The AI Technology Page:

OpenAI Is Working On A New Project That Will Produce A Model that Has Better-Than-Human Intelligence:

According to leaked internal documentation and information from an insider, OpenAI is working on a new model that has human-like reasoning capabilities under the code name “Project Strawberry”.

🔑 Key Points:

Project Strawberry is being trained to “navigate the internet autonomously” and undertake “deep research”—independently, without being told to do so by the user—to solve complex problems.

Because it can undertake “deep research”, Project Strawberry will serve up fewer hallucinations, as models like ChatGPT often resort to making up information when they lack training data.

Project Strawberry uses a Self-Taught Reasoner (STaR) method, which means it improves itself by learning from its own reasoning, meaning it could reach better-than-human intelligence over time.

🤔 Why you should care: 

Although it’s unknown if Project Strawberry will launch with the upcoming GPT-5 (which is believed to be as intelligent as someone with a PhD), this news comes after concerning reports from ex-OpenAI employees revealed that OpenAI continuously prioritizes new launches over safety, and after multiple OpenAI employees wrote a letter to the board warning of a new AI breakthrough that “could threaten humanity.”