Not using NVIDIA? Apple's large models run on servers built with "in-house chips"

At the highly anticipated WWDC conference last night, Apple played a clever pun: the term "Artificial Intelligence (AI)" was replaced with "Apple Intelligence".

Although the market was not entirely impressed with Apple, and even found the AI capabilities demonstrated somewhat lacking (closing market value dropped by three Googles), Apple still showcased a technological highlight at the WWDC conference: privacy protection.

Apple emphasized at the conference that its new Apple Intelligence system will use a new technology called "private cloud computing" to ensure that any data processed on its cloud servers is protected in a transparent and verifiable manner. Moreover, these data centers do not rely on NVIDIA chips, but exclusively use Apple's own chips.

According to Craig Federighi, Apple's Senior Vice President of Software Engineering, at the WWDC conference, part of Apple's artificial intelligence system is implemented through edge computing.

Federighi stated:

Many of Apple's generative AI models can run entirely on devices powered by A17+ or M series chips, eliminating the risk of sending personal data to remote servers.

If more powerful, cloud-dependent large models are needed (such as OpenAI's GPT large model), Apple will use servers specifically driven by its own chips. Federighi emphasized that these servers are equipped with secure tools written in the Swift language, and Apple's AI "only sends the relevant data needed to complete the task" to these servers, without granting full access to contextual information on the device.

Furthermore, Apple stated that this minimal data will not be stored for future server access or used for further training of Apple's server-based models.

Federighi said:

Your data will never be stored or accessed by Apple, it is only used specifically to fulfill your requests.

Most boldly, Apple emphasized that the integrity of privacy protection can be easily verified by outsiders. Because the server code used in private cloud computing is open source, third-party experts can inspect the code running on these servers at any time to verify Apple's privacy commitments.

However, despite this, there are still skeptics.

Billionaire Musk, who detests OpenAI, tweeted after the WWDC conference ended that if Apple integrates OpenAI's large models, he will ban Apple devices in his companies.

Musk also sarcastically remarked:

Apple lacks the intelligence to create its own artificial intelligence, yet has the ability to ensure that OpenAI can protect your safety and privacy, which is clearly absurd!

Once you hand over your data to OpenAI, Apple has no idea what actually happened. They are selling you down the river