Archive for October 15, 2024

Tuesday, October 15, 2024

iPad mini (7th Generation)

Apple (MacRumors, Hacker News):

With a beloved ultraportable design, the new iPad mini is available in four gorgeous finishes, including a new blue and purple, and features the brilliant 8.3-inch Liquid Retina display. A17 Pro delivers a huge performance boost for even the most demanding tasks, with a faster CPU and GPU, a 2x faster Neural Engine than the previous-generation iPad mini, and support for Apple Intelligence. The versatility and advanced capabilities of the new iPad mini are taken to a whole new level with support for Apple Pencil Pro, opening up entirely new ways to be even more productive and creative. The 12MP wide back camera supports Smart HDR 4 for natural-looking photos with increased dynamic range, and uses machine learning to detect and scan documents right in the Camera app.

The new iPad mini features all-day battery life and brand-new experiences with iPadOS 18. Starting at just $499 with 128GB — double the storage of the previous generation — the new iPad mini delivers incredible value and the full iPad experience in an ultraportable design.

John Gruber:

A17 Pro is the chip from last year’s iPhone 15 Pro models, and, notably, there was no non-“Pro” variant. Still, though: an interesting chip to use for iPad Mini.

I thought the rumor was that was to be a temporary chip because it was much more expensive to manufacture.

Ryan Christoffel:

For Wi-Fi, the previous mini offered Wi-Fi 6 compatibility, but the new mini takes it further by supporting Wi-Fi 6E.

The previous mini supported Bluetooth 5, but the new mini adds the more modern Bluetooth 5.3 spec.

[…]

Both models support USB 3. However, the old mini only offered speeds up to 5 GB/s, whereas the new model doubles that and goes up to 10 GB/s.

M.G. Siegler:

Overall, the update to the iPad mini would seem to be a good one – as tends to be the case when you wait three years between product refreshes.

Federico Viticci:

I’m not even sure that “disappointing” would properly describe this iPad mini update.

Three years for a chip bump and Apple Intelligence, and this thing will likely be replaced in 2027? Cool.

Rui Carmo:

[No] matter how they sugarcoat the A17 Pro, it’s not the upgrade I wanted for my mini 5 in either CPU, display, camera or anything else short of the USB-C port and TouchID (yes, I prefer TouchID).

Given the PR-only prerelease and outrageously spaced out refreshes it’s obvious the mini isn’t a priority for Apple, so I have to figure out if I want to address the fact that the 256GB cellular model is closer to €1000 than I would like or wait another two years to upgrade.

René Fouquet:

Still no Pro Motion display. I tried the last iPad Mini and had to send it back because I couldn’t get used to the low refresh rate.

Steve Troughton-Smith:

iPad mini battery life is pretty miserable as-is, without Stage Manager or ProMotion or Face ID. While I would love to see an M-series iPad mini Pro, with all the bells and whistles, I’m not convinced it can be done to that level with current battery technology.

Adam Overholtzer:

The big problem for me remains the price. The iPad mini isn’t worth $500–600 and I don’t want to pay $500+ for an iPad.

Steve Troughton-Smith:

I would honestly fully support Apple splitting the iPad mini into two separate lines — remove some stuff to make the mini even cheaper than it is today, but have an iPad Pro 8.3-inch (M4) with everything the bigger models have. Give it that 5.1mm OLED design to make it the ultimate notepad/sketchpad

Previously:

Ward Christensen, RIP

Benj Edwards (Hacker News):

Ward Christensen, co-inventor of the computer bulletin board system (BBS), has died at age 78 in Rolling Meadows, Illinois. He was found deceased at his home on Friday after friends requested a wellness check. Christensen, along with Randy Suess, created the first BBS in Chicago in 1978, leading to an important cultural era of digital community-building that presaged much of our online world today.

[…]

Despite creating one of the foundational technologies of the digital age, Christensen maintained a low profile throughout his life, content with his long-standing career at IBM and showing no bitterness or sense of missed opportunity as the Internet age dawned.

[…]

Prior to creating the first BBS, Christensen invented XMODEM, a 1977 file transfer protocol that made much of the later BBS world possible by breaking binary files into packets and ensuring that each packet was safely delivered over sometimes unstable and noisy analog telephone lines. It inspired other file transfer protocols that allowed ad-hoc online file sharing to flourish.

Matt Keeter (Hacker News):

How did I find myself writing a new implementation of a 45-year old protocol?

Previously:

Apple’s Stale Mac Displays

Joe Rossignol:

Apple sells two external displays, including the Pro Display XDR and the Studio Display, but neither has received hardware upgrades in years. In fact, the Pro Display XDR is nearly five years old, having been released all the way back in December 2019.

