Skip to content

Apple Breaks Silence: Open Sources Ajax GPT Training Code, Signals Shift in AI Strategy

In a groundbreaking move, Apple has open-sourced the training code for Ajax GPT, an AI model believed to surpass GPT-3.5 with over 200 billion parameters. This challenges Apple's traditional closed-source approach and could foster collaboration within the AI research community.

Apple, renowned for its secrecy, has disrupted the AI landscape by open-sourcing the training code for Ajax GPT, a formidable AI model. This transformative shift from closed-source to open-source could have far-reaching consequences for AI development and collaboration.

Apple's Ajax GPT, trained on over 200 billion parameters, is poised to outshine GPT-3.5, marking a significant advancement in AI capabilities.

In a surprising turn, Apple uploaded the code for AXLearn, its internal software used to train Ajax GPT, on GitHub. This open-sourcing move allows the public to harness pre-built tools for training large language models, breaking away from Apple's traditional secrecy.

Apple's reputation for closed-source technologies is well-known. This shift in strategy represents a significant departure from its conventional approach, one that has been criticized for reaping the benefits of big tech research while keeping its own developments hidden.

Apple's decision to open-source AXLearn could foster collaboration and innovation within the AI research community. This move aligns with a broader trend of openness in AI development.

With substantial daily investments in AI development, Apple is determined to compete vigorously in the AI race. Open-sourcing its training software signals its commitment to staying at the forefront of AI innovation.

Apple's head of AI, John Giannandrea, and lead of the conversational AI team, Ruoming Pang, both hail from Google. They have brought Google's culture of research freedom and innovation to Apple, fostering talent growth and recruitment.

Apple has recruited talent from Google and Meta's AI platform teams, indicating a shift towards an open research community approach.

Apple's recent AI research hints at the development of software capable of generating images, videos, and 3D scenes—a sign of progress toward a multimodal AI approach.

While Apple typically integrates new software into its devices, incorporating a 200-billion parameter LLM into an iPhone presents storage and computing challenges. The exact integration plans for Ajax GPT remain undisclosed.

Apple's bold step to open-source Ajax GPT's training code marks a turning point in its AI strategy. The move not only challenges its closed-source tradition but also signifies its determination to excel in the dynamic AI landscape. As the AI community explores the possibilities unlocked by this shift, Apple's role in shaping the future of AI development takes on a new dimension.