OpenAI, the company that developed ChatGPT, has launched a new project called “Strawberry.”which aims to dramatically improve the logical capabilities of artificial intelligence. The project is backed by Microsoft and aims to make AI more effective at designing and solving complex problems. The details of “Strawberry” come from internal documents and a source familiar with the matter, who is Reuters I reported.
Strawberry project. Not only does it generate answers, it allows the AI to navigate the internet autonomously and conduct in-depth research. This would allow models to tackle complex problems that require long-term planning, a capability that models currently lack. Strawberry is a continuation of an earlier project, Q*, which has already demonstrated its advanced capabilities in answering scientific and mathematical questions.
Used in “strawberry”. “After training” A method similar to that developed by Stanford University. Self-made Mind (STAR) A technology that allows AI to continuously improve its performance by generating its own training data. This process could help models achieve higher levels of intelligence than humans. According to internal OpenAI documents, Strawberry will be able to perform complex, long-term tasks, including independent web searches and software application development.
However, the challenges remain significant. Current AI models can quickly summarize dense text and write elegant prose, but they struggle to handle common sense problems and sometimes “hallucinating” wrong information. Strawberry aims to overcome these limitations.This allows AI to plan for the future and solve complex problems reliably. However, creating AI with advanced reasoning capabilities and a high level of autonomy raises serious questions regarding security and control.
If these models get out of control, they could pose a serious risk to humanity, as they can make decisions without oversight and cause unpredictable consequences.
