A very different ChatGPT: OpenAI is already teasing GPT-6, the next version of AI

© 01net avec Photoshop

Also read: OpenAI is considering launching an essential security feature on ChatGPT

GPT-6 will arrive earlier than expected

Shortly after the launch of GPT-5, Sam Altman, CEO of OpenAI, mentioned the arrival of the next version of ChatGPT, which is GPT-6. In front of a group of journalists, he assured that GPT-6 was already in development and that it won't take too long to arrive, according to CNBC.

In fact, Altman states that OpenAI won't wait as long as between GPT-4 and GPT-5. The release of the two models was spaced more than two years apart, with GPT-4 being unveiled in March 2023, a few months after the ChatGPT wave began. During this time, the company has nonetheless occupied the market by releasing a multitude of intermediate models. In short, one won't have to wait until 2027 to see GPT-6.

Better memory for ChatGPT

The CEO of OpenAI reveals that GPT-6 will be very different from its predecessor. The next version of ChatGPT will mainly adapt more to its users. OpenAI wants internet users to be able easily personalize the chatbot according to their personal preferences. In this regard, the AI giant has already launched a range of features, starting with different personalities that influence how ChatGPT answers questions.

For Altman, the key to ChatGPT's personalization is memory. To offer a truly personalized experience, the conversational robot must remember who you are, along with your preferences, habits, and a bit of quirky mannerisms, estimates the executive.

« People want AI to have memory. They expect features that require us to be able to truly understand them », says Sam Altman.

OpenAI has already significantly improved ChatGPT's memory through the versions. ChatGPT is indeed capable of referencing all your past conversations to tailor its responses. For example, ChatGPT can remember that a user is allergic to a fruit or prefers a certain response format, thus avoiding having to remind them in each new conversation. Of our side, we asked the chatbot to prioritize answers in bullet points.

Finally, Altman promised that future versions of ChatGPT will be ideologically neutral. In line with the requirements of the U.S. government, OpenAI will allow each user to calibrate the chatbot's ideological positioning. In short, ChatGPT can be « super woke » or rather conservative.

👉🏻 Follow tech news in real time: add 01net to your Google sources, and subscribe to our WhatsApp channel.

Source: CNBC