News

Meta and OpenAI are reportedly preparing new AI models with reasoning and planning capabilities

Meta and OpenAI are on the verge of releasing the next versions of their AI models, claiming that the next models are capable of reasoning and planning. However, this may be a marketing claim with a bit of exaggeration.

No exact release schedule has been announced. According to the Financial Times, OpenAI recently mentioned its next AI model, probably called GPT-5, which will be released soon. On the other hand, Meta says the next version of Llama 3 will be introduced in the coming weeks.

Joel Pineau, Meta’s vice president of AI research, says: “We are working hard to understand how to use these models not just for conversation, but for reasoning, planning… and for having memory.”

If the claims are accurate, both models could be a major step towards achieving Artificial General Intelligence (AGI). This version of AI will likely be human-level or even smarter.

Brad Lightcap, COO of OpenAI, says: “We are seeing an AI model that can perform more complex tasks in a more sophisticated way.” He also believes that the reasoning level configuration of these models is just the beginning.

Meta plans to integrate the more powerful AI assistant and Ray-Ban augmented reality glasses with Llama 3 into WhatsApp. However, OpenAI’s plans for the release of GPT-5 are unclear.

Lightcap believes that over time, models will move towards longer and more complex tasks, which implicitly requires them to improve their reasoning ability.

The claims made by both companies are frustratingly vague; however, this is not the first time AI leaders have made such big claims.

Last summer, Microsoft researchers claimed that GPT-4 showed signs of AGI. OpenAI CEO Sam Altman has been talking about the effort to develop AGI for a long time. Meta CEO Mark Zuckerberg has explicitly stated that the ultimate goal of the company’s expensive AI business is to develop AGI or human-like AI.

Various studies have shown that AGI progress is more than just a mirage or illusion. In general, there is a lot of skepticism surrounding this idea. Perhaps there is more skepticism now than ever before.

On the other hand, other experts, including Geoffrey Hinton, known as the godfather of AI, expect AGI to be released in the next few years. Ultimately, the claims of OpenAI and Meta seem more like rumors than anything else.

Back to top button