OpenAI develops Strawberry, the artificial intelligence that predicts our thoughts

OpenAI , a company which produces ChatGPT and which is considered somewhat of a flagship of artificial intelligence made in America, is working on a new approach for its artificial intelligence models. And the project already has a well-defined name: 'Strawberry'. The rest is fairly top secret, with the management of the San Francisco-based company intent on maintaining strict secrecy about an innovation that could be a new step forward in the field of generative artificial intelligence.

According to a document viewed by the Reuters news agency, the AI models based on Strawberry would aim to enable the artificial intelligence of Sam Altman 's company to not only generate answers to questions posed by users, but to plan in advance the Internet surfing autonomously and reliably to perform what OpenAI calls 'deep reseach'.

Prediction is an aspect hitherto unexplored by the GenAI giants. But it also seems to be an already defined frontier. And OpenAI wants to put its own flag on it, a bit like it did with ChatGPT a year and a half ago, when it gave the necessary jolt to the generative artificial intelligence boom.

When asked about Strawberry and the details of this story, an OpenAI spokesperson responded in a statement rather evasively: "We want our AI models to see and understand the world more than we do. Continued research into new AI capabilities is a common practice in the industry, with a shared belief that these systems will improve reasoning over time."

According to rumours, Strawberry is nothing more than an evolution of Q*, OpenAI's top-secret project that was much talked about a few months ago. Several sources told of seeing what OpenAI employees said earlier this year were demonstrations of Q*, capable of answering difficult scientific and mathematical questions that are out of reach of the models available on the market today.