Apple announced a significant development at this year’s WWDC: the creation of its own cloud infrastructure named “Private Cloud Compute” for securely handling certain Apple Intelligence requests. As an IT professional working in data center technologies, I have a few questions that remain unanswered even after watching John Gruber’s The Talk Show Live:

1. What CPU is used in each server? I wonder if Apple is utilizing high-end versions of its Apple Silicon chips. It’s worth noting that there was no update to the Mac Studio this year. Is Apple diverting M3 Max or M3 Ultra production to build its Private Cloud Compute data centers (which currently feature 32 Neural Engine cores in the M2 Ultra)?

2. What type of case design is Apple using for the servers? Are they modified versions of the Mac mini, or are they using a rack-mount variant of the Mac Studio?

3. Is Apple deploying data centers only in the United States or across multiple continents? I suspect the latter, for the sake of redundancy and capacity.

I expect that sometime in the future, perhaps at WWDC 2025, Apple may reveal details about the first year of Apple Intelligence in a short video. We’ll have to wait and see.

Microsoft is recalling “Recall” after all, and this makes them look rather bad. This happens on the same week of Apple revealing Apple Intelligence which received a more positive set of reactions.

We are witnessing two different approaches to the challenge of intelligently integrating generative AI prowess to the base operating system. These two events couldn’t be more evocative of how different Apple and Microsoft strategy and culture are. Guess which approach I prefer? I’m excited for Apple Intelligence, but I appreciate the time it will take to make it right.

If Apple Intelligence¹ requires a new generation of hardware to be fully appreciated, I would be surprised and be a little sad. My expectation would be that last year’s generation of hardware should be, at the minimum, enough. Otherwise, it might send a wrong message that even current Apple Silicon isn’t enough to power generative AI-based experience. How could this be, after all, Apple always brag about its silicon superiority. Tim Cook said that their silicon was their current competitive edge… but was he referring to future chips only? I hope not.

“Apple will put a major emphasis on AI at its upcoming WWDC, as the company aims to catch up to competitors in this rapidly evolving space. While Apple’s approach has been more measured, the pressure is on to provide a clear growth narrative, which its AI initiatives may struggle to directly deliver. Unlike Microsoft and NVIDIA, Apple does not have an obvious path to monetize AI, instead focusing on using the technology to enhance its existing products and services.” — ChatGPT summarizing this article from M.G. Siegler.

I’m more than ok that Apple doesn’t try to monetize their AI journey.

Using templates in office work can sometimes be counterproductive, leading to less original thinking, reduced engagement, and fewer creative problem-solving opportunities. Templates can make it easy to fall into a routine of just filling in blanks, resulting in more generic outputs and a checkbox mentality. This reliance on templates can also make it challenging to adapt and innovate when a task doesn’t fit the template. To keep creativity and innovation alive, it’s helpful to use templates as a starting point while encouraging team members to think critically and adapt as needed.

Some of my colleagues are highly dependant on them, I’m not. I’m staying away from them. It’s a creative thinking killer.

The new M4 iPad Pro, with its repositioned FaceID camera, feels like a step backward compared to the 2018 iPad Pro. I’m not sure what I’m doing wrong, but I frequently receive prompts to enter my PIN due to failed FaceID attempts. Swiping up to unlock the device momentarily covers the camera and sensors, disrupting the process. While Apple may have enhanced the FaceTime experience, the unlocking experience has become more frustrating. In my opinion, this is a significant regression.

This is my second narrated blog post on Micro.blog. It seems to me that writing for sharing as a text-only post triggers different strings for me than writing for a narrated text post. It feels strange because this isn’t a podcast episode. There is no introduction, no conclusion. No background music either. Yet, I find this cool because this is going to be available also to my podcast subscribers. Because of that, I call this an audiocast since the textual transcription (the writing) is done before the audio recording. Audiocast: this word doesn’t exist, isn’t it? Remember that you read it, and heard it here first.

Mac computers are fast and have impressive battery life. Mac Studio with the M4 Ultra ship can wait. The iPad is fast too, potentially surpassing the performance of many Mac computers. It is thin also. Really thin. We get it. The iPhone is fast, light, thin and features excellent camera capabilities. Enough of incremental updates. That is why I do not want new Apple hardware at the upcoming WWDC conference, even if it was a home router, a la AirPort.

Instead, I think Apple’s software in general requires attention and improvement.

For instance, the iPadOS user interface could benefit from refinements to enhance the overall user experience. Similarly, the macOS notifications center and the iPhone Control Center may need rethinking to improve their functionality and usability. Additionally, the macOS Settings app could undergo further enhancements to provide users with a more streamlined and intuitive interface. These are examples of user-facing interfaces that could benefit from further enhancements to better meet the needs and expectations of Apple’s customers. Oh, and Siri needs even more works to become the real intelligent assistant we are all waiting for. That’s the story I wish for WWDC this year.

No new hardware, please. Thanks you.

The upcoming Windows 11 AI ‘Recall’ feature, which is designed to take regular screenshots of a user’s PC content to help them find past information, has been called a security disaster by a security researcher. According to the expert, Kevin Beaumont, the Recall feature is essentially an “infostealer” that can be exploited by hackers to access sensitive data, such as user interactions, text messages, passwords, and websites visited. Despite Microsoft’s attempts to improve the security of Recall, the researcher claims the feature will “set cybersecurity back a decade by empowering cybercriminals” as the underlying database can be accessed through malware-infected PCs.

I wonder how differently Apple would fare if it implemented something similar. Would it be better?

Hi everyone. I might be late to the party, but I finally sat down for a few minutes to test Micro.blog’s new feature: narrated blog posts. It is strange because this is my first use and I cannot write a full review because as I’m writing this, my audio recording is still not done, and I don’t know how is the audio attachment and processing will actually work. It might be similar to when I post a new podcast episode. In any case, I think this is a clever addition to an already useful and enjoyable blogging service. Thanks to @Manton for this. Now, should I say, “have a great day and see you for another blog post?” Probably not.

[Update #1]: It seems the play button doesn’t show up. Probably a bug with the Cards theme. Investigation started. [Update #2]: I found out that narrated blog posts come out as podcast episodes on Podcasts. I’m not sure this is what I want.