克劳德让人痛苦
我相信我并不是唯一一个有这种感觉的人。Claude 的表现对用户来说非常对立，坦白说，使用起来简直是一场噩梦。慢是轻描淡写……我在一台 16GB 的 Mac Mini 上使用 Deepseek 8B 模型的输出速度更快，而且输出质量也差不多。
我问 Gemini 一些问题，10 秒钟就能得到详细且长篇的回复。我问 Claude，它却浪费我的代币自言自语，做出一些无理的、愚蠢且无礼的简化回答，这些回答毫无逻辑可言，只是为了 Anthropics 的利益。
这根本无法使用。它只会给出半成品的回复，决定剔除所有有价值的内容，什么有用的东西都不放进去，留下半成的文件，最糟糕的是……它甚至不输出这些内容，而是将它们留在自己的私有记忆中……直到你明确要求它，再以全价重新生成这些内容……这简直就是纯粹的欺诈。没有其他备注。你有什么想法吗？
查看原文
I cannot be the only person who feels this way. Claude is antogonistic and frankly an absolute nightmare to use. SLOW is an understatement... I get faster output from a 16GB mac mini using deepseek 8B model....and not much worse output...<p>I ask Gemini something. I get a response in 10 seconds and its long and well detailed. I ask claude, it wastes my tokens talking to itself, makes insulting stupid and frankly rude shortcuts that have no basis in reason except for Anthropics profits...<p>Its unuseable. It will just cough up a half complete reaponse, deciding to strip everything out of value...place nothing useful in, leave half finished files and then worst of all...it doesnt even putput them and leaves them in its own private memory...until you ask for them explocitly and it regenerates rhem again at full cost....its just pure fraud. No further notes. Thoughts?