Brin said that threatening AI can improve the quality of answers, and Altman responded that polite language "is worth tens of millions of dollars".

Brin Says Threatening AI Can Improve Answer Quality, Altman Responds that Polite Language is "Worth Tens of Millions of Dollars"

Sergey Brin, the co-founder of Google, said at the All-In-Live event in Miami that most generative AI models perform better when "threatened" (even using words related to physical violence), "not just our models, but almost all models." He said this phenomenon is "not often publicly discussed" in the industry.

Previously, when Sam Altman, the CEO of OpenAI, was asked by a netizen "how much electricity does it cost to say 'please'/'thank you'", he responded: "It's worth tens of millions of dollars - you never know." One "threatens" and the other is "polite", the two giants highlight the subtle impact of the tone of prompt words on the cost and effectiveness of the model from different perspectives.

IT Home | Benzinga | Futurism

📮Contribute ☘️Channel 🧧JD

via Tech Circle🎗 Zaihua Channel📮 - Telegram Channel