IT Home reported on May 31 that at the All-In-Live event in Miami, Google co-founder Sergey Brin introduced a provocative idea about artificial intelligence interaction: sometimes threatening AI models can actually stimulate and improve their performance. This insight applies not only to Google’s AI models but extends across most AI models currently available on the market.
- Xiaomi 17T India price and storage options tipped ahead of 4th June launch
- Google AdSense Expands Dynamic Anchor Ads to Desktop for Higher Publisher Revenue
- [Exclusive] iQOO 16 5G Appears on IMEI database once again, confirms Global and India launch
- HMD Vibe 2 5G Launched in India with 120Hz Display, 6000mAh Battery and Unisoc T8200 Starting at ₹10,999
- Xiaomi 17T and Xiaomi 17T Pro Launching on May 28 With Leica Cameras, HyperOS 3 and Massive Batteries
Brin’s statement, as translated by IT Home, challenges conventional beliefs around AI communication. He said, “I’m not sure how many people in the AI community know about this, but based on our tests, the performance of almost all models is better after being threatened.” This overturns the common practice where users tend to include polite phrases like “please” and “thank you” in their AI prompts.
This unconventional approach contrasts sharply with the general user etiquette towards AI systems. Adding to this discussion, OpenAI CEO Sam Altman has humorously remarked that processing polite words such as “please” and “thank you” costs tens of millions of dollars in electricity.
Brin’s comments open up fresh perspectives on optimizing AI interactions and performance, signaling new directions for developers and users exploring artificial intelligence.
For more insights on AI developments and technology updates, stay tuned to PassionateGeekz.com.
0 Comments