We’ll Explore Further. Coming Soon™
As per the report by the Financial Times, OpenAI and Meta are purportedly gearing up to unveil the next iterations of their AI models. However, given the typical hype surrounding major software releases, it’s advisable to approach this news with caution.
Notably, there is no concrete timeline provided for the rollout. OpenAI vaguely mentions the swift arrival of its anticipated GPT-5 model, while Meta anticipates the imminent launch of its Llama 3 variant.
Joelle Pineau, Meta’s VP of AI research, expressed the ambition to enhance these models beyond mere speech capabilities, aiming to imbue them with reasoning, cognitive approaches, and memory retention.
The potential advancements in these versions could signify a significant leap towards achieving artificial general intelligence (AGI), a concept that envisions AI entities possessing human-like or even superior cognitive abilities. However, the credibility of such claims remains questionable, given their superficial nature.
Solemn Commitment
Brad Lightcap, the COO of OpenAI, foresees AI tackling more complex tasks with increased efficacy, emphasizing the nascent development of reasoning capabilities within these models.
Meta plans to integrate a more robust AI assistant into WhatsApp and its augmented reality Ray-Ban glasses through the forthcoming Llama 3 release, whereas OpenAI’s roadmap for GPT-5 appears less defined.
Lightcap envisions a progression towards intricate and extended tasks for these models over time, underscoring the necessity for enhanced reasoning abilities to support such evolution.
The Quest for AGI
Both parties make somewhat vague assertions, consistent with the customary approach of AI industry leaders. The tantalizing prospect of achieving sentient AI has been a recurring theme, with Microsoft researchers hinting at AGI potential in GPT-4, and figures like Sam Altman and Mark Zuckerberg championing AGI initiatives.
While some experts caution that the pursuit of AGI may be illusory, citing potential data limitations for AI training, others like Geoffrey Hinton foresee AGI realization within the next few years.
Despite the optimistic projections from OpenAI and Meta, these latest announcements may still sound more like hearsay until tangible outcomes materialize, warranting a cautious stance amidst the speculative fervor.