[…]

In December 2022, Bloomberg’s Mark Gurman said Apple was working on an updated version of the Pro Display XDR with an Apple silicon chip, something the current model lacks. In the Studio Display, an A13 Bionic chip powers features such as Center Stage camera framing, spatial audio, and Siri voice commands. However, there have not been any recent rumors about a new Pro Display XDR, so it’s unclear what Apple’s current plans are.

[…]

There have been on-again, off-again rumors about Apple planning a new 27-inch external display with mini-LED backlighting, which would allow for increased brightness and higher contrast ratio. In April 2023, Apple supply chain analyst Ming-Chi Kuo said the display was slated for mass production in 2024 or early 2025.

This is not surprising, since Apple has historically taken a long time to update its displays. I don’t think the panels necessarily need to be updated. But it’s disappointing because the Studio Display has well documented camera problems and power issues. I had high hopes that, coming from Apple, it would be reliable as a USB hub, but I end up directly connecting as many storage devices as possible to the meager ports on my MacBook Pro.

Also, it would be nice to have a mid-range display in the lineup. You can get an M3 iMac with a built-in 4.5K display for $1,299, yet the Studio Display by itself starts at $1,599. It’s still hard to find good third-party Retina displays. Why can’t Apple sell an iMac-less panel for a reasonable price?

Previously:

Understanding the Limitations of Mathematical Reasoning in Large Language Models

Hartley Charlton (Hacker News):

The study, published on arXiv, outlines Apple’s evaluation of a range of leading language models, including those from OpenAI, Meta, and other prominent developers, to determine how well these models could handle mathematical reasoning tasks. The findings reveal that even slight changes in the phrasing of questions can cause major discrepancies in model performance that can undermine their reliability in scenarios requiring logical consistency.

Apple draws attention to a persistent problem in language models: their reliance on pattern matching rather than genuine logical reasoning. In several tests, the researchers demonstrated that adding irrelevant information to a question—details that should not affect the mathematical outcome—can lead to vastly different answers from the models.

Gary Marcus:

Everyone actively working with AI should read it, or at least this terrific X thread by senior author, Mehrdad Farajtabar, that summarizes what they observed.

[…]

Another manifestation of the lack of sufficiently abstract, formal reasoning in LLMs is the way in which performance often fall apart as problems are made bigger.

[…]

What I argued in 2001, in The Algebraic Mind, still holds: symbol manipulation, in which some knowledge is represented truly abstractly in terms of variables and operations over those variables, much as we see in algebra and traditional computer programming, must be part of the mix. Neurosymbolic AI — combining such machinery with neural networks – is likely a necessary condition for going forward.

Dare Obasanjo:

This is a problem for anyone who belueves they can build autonomous AI agents on this foundation since it means anytime the “agent” sees a pattern it doesn’t recognize, it will fail hilariously or even catastrophically.

Nick Lockwood:

The most surprising part of the news that Apple researchers have discovered that LLMs can’t reason is that anybody who had even a layman’s understanding of LLMs thought they could in the first place.

Pierre Habouzit:

I think that what [LLMs] do is similar to our human so called “intuition”: they recognize “patterns they’ve seen before and intuitively go to the answer that worked then.”

This is an important aspect of how Inthink and a lot of the creative process I have at work is a back and forth between “intuition” and verifying that it sustains a more rigorous model.

[…]

LLMs have a role into an actual form of AI. It just can’t be on its own.

Dave Rahardja:

LLMs can’t do math because they don’t actually understand concepts; they are just really fancy autocomplete engines.

We knew that already, but this paper quantifies it. The math performance is really pretty dismal even with training that tries to optimize for math. The best performance was by OpenAI’s GPT-4o, which scored around 95% for the most basic of grade-school word problems, which means it got 1 in 20 questions wrong, which means it’s not usable for anything in production.

[…]

IMO the biggest problem with LLMs is not that performance is poor, but that there is no way to tell when they get it wrong. The models may make one mistake in a million, but *which output is the wrong one*?

Jason Koebler:

In December, NARA plans to launch a public-facing AI-powered chatbot called “Archie AI,” 404 Media has learned. “The National Archives has big plans for AI,” a NARA spokesperson told 404 Media. It’s going to be essential to how we conduct our work, how we scale our services for Americans who want to be able to access our records from anywhere, anytime, and how we ensure that we are ready to care for the records being created today and in the future.”

Employee chat logs given during the presentation show that National Archives employees are concerned about the idea that AI tools will be used in archiving, a practice that is inherently concerned with accurately recording history.

Previously: