Link: OpenAI plans to release its next big AI model by December

OpenAI is set to launch a new model named Orion in December, as reported by The Verge. Unlike previous models, Orion will not initially be available on ChatGPT but will first be accessible to select companies to integrate into their products.

Microsoft engineers are reportedly preparing to support Orion on Azure starting November. Orion, considered an upgrade to GPT-4, may not officially be labeled GPT-5.

OpenAI's ambition with Orion extends to enhancing its line of large language models (LLMs) toward developing artificial general intelligence (AGI). Orion is described as potentially 100 times more powerful than GPT-4.

In preparation, OpenAI used another model, code-named Strawberry, to feed synthetic data for Orion’s training. A celebration marked the completion of Orion’s training phase in September.

OpenAI CEO Sam Altman hinted at Orion’s release in a cryptic social media post about the "winter constellations rising soon." ChatGPT links Altman's hint directly to the constellation Orion, visible from November to February.

The launch of Orion comes after OpenAI secured a significant $6.6 billion funding round, amidst major changes including transitioning to a for-profit organization and high-profile staff departures #

--

Yoooo, this is a quick note on a link that made me go, WTF? Find all past links here.