ChatGPT developers started working on a new AI project called Strawberry
OpenAI, a startup that has gained worldwide fame for its ChatGPT AI product, has started working on the Strawberry project.
According to Reuters sources, this initiative is aimed at forming a neural network capable of conducting research on the Internet on its own (and not just answering questions). If the agency’s information is true, ChatGPT’s analytical abilities will come close to human ones.
Modern language models can summarize and compose texts faster than humans, but they cannot cope with tasks that require intuitive thinking (recognizing logical errors, etc.).
Strawberry’s direct predecessor is the so-called Project Q, which led to the dismissal of the startup’s CEO, Sam Altman, last November. OpenAl employees called the project a “breakthrough” in the field of powerful AI, but expressed concern about the lack of security measures for the controlled development and commercialization of such models.
Altman already declared the inevitability of AI’s “ability to reason”.
It should be noted that the scientific community has long been discussing the dangers of highly intelligent machines, which may one day conclude that it is advisable to destroy humanity.
