SUPERNEWS

World Wide

AI Giants OpenAI and Meta Unveil GPT-5 and Llama 3 Models, Google Struggles in AI Race

OpenAI and Meta to launch advanced AI models, signaling a leap towards machines with reasoning capabilities.

By Alex P. Chase

ᐧ

4/9, 21:10 EDT
article-main-img

Key Takeaway

  • OpenAI and Meta announce new AI models, GPT-5 and Llama 3, enhancing reasoning capabilities, with OpenAI's revenue hitting $2bn annually.
  • Google struggles in the AI race due to internal issues and competition, while OpenAI expands with GPT-4 Turbo and customizable ChatGPT options.
  • Meta to integrate its AI into WhatsApp and Ray-Ban smart glasses; OpenAI reduces fees for its software to lead in the generative AI market.

AI Advancements Announced

OpenAI and Meta have unveiled plans to launch new artificial intelligence models, signaling a significant leap towards machines capable of reasoning and planning. Meta is set to introduce Llama 3, while OpenAI, with Microsoft's backing, anticipates the release of what is expected to be named GPT-5. These advancements are seen as crucial steps towards achieving superhuman cognition in AI. Joelle Pineau of Meta emphasized the focus on enabling models to reason and plan, while OpenAI's Brad Lightcap highlighted the upcoming models' capabilities to tackle more complex tasks with sophisticated reasoning. This development is part of a broader wave of innovation in the AI sector, with companies like Google, Anthropic, and Cohere also releasing new large language models.

Strategic Applications and Business Models

Meta plans to integrate its new AI model into applications such as WhatsApp and Ray-Ban smart glasses, aiming to offer a range of model sizes for various applications and devices. OpenAI, on the other hand, has seen its business model thrive, with revenues surpassing $2bn on an annualized basis. The company's products, including the viral ChatGPT and its underlying model GPT-4, are used by a significant portion of Fortune 500 companies. OpenAI's strategic partnerships and product offerings have positioned it as a leader in the rapidly growing generative AI market, despite facing challenges such as the high costs of training and running large language models.

Google's AI Challenges

Google has encountered difficulties in maintaining its lead in the AI race, with internal challenges and competition from companies like OpenAI and Microsoft. The company's efforts to launch a competitor to OpenAI's GPT-4 were hampered by internal competition and a lack of a unified approach. Google's Bard chatbot, intended as a response to ChatGPT, faced criticism for its lack of sophistication. Despite these challenges, Google is working on enhancing its generative AI services, including adding a premium tier to its search engine. However, the company's organizational structure and big-company inertia have been obstacles to its AI ambitions.

OpenAI's GPT-4 Turbo and Business Expansion

OpenAI announced GPT-4 Turbo, a more powerful AI model, along with a new option for users to create custom versions of ChatGPT. The company is also reducing the fees for running its software, aiming to stay ahead in the AI arms race. GPT-4 Turbo offers updated knowledge, accepts longer inputs, and supports AI-generated images and text-to-speech. OpenAI's move to allow personalized chatbots and the introduction of the GPT Store represent significant steps towards democratizing AI development and creating new revenue streams. Microsoft CEO Satya Nadella's support underscores the collaborative effort to advance AI technology.

Management Quotes

  • Joelle Pineau, Vice-President of AI Research at Meta:

    "We are hard at work in figuring out how to get these models not just to talk, but actually to reason, to plan . . . to have memory."

  • Brad Lightcap, COO of OpenAI:

    "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... Today’s AI systems are really good at one-off small tasks."

  • Yann LeCun, Chief AI Scientist at Meta:

    "Current AI systems produce one word after the other really without thinking and planning... Because they struggle to deal with complex questions or retain information for a long period, they still make stupid mistakes." "Adding reasoning would mean that an AI model searches over possible answers", “plans the sequence of actions” and builds a “mental model of what the effect of [its] actions are going to be"... It was working on AI “agents” that could, for instance, plan and book each step of a journey.

  • Chris Cox, Chief Product Officer at Meta:

    "[The cameras in Meta’s Ray-Ban glasses] could be used to look at, for instance, a broken coffee machine, and an AI assistant — powered by Llama 3 — would explain to the wearer how to fix it."