eXtensions - Wednesday 22 February 2023
By Graham K. Rogers
Apple may have to stop selling the Apple Watch in the USA; or maybe not. Recent news and movies from AppleTV and Netflix. Concerns over AI Writing Tools - the potential for good and for risk with a few useful links
It is remiss of me for being so long between these online comments. Mea culpa. By way of exuse, two thesis attempts by students and a long scholarly article came my way this week for editing, while I also wrote a discussion document at work about the need for policy decisions regarding the use of AI-writing tools. See below. . .
The decisions here are important in several ways: patents are being upheld even though it is David v. Goliath; a large company (and Apple is significant here, almost symbolic) is being held to account; whether by oversight or deliberate (see below), Apple will either have to remove the technology or pay a license fee. The last, I suspect is the best route for all, but there could be other side-effects concerning the publicity, both for Apple and AliveCor.
There are still twists and turns, Disotto explains, notably "AliveCor's appeal of the U.S. Patent Trial and Appeal Board's decision that found the asserted patents unpatentable" which may be why Apple used the technology in the first place. If that appeal is upheld, Apple will probably also use its deep pockets to delay any action, but that should include paying a fee if the patent is held to be enforceable.
Note also that the decision also protects Apple considering the number of patents it is granted each week: Patently Apple is a good source for many of and gives a good idea of what types of technology Apple is examining. However, Jack Purcher (Patently Apple) is less convinced that AliveCor has a solid victory here, particularly as Apple is counter-suing in a San Francisco court. The President's non-decision was just one step on a long road.
On 13 March, which is a Wednesday, series 3 of Ted Lasso will appear and it is expected to be the last. A trailer has been released: "If seeing is believing, I believe we have been seen". Apple usually releases its new series and subsequent episodes on a Friday but that means the day can become a bit cluttered with other series running concurrently; and in any case. They are also doing this with a new series, The Big Door Prize, which opens 29 March (Dennis Sellers, AppleWorld Today). Ted Lasso already has such a big following that the day iOS hardly likely to matter. However with a new show, Extrapolations, coming the same week (17 March) that works rather well. This is expected to be the final season of Ted Lasso, but Dennis Sellers reports that "Jason Sudeikis, star, executive producer and co-creator of the hit show, may be looking to continue the show to a fourth season. . . ."
Another new show is Tetris, starring Taron Egerton and Apple has just released a trailer for this. This is based on the story of how rights for the original game - developed in Russia in 1984 by Alexey Pajitnov - were acquired with some interesting and quite funny scenes set in a crumbling Soviet Union. There are comments on the 31 March release and the trailer in an item by Tammy Rogers (iMore). This makes quite a departure for Egerton after his recent performance in Black Bird.
There are a few things coming over on Netflix, particularly some European offerings and I particularly enjoyed the Norwegian movie, Narvik, which tells the story of an early battle in World War 2 which was known as Hitler's first defeat. However, with Allied troops withdrawing quite quickly the Norwegians were conquered again soon after. As a note, Vidkun Quisling who was made Prime Minister lent his name to the term, Quisling, that has become synonymous with "traitor" for his collaboration. Also reported to be on its way is a new show from Ridley Scott. Tammy Rogers (iMore) writes that the expensive endeavor is called Sinking Spring, and the pilot alone is reported to have cost $50 million (Gladiator was $103m).
The MacPro keeps appearing in the news as not-a-real-product from Apple. Apple has said it will come, but there is considerable speculation. My own view is that the landscape has changed and that this is a dinosaur: technology from the past, despite the incredible legacy. The new generations of Macs, particularly the beefed up M2 Mac mini and the Mac Studio (even with its current M1 chips) show how a device with a smaller footprint can outperform much larger machinery.
I do not think Apple needs a Mac Pro, but a lot of people see this as a test, despite Apple silicon putting engineering constraints on the latest devices. It has not been possible for a few years to update the memory in several Macs. The M1 and M2 come with memory pre-installed too. The lack of upgradability is a sticking point for some commentators, but I have not seen any input from potential users.
The Mac Pro is, as Jason Snell (MacWorld) writes, a niche product within a niche category. While Michael Simon (also at MacWorld) thinks that the only way to "save" the Mac Pro is to stick with Intel: an upgraded Intel model is probably preferable to an M2 Ultra-based model, he writes. Despite the power efficiency he suggests that the new Xeon W-3400 or Xeon W-2400 chips that go up to 350Watts would be preferable to some users with the expandability and after-market upgrades. It is a thin argument methinks.
My feeling is that, with the prices of the devices that are here now, and the potential price range of a new Mac Pro, it may be just as economic for users to upgrade to the most recent Mac Studio with maxed out memory that the job needs. There have sometimes been chip- and memory-matching problems in upgrades so the SoC where everything is integrated makes more sense.
Apple however is in a damned-if-you-do and damned-if-you-don't situation. If they produce a Mac Pro within whatever limitations Apple silicon creates, the sales will not be massive (they never are); but if they decide not to go with a new Mac Pro, despite claiming - nay promising - that they would, there will be howls of criticism and derision. The shares will fall. They always do. Apple will carry on. It always does.
While Apple employees seemed to have escaped the mass layoffs that had afflicted several other big tech companies in recent months, it is reported that a few hundred contractors are now being laid off. Dennis Sellers (AppleWorld Today) comments that these are "technically employed by outside agencies who work alongside Apple employees on project" built the contracts are being terminated now: Sellers has more on this. Use of contractors is a way round budget constraints on hiring personnel. The CIA does this and one of its more famous outside hires was Edward Snowden.
Although I have been trying out GPTZero (with some success) to discover AI-created content, it is reported by Tom Williams (Times Higher Education), that the plagiarism detector, Turnitin, has announced an AI detector with "97 per cent accuracy". This is expected to be available from April (my source was Retraction Watch, which is well-worth following for news of unusual and/or unfair practises in journal publication).
I have talked (and written) a fair amount recently about AI writing and detection methods which I think will become increasingly important with regard to academic writing. If a machine is creating your content, you are not. Indeed, the process by which AI learns is to hoover up what has already been written, so it ain't yours, whatever you claim. I have been looking at AI-detection tools, particularly GPTZero. Kyle Wiggers (Tech Crunch) has compared 7 detection tools and compares the results. GPTZero is good, but not perfect. I will now also be looking at CatchGPT.
A comment on this that I saw on Twitter reported that a student had been experimenting with GPTZero and had managed to produce false results from the writing submitted. If you are deliberately trying to catch out software, it could be relatively easy to write using repetitions and bland stock phrases to produce a false negative. I feel that, as ever, the answer (in an academic situation) is the responsibility of the teacher: further checks if the software indicates a problem; and proper supervision throughout the writing process. Not, as the writer of the Tweet suggested, suing the developer for libel if the child is accused of such plagiarism.
Before these AI writing tools became available, I might occasionally find what I suspected was plagiarism in student writing. It is a little easier when a native speaker switches from halting English to a paragraph that has sophisticated grammar and sentence structures. But first I would check: try and find the source on the Internet. Sometimes I would track it down easily and the case would be proven. On other occasions, like the recent class which used a reference text but instead of paraphrasing themselves used an AI site. Although I had suspicions (and some evidence), the main tool for finding out was to ask the students. Questioning is essential.
Bajarin, Tim. " Why Apple Is Not Rushing To Create A More Public Representation Of Its AI Capabilities". Feb 21, 2023. Forbes.
Geigner, Timothy. " Sports Illustrated Sure Looks Like It's Trading Human Journalists for AI". Feb 21st 2023. TechDirt.
MacDaily News (additional comments on Bajarin). " Apple's Artificial Intelligence and Machine Learning are more advanced than many believe." February 21, 2023. MacDaily News.
Graham K. Rogers teaches at the Faculty of Engineering, Mahidol University in Thailand. He wrote in the Bangkok Post, Database supplement on IT subjects. For the last seven years of Database he wrote a column on Apple and Macs. After 3 years writing a column in the Life supplement, he is now no longer associated with the Bangkok Post. He can be followed on Twitter (@extensions_th)
For further information, e-mail to
Back to
eXtensions
Back to
Home Page