OpenAI And Meta Models Will Soon Have 'Reasoning' Capabilities The upcoming major language model releases from Meta and OpenAI are expected to significantly improve reasoning and multi-step planning.
By Kavya Pillai
You're reading Entrepreneur India, an international franchise of Entrepreneur Media.
According to reports, OpenAI and Meta are about to release their most advanced artificial intelligence models, which will push the envelope even farther.
According to reports, the latest AI systems from OpenAI and Meta are significantly improving their capacity for planning and reasoning. These improved language models are being developed to truly think through problems, draw logical conclusions, and outline multi-step solutions, as opposed to just spitting out words based on patterns.
"We are hard at work in figuring out how to get these models not just to talk, but actually to reason, to plan, and to have memory," says Joelle Pineau, vice president of AI research at Meta.
COO Brad Lightcap of OpenAI revealed to the press that significant progress will be made on these "hard problems" of reasoning with their next version, which is probably going to be called GPT-5. Though still somewhat limited in its application, Lightcap recognized that existing AI is "really good at one-off small tasks."
These new AI assistants are intended to be able to perform intricate, multi-step tasks. Chief AI scientist at Meta Yann LeCun imagines AI "agents" that might arrange a whole trip, including all necessary hotel rooms, flights, and ground transportation. According to the research, major players in the sector including Google and Anthropic are also fervently developing their own updated large language models.
In the upcoming weeks and months, it is anticipated that both OpenAI and Meta will begin rolling out their new reasoning-enabled models. With the release of GPT-5, OpenAI's Lightcap hinted at "more to say soon," while Meta is getting ready to release Llama 3 in multiple sizes that are optimized for different apps and devices, such as their Ray-Ban smart glasses and WhatsApp.
Even though artificial intelligence (AGI) has no set definition, businesses are racing to develop it by attempting to solve technical challenges related to simulating reasoning, abstraction, and contextual awareness in machines. Proficiency in multi-step planning and open-ended reasoning is considered essential by researchers.
