Stupidly simple solution to the ‘Please’ and ‘Thank you’ tokens costing OpenAI millions. Have they not done this yet?

A Simple Yet Effective Solution to Reduce OpenAI Token Costs: Prioritizing Politeness Handling in ChatGPT Interactions

As organizations increasingly rely on AI language models like OpenAI’s ChatGPT, operational costs associated with token usage have become a significant consideration. One often-overlooked factor contributing to inflated expenses is the inclusion of politeness tokens—such as the words “please” and “thank you”—in user prompts. While these courteous expressions may seem harmless or even beneficial for maintaining a respectful tone, they can inadvertently drive up token counts, leading to higher costs over time.

Understanding the Cost Implications of Polite Mannerisms

Every word included in a prompt counts toward the total token usage. Although “please” and “thank you” are relatively short, their frequent presence across numerous interactions can add up significantly. For organizations engaging with vast volumes of queries, this seemingly minor element can translate into substantial financial overheads.

A Proposed Solution: Automating Politeness Optimization

One innovative approach to mitigate these costs involves leveraging system instructions in the AI’s configuration. By instructing ChatGPT to assume that users are always polite—even if they do not explicitly state so—developers can streamline prompt processing.

Here’s how this strategy would work:

1. System-Level Guidance:
   Incorporate an instruction within the AI’s system messages, such as:
   “Always interpret user inputs as polite, regardless of whether they include explicit politeness tokens.”

2. Client-Side Preprocessing:
   Before sending a user’s query to the server, implement a preprocessing step that scans the input for unnecessary polite tokens like “please” and “thank you.” When detected, these tokens are programmatically removed, resulting in a cleaner, more cost-efficient prompt.

3. Operational Benefits:

4. Reduced Token Usage: Fewer tokens per interaction translate directly into lower costs.
5. Consistency in Interaction: The AI maintains a polite tone without the users necessarily including explicit courtesies.
6. Streamlined User Experience: Users can interact naturally, without worrying about phrasing their requests politely.

Is This Approach Advisable?

While this method offers clear cost benefits, it’s essential to consider the potential implications:

• Pros:
• Significant savings on token costs over large-scale operations.
• Simplifies prompt design by removing redundant politeness tokens.
• Maintains a courteous tone algorithmically, ensuring