The companies OpenAI e Meta are about to launch new artificial intelligence models capable of reasoning, marking a significant advance towards superhuman cognition in machines, reported the Financial Times
ADVERTISING
Both companies have signaled the next version of their big language models, which power generative AI applications like ChatGPT.
Meta announced the release of Llama 3 in the coming weeks, while OpenAI has indicated that its next model, expected to be called the GPT-5, is coming soon.
The focus is on advancing the development of AI models that not only generate text, but are also capable of reasoning, planning and memory.
ADVERTISING
🇬🇧 Subscribe to the best newsletter about AI (in English 🇬🇧) (I.e. Sign it the best newsletter about AI (in Portuguese 🇧🇷)
Releases in 2024
The companies' advances are part of a wave of big new language models being released this year, including companies like Google, Anthropic and Cohere.
The ability to reason and plan is crucial to achieving what researchers call “artificial general intelligence,” as it allows virtual assistants to complete sequences of related tasks and predict the consequences of their actions.
Second. O Financial Times, both the Target and the OpenAI are working on AI models that can plan and execute complex sequences of actions, like planning a trip from Paris to New York.
ADVERTISING
At an event this week in London, Meta's chief AI scientist Yann LeCun said that current AI systems “produce one word after another, really without thinking and planning.”
Because they have difficulty dealing with complex issues or retaining information for a long period of time, they still “make stupid mistakes,” he said.
Whatsapp and Ray-Ban
🇬🇧 Subscribe to the best newsletter about AI (in English 🇬🇧) (I.e. Sign it the best newsletter about AI (in Portuguese 🇧🇷)
Meta plans to integrate its new AI model into WhatsApp and its Ray-Ban smart glasses, with the launch of Llama 3 in different model sizes for different applications and devices.
ADVERTISING
A OpenAI promeSee more details soon about the next version of GPT, indicating that models tend to evolve to perform longer and more complex tasks, which requires an improvement in their reasoning capacity.
The vision is that AI assistants, powered by these new models, will be present in many aspects of people's digital lives, assisting with everyday tasks and mediating online interactions.
