Apple is developing a large array of features that use generative AI, including a new version Siri, that could launch next year, according to Bloomberg's Mark Gurman.
In his latest "Power On" newsletter, Gurman said that despite Apple CEO Tim Cook's claim that Apple has been working on generative AI technology for years, Apple's executives were "caught off guard" by the industry's sudden interest in AI and have been anxiously "scrambling since late last year to make up for lost time."
Apple's senior vice president of machine learning and AI strategy, John Giannandrea, and senior vice president of software engineering Craig Federighi are apparently leading the company's AI efforts. Eddy Cue, Apple's senior vice president of services, is also involved in the push.
Giannandrea is said to be overseeing development of the underlying technologies for a new AI system. Most notably, his team is working on a new, "smarter version of Siri" that is deeply integrated with AI. Gurman says it could be ready as soon as next year.
On the other hand, Federighi is supervising the integration of AI into to the next major version of iOS by introducing features running on Apple's large language model (LLM). Among the new features is a revamped interaction between Siri and the Messages app, enabling users to field complex questions and auto-complete sentences more effectively. Federighi's team is also looking at integrating AI into Xcode to help developers write code more quickly, bringing it in line with services like Microsoft's GitHub Copilot.
Cue is reportedly attempting to add AI to as many Apple apps as possible, including features in Apple Music, such as auto-generated playlists, and productivity apps like Pages and Keynote, where AI could assist with writing or creating slide decks. Apple is also apparently testing the use of generative AI for internal customer service apps within AppleCare.
Apple is purportedly on course to spend $1 billion per year on AI research. Whether Apple should deploy generative AI as a completely on-device experience, a cloud-based setup, or a hybrid approach, is currently a source of debate internally.
Gurman previously reported that Apple is working on a chatbot tool that some engineers have dubbed "Apple GPT" to rival OpenAI's ChatGPT.
Top Rated Comments
Some have discovered that GPT4 was just 8 instances of GPT3 stacked on top of each other. That makes it extremely inefficient for what you are getting out of it. Literally scorching the planet to death just to play with a word machine.
Some have also noticed model decay. This is what happens when AI models scrape data created by AI models. It's similar to what happens when a VHS tape is copied again and again and again. The quality degrades over time.
They go from being trained on high quality data to being trained on lower and lower quality data.
So imagine a future where an AI is surfing an internet full of AI generated garbage, fake articles, fake images, and general enpooification.
Steve Jobs could see some unfinished technology and understand where that would pay off in 5, 10 even 20 years. The mouse and cursor UI tech was just sitting unused at Xerox. Steve's genius wasn't necessarily inventing brand new things from nothing. It was in seeing disparate parts and putting them together into something innovative that others hadn't yet seen.
Jobs' last big move was acquiring Siri and quickly implementing it into the iPhone. When he died, Tim Cook let it sit, like Xerox did with the mouse, unable to see the massive importance it would play in a shift to conversational UI ('https://forums.macrumors.com/threads/the-ai-wars-are-here-whats-apples-response-to-chatgpt-google-bard.2379877/') in a decade.
Apple has the resources to catch up but it was inevitable that sooner or later, the next Steve Jobs was going to outwit Cook like Sam Altman at openAI did. Apple needs a dreamer.