Why is ChatGPT Getting Worse? — Answer Revealed

Discover why ChatGPT is getting worse over time.
This article discloses:

• Evidence that GPT 4 is getting worse over time.
• Demonstration that explains why this is happening.
• How to protect your company and yourself against devolving LLM performance.

GPT 4 is Getting Worse

A Stanford/Berkeley study empirically documented “that the performance and behavior of both GPT-3.5 and GPT-4 can vary greatly over time. For
example, GPT-4 (March 2023) was reasonable at identifying prime vs. composite numbers (84% accuracy) but GPT-4 (June 2023) was poor on these same questions (51% accuracy).” This study provided “evidence that GPT-4’s ability to follow user instructions has decreased over time, which is one common factor behind the many behavior drifts. Overall, our findings show that the behavior of the ‘same’ LLM service can change substantially in a relatively short amount of time.” (http://arxiv.org/pdf/2307.09009)

This is a tremendous problem for companies and developers alike. An app built on ChatGPT might work today and then miserably fail tomorrow. In a moment, we’ll cover how to escape this game of Whac-A-Mole. However, let’s first dig into why this happening. What’s going on…