A few weeks ago, the AI community was taken aback when OpenAI CEO Sam Altman revealed that polite phrases like “please” and “thank you” used with ChatGPT were driving up electricity costs by millions of dollars. Interestingly, some AI experts believe such courteous language can actually yield better responses from large language models.
However, Google co-founder Sergey Brin recently made a startling counterpoint. Speaking on the All-In podcast, Brin remarked that, strangely enough, AI models—not just Google’s but others as well—seem to perform better when they’re threatened, even with suggestions of physical harm. He added that this isn’t something openly discussed within the AI community.
“But, you know, people find that kind of thing uncomfortable, so it’s not something we really talk about,” Brin added. “In the past, you might just say something like, ‘I’ll kidnap you if you don’t do this or that…’”
Sergey Brin explains what motivated his return from retirement.
Although Sergey Brin stepped back from his daily role at Google/Alphabet after passing leadership to Sundar Pichai, he has recently become actively involved in advancing Google’s Gemini AI model. In an interview with Big Technology, Brin said the current moment in tech is too important to ignore, emphasizing that now is not the time for computer scientists to remain on the sidelines.
“Honestly, if you’re a computer scientist, this is not the time to be retired,” Brin said. “We’re at an unprecedented turning point in technology, filled with both challenges and opportunities.”
Brin also made an unexpected appearance at the I/O 2025 conference last week, joining Google DeepMind CEO Demis Hassabis on stage to discuss the company’s advancements in AI. Although Google initially lagged behind when ChatGPT launched in late 2022, it has since made a strong comeback—integrating AI across its ecosystem and reestablishing itself as a major contender in the race with OpenAI.