A new release of OpenAI, Strawberry, the reasoning-centered artificial intelligence to integrate into ChatGPT, will be released in the next two weeks. Two people gave the above information and got the chance to utilize this model.
While Strawberry, like other conversational AI models, can answer questions posed to it, what differentiates it is its capacity to ponder questions before responding to them. Still, more concretely, it will be embedded into (and named) ChatGPT-proper; Strawberry is also expected to be offered as a distinct product type, although the precise form of it is unknown at this time.
One possible way users could interact with Strawberry is to add it to the list of AI options that the website offers and which can integrate with ChatGPT to expand its capabilities.
A Reuters report in July suggested that under project Strawberry, OpenAI—an organization in which Microsoft is an investor—was working on this new and exciting technique. The project’s goal is to enable AI to come up with new responses and do the planning and Internet searches on its own, something that the previous models could not do.
In addition, Strawberry is said to utilize a distinct processing approach after pre-training on large data sets, much to improve the reasoning of OpenAI’s AI models. An OpenAI spokesman said that the idea is to make AI systems observe and interpret the world ‘in the manner of the rational animal’.
The foremost scholars suggest that the enhancement of reasoning is a prerequisite for reaching such (human/superhuman) AI. This is important because big language models have been proven to have deficits in common-sense reasoning or the tasks that people consider obvious, like recognizing the fallacies and playing the tic-tac-toe.
