OpenAI is thinking about a new pricing model for ChatGPT, as shared by CEO Sam Altman. In a recent interview with Bloomberg, Altman mentioned that the current pricing system for ChatGPT isn’t very detailed.
When they first introduced paid options, they only tested two prices: $20 and $42 a month. Users were not happy with $42 but liked $20, so they chose that price. There was no market research or focus groups involved; it was just a decision made based on instinct in late 2022.
Now, OpenAI is looking into a more flexible pricing method. Altman explained that many customers want to pay based on how much they use the service.
Some months, they might spend $1,000 on computing, while other months, they want to spend much less. However, he ruled out charging based on time, calling that an outdated idea.
This change seems logical, especially with OpenAI’s new “o” models, which require more computing power for better results, leading to higher costs.
The company has already entered the premium market with ChatGPT Pro, which offers enhanced computing power for $200 a month—about ten times the regular subscription price.
Despite the high price of ChatGPT Pro, Altman revealed on X that the service is actually losing money. He admitted that he set the price thinking it would be profitable, which turned out to be a mistake.
This situation adds urgency to their need to revise their pricing strategy, considering how expensive running the business is.
Even with all the discussions about pricing and products, Altman insists that OpenAI remains focused on its main goal: creating Artificial General Intelligence (AGI) and superintelligence (ASI).
Currently, the research team works in a separate building from the rest of the company, but Altman noted that this was just a matter of space planning. They plan to bring everyone together eventually while keeping research in its own area to protect its core.
What does AGI mean? Altman describes it as AI that can do the work of highly skilled human workers. He believes that if AI could perform as a great software engineer, many would consider that AGI. He thinks we might see something like this in the next four years.
However, Altman acknowledges that the term “AGI” has become unclear. There are still many unanswered questions about autonomy, and the definitions keep changing as AI technology evolves.
To address this, OpenAI has started discussing AI development in terms of specific levels to better show progress. Altman suggests that one sign of superintelligence would be AI’s ability to speed up scientific discoveries.
In the past year, OpenAI has faced a lot of internal criticism for what some see as weak safety measures, especially from its AGI and ASI teams.
The company has three safety oversight groups: the Safety Advisory Group for technical studies, a board-level Safety and Security Committee, and a joint Deployment Safety Board with Microsoft. Altman believes this three-tiered system can cause confusion within the organization.
Other Relevant AI News You May Like