When GPT-5 Comes Out, Current ChatGPT Will Be Laughingly Bad

MetaversePlanet May 9, 2024

0 1 minute read

OpenAI is reportedly working on the GPT-5 language model, although it has not officially confirmed this yet. According to a company executive, a year from now, today’s ChatGPT will seem ‘laughably bad’.

ChatGPT began to gain popularity with OpenAI’s GPT-3 artificial intelligence language model, and its capabilities were significantly enhanced with the GPT-4 model. OpenAI is believed to be developing the GPT-5 language model, expected to be the next iteration of GPT-4.

According to the company’s top executive, future versions of ChatGPT will undergo significant upgrades.

With GPT-5, ChatGPT will improve significantly

OpenAI has not specified a release date for GPT-5, but a company executive suggests that future versions of ChatGPT will represent such significant upgrades that older versions will seem inadequate. “I think in a year when we look back, we’ll realize how ridiculously bad ChatGPT was,” stated OpenAI COO Brad Lightcap.

According to Lightcap, the current turn-based, almost oracle-like system that ChatGPT uses is not viewed as a viable long-term approach. Instead, he envisions a future where AI can handle more complex tasks and relate more closely to users, essentially serving as a teammate. This advancement will be so significant that interacting with AI models will feel like talking to a friend or colleague.

OpenAI CEO Sam Altman has made similar remarks about future versions of ChatGPT, stating, “GPT-4 will be the dumbest model any of you will ever have to use again.”

Lightcap also mentioned that artificial intelligence is expected to support economies by creating new job sectors rather than causing job losses. However, he notes that the biggest risk associated with AI is its “energy consumption.” As OpenAI and other companies work to develop smarter AI systems, the demand for GPUs and their power consumption increases, posing a challenge that needs to be addressed in the near term.