Meta and OpenAI to launch AI models with ‘reasoning’ skills
One of the characteristics that make us human is the ability to reason. If we want an AI model with human intelligence and beyond, that is one of the most essential abilities it must have.
Many believe that having AI that can do this will take at least a decade despite ChatGPT and its AI revolution. Fortunately, two companies may move us significantly closer to that future.
READ: How to record a phone call
The Financial Times reported that OpenAI and Meta are about to release AI models capable of “reasoning.”
Moreover, Meta’s chief AI scientist Yann LeCun believes artificial intelligence will become a part of our daily “digital diet.”
More info about the AI models
Executives at OpenAI and Meta signalled that they are preparing to launch the next versions of their large language models that will be capable of reasoning and planning https://t.co/BB8KsazsDB pic.twitter.com/KuAj3XdMDg
— Financial Times (@FT) April 9, 2024
The Financial Times shared its announcements from Meta and OpenAI, two of the most prominent AI firms worldwide.
OpenAI chief operating officer Brad Lightcap told FT in an interview that the next generation of GPT would “show progress on solving hard problems, such as reasoning.”
“We’re going to start to see AI that can take on more complex tasks in a more sophisticated way,” he said.
“I think we’re just starting to scratch the surface on the ability that these models have to reason,” Lightcap added.
In contrast, modern AI models are “really good at one-off small tasks” but are still “pretty narrow” in their capabilities.
Meta, formerly known as Facebook, said it will release LLaMA 3 in the coming weeks. Meta chief AI scientist Yann LeCun shared more details about the Facebook company’s developing project.
He said it was working on AI “agents” that could “plan and book each step of a journey, from someone’s office in Paris to another New York, including getting to the airport.”
Meta plans to integrate this LLaMA 3 into WhatsApp and its Ray-Ban smart glasses. Also, the company will release it in various model sizes to fit different applications and devices.
LeCun explained that adding reasoning to AI models will let them perform the following:
- Plan a sequence of actions.
- Search over possible answers.
- Build a mental model of what the effects of its actions are going to be.
“We will be talking to these AI assistants all the time,” the chief AI scientist stated. “Our entire digital diet will be mediated by AI systems.”
