Apple is trying to get ‘LLM Siri’ again
There is Apple Intelligence It was a wreck Since the first feature was released last year Big new report From ~ BloombergMark Gurman’s Mark Gurman describes why and how Apple is trying to carve again again. And most of the effort depends on rebuilding Siri from the beginning.
Gurman reported it in the past Apple is working Internally called ‘LLM Siri’ -Ai version of the created AI version of the company’s digital assistant. The previous approach to Apple, which mergates existing Siri and Assistant, did not work. Gurman describes a lot of reasons in detail, but the following is a summary:
The company is now trying to rebuild the approach. Some of them are not intended to work with old Siri to work with AI. According to Gurman, Apple is studying the AI team in Zurich, “a new architecture that will completely build an LLM -based engine.” Gurman reported that the company was doing this work in November last year, and this idea is that the assistant is “an incredible conversation and better conversation in synthesizing information.”
Another part of the solution is Use iPhone Differential Privacy to improve Apple’s synthetic data compares fake education data in the e-mail of IPONE users, but uses the device and sends only the data synthesized for AI education. And one way the company discusses SIRI improvement is to loosen the LLM version on the web and “grab and synthesize data from multiple sources.” By default, Siri as an AI web search tool is one of the companies that Apple approached the partnership of AI search in Safari, unlike Perplexity.
Whatever the results, Giannandrea is not a direct part of the product development, Siri and Robotics projects in the spring. According to Gurman, Apple EXECS talks about putting him on the way to retirement, but he is concerned that some research and engineering people will leave with him. Gurman says he plans to insist on saying that “Siri is a problem for others.”