"Apple GPT" is Built on the Secret Large Language Model Ajax

But, does this even matter?

Bloomberg reported today that a team of Apple engineers has created a large language model (LLM) enabled chatbot called Apple GPT. While Apple employs machine learning in many product features, the company may be unsure about how to use its new LLM technology.

The iPhone maker has built its own framework to create large language models — the AI-based systems at the heart of new offerings like ChatGPT and Google’s Bard — according to people with knowledge of the efforts. With that foundation, known as “Ajax,” Apple also has created a chatbot service that some engineers call “Apple GPT.”

In recent months, the AI push has become a major effort for Apple, with several teams collaborating on the project, said the people, who asked not to be identified because the matter is private. The work includes trying to address potential privacy concerns related to the technology.

….

Ajax was first created last year to unify machine learning development at Apple, according to the people familiar with the effort.

Missing the Generative AI Moment

Apple does its own thing. It was not the first smartphone, but it has been the most recongizeable device in the market for more than a decade. The same is true for the smart watch and wireless earbuds. When everyone was talking about smart speakers, the company waited, then came to market with an expensive and unpopular device that was eventually discontinued in favor of a more affordable version.

While the world was swooning over genertive AI, Apple doubled down on its virtual reality bet and announced the Vision Pro VR headset will be available next year. It was not surprising they announced the headset they have been working on for several years. They need some lead time for developers to build products for the new device. However, they also went through an entire developer conference without mentioning generative AI. There were the obligatory mentions of machine learning-based features in Apple devices, but the company had nothing to say about generative AI.

This led to the Synthedia Headline “Apple's Reality Distortion Field - Go Long on VR. Ignore Generative AI.” Bloomberg noted that Apple was not exactly enthusiastic about embracing generative AI.

Publicly, Chief Executive Officer Tim Cook has been circumspect about the flood of new AI services hitting the market. Though the technology has potential, there are still a “number of issues that need to be sorted,” he said during a conference call in May. Apple will be adding AI to more of its products, he said, but on a “very thoughtful basis.”

In an interview with Good Morning America, meanwhile, Cook said he uses ChatGPT and that it’s something that the company is “looking at closely.”

Behind the scenes, Apple has grown concerned about missing a potentially paramount shift in how devices operate. Generative AI promises to transform how people interact with phones, computers and other technology.

Knowing and Doing Engines

Two of Siri’s co-founders joined me on the Voicebot Podcast earlier this year. Before the offer from Steve Jobs to acquire Siri and long before its debut in the iPhone 4s, the team initially struggled to get people to understand the voice assitant’s value. Dag Kittlaus, the CEO of Siri, said that changed when they explained that the voice assistant was a “do engine.” It could do things for you; save you time; make your life easier.

Adam Cheyer, Siri’s co-founder and VP of engineering, says that virtual assistants have two general tasks they can execute for users. He characterizes them as “knowing” and “doing.” ChatGPT checks the “knowing” box but he suggests will not soon be competent at a whole lot of “doing.” Assistants such as Siri still focus on “doing” but an LLM as competent as GPT-3.5 could fill the gap and fulfill his original virtual assistant vision.

However, that is unlikely. Apple is not in the “knowing” business. It is in the “doing” business. All of its products are about helping users do things. It is not likely to come along and perfect the “knowing” chat assistant paradigm. We have several very good examples already. And, ChatGPT is widely used on iPhones already. According to Bloomberg:

The chatbot app was created as an experiment at the end of last year by a tiny engineering team. Its rollout within Apple was initially halted over security concerns about generative AI, but has since been extended to more employees. Still, the system requires special approval for access. There’s also a significant caveat: Any output from it can’t be used to develop features bound for customers.

Even so, Apple employees are using it to assist with product prototyping. It also summarizes text and answers questions based on data it has been trained with.

…

While the company doesn’t yet have a concrete plan, people familiar with the work believe Apple is aiming to make a significant AI-related announcement next year.

Don’t count on it. Apple is variously predicted to be rolling out televisions and new cars. It also was going to introduce its new VR headset long ago and turn Siri into a juggernaut the would far outpace Alexa and Google Assistant. This didn’t happen. The rumors almost never come true.

Generative AI functionality is now widely available. Apple, like everyone else, will implement some generative AI features in its devices and use the tools internally. The likelihood that it will do some more significant than its tech giant peers in the next couple of years is very low. That is not how Apple works. It is slow and methodical. It does best in segments that have largely been ignored or abandoned as opposed to those with already high growth and innovation rates.