Openai CEO Sam Altman pulled the curtains back with the strange and attractive cost driver behind ChatGpt: Polityness.
In a recent statement, Altman revealed that simple phrases like “please” and “thank you” are quietly earning millions of computing costs. why? All words entered into chatgpt will be processed as tokens. The model bites and generates a response. Adding some additional words to billions of queries, suddenly, you’re looking at a serious tab.
Certification to ChatGpt costs millions
The conversation began with an X when a user named Tomie asked, “How much money has Openai lost in electricity costs from people saying “please” and “thank you” to the model?” ”
Sam Altman didn’t miss the beat. Openai CEO replied:
“Ten millions of dollars have been spent well. I’ll never know.”
Tens of millions of dollars have been spent well. You never know
– Sam Altman (@sama) April 16, 2025
Why people continue to be polite to AI anyway
Despite the cost, users are not in a hurry to stop saying “please.” And it’s not just nostalgia and muscle memory. Many posts on X show that people use courtesy as a way of staying based on human experience, even when talking to machines.
Tomie said, “Sama I just want to thank Openai. Your model is very useful for helping with complex tasks like understanding their names.”
Back in the summer, Scientific American explored why so many people are polite to chat with. This article pointed out the possibility that social norms and courtesy can improve AI responses. “The benefits of being polite to AI may include encouraging better chatbot responses and fostering our humanity,” the article states.
Those feelings also manifested in real-world behavior. In an informal online survey by University of Pennsylvania professor Ethan Morrick, nearly half of respondents say chatbots are often polite. Only about 16% agreed to “just order.”
Even developers reflect this habit. On the Openai forum, one user explained: [ChatGPT] Because that’s how I talk to the real people who helped me.
AI token economy and technical sacrifice
The backend story is very simple: more tokens = more processing. ChatGpt’s response is equipped with a GPU-rich infrastructure. Even a small addition to an input like “please” will require a little extra calculation. One token doesn’t cost much, but spanning billions of interactions, mathematics catches up quickly.
Large language models like ChatGpt break down user input into tokens. This is a small text that contains words, punctuation and even spaces. Each token taps on GPU cycles, electricity, and server time computing resources. So, when users add polite phrases like “please” or “thank you,” it’s not just a word. That’s an extra treatment.
On a large scale, it is summed up. ChatGpt processes millions of queries every day. Multiplying a few extra tokens per prompt results in significant costs.
A March 2025 report from MIT Technology Review notes that individual queries consume only energy slivers, but volume is growing concern. When billions of interactions accumulate, even small inefficiencies are accompanied by a large price tag.
Therefore, these words have a ripple effect. It’s not huge for everyone. But is it large? That’s real money.
Efficiency vs. Etiquette
Altman’s comments highlight a bigger question for AI platforms. How do you scale efficiently without alienating users who value natural interactions?
The Bloomberg report also pointed out that AI companies are quietly testing ways to reduce token bloating. But there are trade-offs. If it’s too far, you start to feel like you’re being told how to talk.
This is where things become tricky. If people start treating AI more like a utility and treating it like a conversation partner, what do we lose in the process?
So… should we stop?
That’s a million dollar question. When I ask Openai’s finance team, it looks like millions. Openai could eventually build a system that trims filler words before hitting the GPU. Or they may launch an awareness campaign on token restrictions. However, depriving pleasure comes with its own costs. It makes the interaction feel colder and more mechanical.
For now, users are split. Some people trim their fluffs to keep things lean. Others are stuck to manners, out of principles, habits, or superstitions.
Whether it’s efficiency, value, or future robotic karma, one thing is clear. After all, small words aren’t that small.
🚀Want to share the story?
Submit your stories to TechStartUps.com in front of thousands of founders, investors, PE companies, tech executives, decision makers and tech leaders.
Please attract attention
Source link
