Apple hasn’t said much publicly about its plans to join the many companies offering generative AI products. Still, this week it opened a window into its behind-the-scenes work on the kind of system that powers AI chatbots and image generators.
On Monday it launched OpenELM, which it calls a “next-generation open language model.” Language models are massive sets of information that tools like ChatGPT, Gemini, Perplexity, and Dall-E use to respond to prompts you type when you want an AI to prepare an email, write computer code, or create a fantastic image.
So it’s still not the Apple AI product we’ve all been waiting for. Still, it’s a logical step in that direction and potentially hints at the AI capabilities Apple could offer in its upcoming iOS 18 software for iPhone.
The launch of OpenELM comes just weeks before Apple’s WWDC event in early June, where the company traditionally talks about its next wave of software offerings.
Apple did not respond to a request for comment.
But during a quarterly earnings conference call in February, CEO Tim Cook hinted that Apple would reveal its plans for generative AI sometime in 2024. Also around that time, Apple was reported to have shut down its long-running electric car project. to focus on generative AI and the Apple Vision Pro, the wearable that went on sale that same month and that Keynote USA reviewer Scott Stein calls “Apple’s strangest, wildest device.”
It is not yet clear how OpenELM fits into these plans. However, in a research paper published in March, Apple looked at large multimodal language models, or those that can generate a variety of content formats.
While Apple has been holding the fire, most tech giants and several startups have already released one or more generations of next-generation AI products. Adobe, Anthropic, Google, and OpenAI are in a race to release increasingly capable models that not only understand a wider variety of queries but also produce more realistic images and videos. They are even interested in highlighting internal research projects in fields such as video games and music composition to show what they can offer to capture your attention and ensure your potential loyalty as users of their AI products. (For the latest AI news, plus product reviews, explainers, tips, and more, check out our new AI Atlas guide.)
Some conclusions about OpenELM
An article posted on Apple’s Machine Learning Research site notes that the OpenELM version includes everything needed to train and evaluate the language model on publicly available data sets, including training logs, checkpoints, and pre-training configurations. (The ELM part of the name stands for “efficient language model.”)
On LinkedIn, Maxwell Horton, a machine learning researcher at Apple, wrote that OpenELM is part of a new open-source repository called CoreNet, which is available through GitHub. This resource also includes code to convert models into a matrix framework for machine learning research on Apple chips, enabling inference and tuning on Apple devices, as well as vision and language models with training recipes and code publishing for 11 Apple research publications.
The OpenELM document says the goal is to pave the way for future open research efforts.
“The reproducibility and transparency of large language models are crucial to advancing open research, ensuring the reliability of results, and enabling investigations into data and model biases as well as potential risks,” the document adds.
Large language models are measured in what are known as parameters, or the number of variables in a mathematical calculation used to produce an output from a given input. As the number of parameters increases, the model’s ability to make sense of language also increases. Models are trained and tuned by Apple instructions and have 270 million, 450 million, 1.1 billion, and 3 billion parameters.
For comparison, ChatGPT 3.5 has 175 billion parameters. The latest version of Meta’s Llama 3 model has 70 billion parameters.
Reports emerged last July that Apple was working on an AI chatbot called Apple GPT and a large language model called Ajax, but the company has not commented.
“We have a lot of work going on internally, as I’ve mentioned before,” Cook said of generative AI during the February earnings conference call. “Our modus operandi, so to speak, has always been to work and then talk about work, and not get ahead, so we’re going to maintain that as well.”
Keynote USA News
For Latest Apple News. Follow @Keynote USA News on Twitter Or Google News.
