OpenAI’s latest version of ChatGPT, GPT-5, was released on August 7 with great fanfare, but its reception was underwhelming due to the removal of key models and rate limits for paying users. In a post explaining the discrepancy, CEO Sam Altman revealed that only 1% of non-paying users and 7% of paying users queried reasoning models like o3 before GPT-5’s release.
This points to a widespread issue: most people aren’t using AI to its full potential. They prioritize speed and convenience over quality, often opting for simpler chatbot interactions instead of more thorough ones. This is reflected in the fact that only 1 in 4 paying users are asking for thoroughness from GPT-5.
The problem stems from a lack of understanding about what reasoning models can offer. These models “think” through problems before answering, providing better results but at a cost – speed and convenience. OpenAI’s attempts to address this by adding a drop-down box for alternative models have been met with criticism.
However, the data suggests that users are still not utilizing AI as intended. Only one in five Americans who have used a chatbot say it’s not useful at all, which is twice the rate among experts. The answer lies in using AI wrong – asking chatbots to handle complex questions without pausing for thought or breath.
OpenAI’s recent moves aim to change this by making more access to models available. Users should take advantage of these changes and use AI with intention, setting it to “think” while remembering its limitations. By doing so, they can unlock the full potential of generative AI.
Source: https://www.fastcompany.com/91385268/most-people-are-using-chatgpt-totally-wrong-and-openais-ceo-just-proved-it