tht

Don't expect productivity, multitasking miracles here.

• As others have alluded, you don't really display multiple apps at the same time on an 11" display. It's a little bit better on a 13" display, but not great. Bigger is better for display multiple apps at the same time. Definitely will hopefully be better on a large external display, but at 11", you won't be displaying a lot of apps simultaneously.
• Having background apps continue running is just about 80% of the ask for iPadOS, imo. Like, you are downloading something large over the Internet or copying something large over USB, those types of tasks should be allowed to run until completion at least, whether it takes 1 minute or 100 minutes.
• Backgrounded apps should not be killed. Theoretically, Apple and the developer will save the state of the app, and if it is removed from memory, it will be brought back to exactly the same state when foregrounded. Too many developers ignore this, and Apple hasn't been able to force them to update properly. It's been like 6 years. Developers aren't going to change now. So, either Apple forces the proper saving of an app's state, or solves it in a way that doesn't require developer updates.
• Stage Manager is not a good multitasking UI. Just add a multitasking app icon in the dock. Upon a tap, a pop-up list of apps can be displays, ordered in terms of last touched and or in-memory. App windows would be a submenu in the pop-up list. I suppose you can set a tap on the multitasking app icon to bring up the task manager and other things too. Stage Manager might be better if it allowed unlimited number of open windows, but the current four windows doesn't improve upon Split View, Slide Over much.
• Get rid of gestures as a UI design. Design the UI for direct manipulation. I'd get rid of long press as a UI input even. Add a meta button in a hot corner. You will need to press on that with one finger and tap on the target get the target's pop-up menu. Two finger tap can do the same thing, or a double tap to get a list of the target's menu items to pop up. When dragging and dropping, there needs to be at minimum a UI target or a delay to execute the action. Like a long press on a URL opens up a preview and a slide up of the preview creates a Slide Over app or a Split View app. That slide has to go to a specific target, like the top of the display, and maybe have a delay timer. iPadOS generally is direct manipulation, but there are some frustrating exceptions.
• Terminal.app and Xcode.app needs to be on iPadOS.

Well, it's the usual. It takes failures at a few levels for situations like this to arise.

Cook never should have structured the org such that AI and software are separate organizations. AI, ML and LLM algorithms seem pretty intricately tied to software. If there needed to be a Siri service owner, it should have been Product Marketing (Joswiak)? Then, Giannandrea was not good at his job. This is me looking from the outside. Maybe he was a good leader, but his org hasn't been very productive. I don't know who's responsible, but my experience with the software keyboard has not been good. It's been passable? At times, its key tapping performance and word prediction performance has regressed. That's an ML task that was under Federighi or Giannandrea? Either way, it's now an ML based keyboard that doesn't seem any better than the past.

Then, the overarching umbrella here is LLMs and chatbots. Everyone is desperate to create AI features. Most of them are of dubious benefit, but everyone has to be in the race. That's basically not Apple's MO, where they try to create beneficial features at their own pace. The company is in conflict with itself with AI. Not a good place to be.

Mike Wuerthele said:

tht said:

It's quite safe to say that buyers of the Mac Studios with Ultra SoCs are not looking at Apple mediarati nor Youtubers for buying advice. They know their workflows and what benefits them.

The latter part about knowing their workflows is true, but as per the first half of this sentence, we already know who reads our Pro hardware reviews, based on feedback and suggestions.

There's a reason why we do it differently than everybody else.

The stoichometry and all the fluid flow calcs are massively multi-threaded, but I don't think the division of labor, per se, is equitable. Without giving anything away, networking overhead, even on Thunderbolt, plus some other operational considerations made clustering them up across multiple machines unrealistic.

I'm confused. You tried running the stoichometry and fluid flow simulations on a mini-clustor of Mac Studios, networked through Thunderbolt?

If they were run on one machine, I would expect a bigger speedup between a 16 CPU core M4 Max and a 32 core M3 Ultra. The chemistry sim had no speedup while the fluid sims had poor speedups. Not expecting a doubling in performance, but 30% to 50% type numbers. So, I have to wonder what's going here. Rosetta limits the number of cores for apps using it? Don't see much reason for a stoichometry or chemistry sim to have no speedup here. Some of these apps expect a certain type of cache, memory and networking architecture, and run bad on hardware they are not optimized for.

I can understand the Xcode compile and image test not scaling as much, so probably not a big deal there. Those could be limited by storage performance.

My experience is that if we spending 10s to 100s of thousands of dollars on workstations or clusters, the vendor will provide a test system.

The headline is a bit too dramatic. 🤪

Work probably began on the M6 in 2023, not now, including logic boards, modem integration, etc. What they are probably doing now is doing some test runs of M6 on TSMC N2 and N3P. Not even pilot production, just testing how to fab M6, getting some test SoCs in very small quantities. Then, testing them out in the various form factors the M6 is going into. The M6 chip design is probably 6 months away from being "frozen" for mass production, but it has likely been a 3 year process of intensive work, if not more.

brianus said:

tht said:

It's probably just a coincidence. There is no such thing as a step function transition of all the product lines. Every product line has their own schedule, and they are updated with the most appropriate hardware per marketing and component availability.

Well I'm not talking about all the lines, just two of them. If they have a glut of M3 chips because they'd been making them for the MacBook Air (and nothing else, for at least 6 months), then suddenly those chips will be available for the iPad Air (and not needed for anything else) once the MBA gets updated to M4. I don't know how many they made for the MBA, but it's their most popular Mac..

Apple, and most OEMs, do not have component inventory anymore. They have channel inventory where they send final product units out to retail sales channels, but they don't buy components and store them at their own warehouse anymore. Apple knows how many M3 or A17 Pro SoCs they need, contracts TSMC to fab the units, they likely adjust on a month to month basis.

IOW, there aren't any SoCs sitting in the warehouse to be used in other products. Apple doesn't even have a warehouse. They are stored at TSMC if TSMC chooses to do it that way, and it's pretty certain they aren't making more than they are contracted to. When Apple knows when a product is transitioning, and they know at least 6 months ahead of time, they tell TSMC to stop and only produce enough to support the final sales. The retail channels will slowly sell their inventory just like they are doing now with MBAs.

If the M3 is going in to the iPad Air, Apple will contract TSMC to fab x millions more. Same answer for A17 Pros.

I meant component availability in terms of the component being available to fab, not available in a warehouse.