我还是忍不住感到一种对无限未开发潜力的恐惧和在探索这个领域时的决策疲劳。
你们是怎么应对的？什么对你们有效？你们有什么建议？
我觉得我现在想要的是一个能够在我面临这种决策疲劳时，代替我做决定的代理，我一直在从一个代理跳到另一个代理，从一个模型跳到另一个模型，说实话，它们都差不多，都能产生类似的结果，是我做错了吗？
如果我试着分析我的沮丧，我会说这是一个用户体验问题。例如，我认为Cursor通过使用新颖的UX重新发明自动补全功能，从而跨越了鸿沟，Claude Code的cli也是如此。但我觉得它们“老化”得不太好，也许只是我个人的感觉。
让我感到不安的是，我们似乎正处于一个次优的中间状态，我们试图进入一种低到中等程度监督的自主计算，但用户体验在我看来，还没有完全与这个方向对齐。——我正在努力措辞，希望在Hacker News上比我更聪明的人能理解我的意思，嘿嘿。
我喜欢我们现在正进入一个云沙盒和异步环境正在出现的时代，但不知为何，它仍然感觉很别扭。半成品，功能工厂，很多尝试一切并看看什么能行。——也许这只是知识差距、信息密度/过载，或者也许我只需要更专注。
有没有人觉得现在有一场比赛，上市时间是首要任务，而不是质量和创新？我们什么时候才能再次慢下来？
好吧……是时候回去在Sora上刷末日信息了 ¯\_(ツ)_/¯
查看原文
Hey folks. Lots of exciting and cool stuff out there these days, incredible time to be alive!<p>That said, I can't help but feel the dread of infinite untapped potential and decision fatigue when it comes to navigating the space.<p>How do you cope? What works for you? What do you recommend?<p>I feel like what I want right now is some agent that can replace my decision making when it comes to this decision fatigue, I keep jumping around from one agent to another, one model to another, they all sort of work and produce similar results if I am being honest, am I doing something wrong?<p>If I try to reason about my frustration, I would say it is a UX problem. For example, I believe Cursor managed to cross the chasm when they reinvented autocomplete using a novel UX, same goes for Claude Code with its cli. That said, I feel like they are not "ageing" well, maybe it's just me.<p>Here's the thing that is upsetting me, it feels like we are in a suboptimal in-between state where we are trying to enter a sort of autonomous computing with low to moderate supervision but the UX is still not quite fully aligned with that direction the way I see it. -- I'm struggling with words, hopefully someone smarter than me on HN can understand what I mean here, heh.<p>I like we're now entering an era where cloud sandboxes and asynchronous environments are emerging, but it still feels awkward for some reason. Half-baked, feature factories, a lot of try everything and see what sticks. -- Maybe it's just knowledge gaps, information density/overload, or perhaps I just need to focus better.<p>Anyone else feels like there is a race where time to market is the priority rather than quality and innovation? Are we ever going to slow down again?<p>Oh well . . . time to go back to doomscrolling on Sora ¯\_(ツ)_/¯