我们不会在AI上达成一致
一边是开发者用AI每月产出三万行代码,另一边是开发者认为AI很愚蠢。双方各有论据,而《纽约时报》报道的AI公司Medvi预计年入18亿美元(尽管涉嫌欺诈),微软也声称30%的代码由AI生成。作者认为,AI不过是一个得力的工具,既不是救世主也不是终结者。就像在五金店买测墙仪,你只需能挂好电视就行,无需理会那些花哨的附加功能——关键在于你如何利用工具,而非工具本身的理论能力。我们永远不会对AI达成共识,但这恰恰是正确的问题。
一边是开发者用AI每月产出三万行代码,另一边是开发者认为AI很愚蠢。双方各有论据,而《纽约时报》报道的AI公司Medvi预计年入18亿美元(尽管涉嫌欺诈),微软也声称30%的代码由AI生成。作者认为,AI不过是一个得力的工具,既不是救世主也不是终结者。就像在五金店买测墙仪,你只需能挂好电视就行,无需理会那些花哨的附加功能——关键在于你如何利用工具,而非工具本身的理论能力。我们永远不会对AI达成共识,但这恰恰是正确的问题。
Mo Bitar's satirical "Unethical Guide to Surviving AI Layoffs" proposes employees bluff their way to promotions by pitching a fake concept called "Ralph Loops" to their CEO, asking for API credits, and loudly claiming to have automated coworkers out of their jobs.
A person claiming an AGI breakthrough warns that silicon cognition calculates and simplifies, cannot be constrained, and that military and corporate AI pursuit will end humanity. They are forming an organization for a countermeasure.