Apple Smart Keyboard Folio for iPad Pro 11-inch (4th, 3rd and 2nd Generation) and iPad Air (5th and 4th Generation) - US English https://www.amazon.com/Apple-Smart-Keyboard-11-inch-iPad-Pro/dp/B08635GZ8H
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iPad Pro The rise of ChatGPT has made the iPad useless
- Thread starter Retskrad
- Start date
- Sort by reaction score
Why does this have anything to do with the iPad? You've been able to use iPads with keyboards for years, possibly even since day one. Or you can dictate. Or scribble with the Pencil. Or use the on-screen keyboard. If this makes an iPad useless to you then I suspect you probably never really needed one in the first place.
I've never heard of anyone describe it as "hallucinating" but that's really hitting the nail on the head. I've been impressed with when ChatGPT is helpful but it has completely made up answers and given me sample code that would not compile several times.
iPad has never been ideal for productivity/content creation. It is mainly for content consumption. Not sure how ChatGPT changes that at all.
Counterpoint: there are more iPads in my junior-year methodology course than ever before. Students are making more efficient use of iPads to collect, organize, and access the unimaginable amount of information coming at them.
Many of my faculty colleagues, on the other hand, are really nervous that ChatGPT will make their roles obsolete. My take on that is they aren’t terribly good at or creative in their assessments. More critical thinking and specific case study application.
Many of my faculty colleagues, on the other hand, are really nervous that ChatGPT will make their roles obsolete. My take on that is they aren’t terribly good at or creative in their assessments. More critical thinking and specific case study application.
I’ve been genuinely surprised at how frequently answers contain information that’s factually incorrect even when accurate information is readily accessible via long-standing resources like textbooks or Wikipedia.
I think looking at a piece of new technology, especially in its infancy, and saying it is going to cause wholesale change to something else that is well established is a fool's errand. Sometimes it does, sometimes it doesn't. The best most accurate predictions... are made after the fact after it has happened.
For example, where is our paperless office? Computers and electronic mail promised us a paperless office. That was all anyone talked about for a while back in the day. Are some offices today essentially paperless? Of course. But, spoiler alert, we're printing more now, by an order of magnitude, than we EVER did before computers, all because of... computers.
iPad is literally for content creators, artists, musicians, and now AI tools.
I don’t envy future teachers having to face or question whether an essay was written by a student or their app.
Imagine claiming that a $40 billion a year product is useless.
Same here. I even thought that it must be referring to a third party library and not a built in language feature but it said it wasn't when I pushed back.
Yep. I have got weirdest of responses when I asked it to convert pythons code to support Apple GPU. Gave me functions and arguments Opencv libraries don’t support, and then CGPT corrected after few iterations.
It does well with front end react frameworks.
It’s still much more advanced than github copilot, which I use a lot.
Funnily enough, "hallucinating" is the accepted term what happens when an AI "invents" new information. I wish I could claim that I came up with it lol
It's best to think of chatGPT and its friends like a super-advanced version of autocomplete on your iPhone. Several years ago, there was a brief fad where people would open up notes on their iPad and type an initial word or phrase - then just tap the autocomplete suggestions repeatedly to see what it would "create". The results were often nonsensical and occasionally funny. That's chatGPT, but with a much larger dataset to work from (and some fancier algebra in the backend).
Here's a good article on the whole thing that makes a great analogy between the information in chatGPT, and blurry JPEGs.
ChatGPT Is a Blurry JPEG of the Web
OpenAI’s chatbot offers paraphrases, whereas Google offers quotes. Which do we prefer?
Last edited:
I won't call iPad useless, because many people have different uses for it. But in my personal opinion, ChatGPT didn't make iPad useless, the lack of running a proper file system and proper multitasking did.
This isn't a "future teachers" problem. It's actually happening today. So far though it hasn't been too hard to catch them thanks to the propensity of AIs like chatGPT to generate incorrect information in their output. Many of the students doing this don't catch the AI's errors because they don't do the work to know/understand the subject well enough in the first place. When the instructor grades it the incorrect info typically stands out.
A big issue is the extra workload on the instructors if they want to pursue an academic integrity case against the student. Many of them just feel it's not worth the time and hassle. Some will let the student squeak by, some will fail the student, in which case they'll just see that student in their class the following year (or if they're lucky, the student will take the same class with a different instructor, passing the buck), and a small number will call the student out ... depends on the academic institution and what their policies and procedures are around academic integrity.
The funny thing is the guy who wrote the article doesn’t really understand AI. ChatGPT isn’t an algorithm, and his understanding is pretty flawed.Funnily enough, "hallucinating" is the accepted term what happens when an AI "invents" new information. I wish I could claim that I came up with it lol
It's best to think of chatGPT and its friends like a super-advanced version of autocomplete on your iPhone. Several years ago, there was a brief fad where people would open up notes on their iPad and type an initial word or phrase - then just tap the autocomplete suggestions repeatedly to see what it would "create". The results were often nonsensical and occasionally funny. That's chatGPT, but with a much larger dataset to work from (and some fancier algebra in the backend).
Here's a good article on the whole thing that makes a great analogy between the information in chatGPT, and blurry JPEGs. It's a very good explanation of what's happening behind the scenes with these systems, and why they can't help but "hallucinate".
ChatGPT Is a Blurry JPEG of the Web
OpenAI’s chatbot offers paraphrases, whereas Google offers quotes. Which do we prefer?www.newyorker.com
People on the internet use such strong language to make a point or draw attention (sometimes both).
“Useless”.
The iPad has plenty of uses for millions of people that are unrelated to ChatGPT. By definition, that means the iPad still has use.
Perhaps you’re a Drama major. Definitely not Communications.
“Useless”.
The iPad has plenty of uses for millions of people that are unrelated to ChatGPT. By definition, that means the iPad still has use.
Perhaps you’re a Drama major. Definitely not Communications.
It's an analogy and not meant to be interpreted literally. It's for the layperson that doesn't understand linear algebra, tensors, vector spaces, training models, etc. The point wasn't to explain the intricate inner workings of ChatGPT & similar models, it was to give the reader something they could relate it to that they were familiar with.
I mean it'd be *great* if the average person *did* understand tensor algebra and all the rest. Because then they likely wouldn't be running around preaching about how wonderful and amazing and magical AI is (as an all-purpose solution to everything), and would understand what it's good at and what it isn't.
But since we're not there yet, education-wise, analogies will have to do.
(While we're at it, I'd also like to see tech reviewers adopt a basic understanding of thermodynamics - it would help them understand why they can't have all the processing power, battery life and minimal heat we want in a wafer-thin device ... lol)
Last edited:
This is just a clickbait thread title. Nothing else. Thanks for the joke though.
It is a terrible analogy. First of all lossy jpeg or Xerox using compression and copying doesn’t link or understand information. It’s a copy, and fills the noise with interpolation.
ChatGPT doesn’t copy or quote verbatim with the information which was fed in training. In fact ChatGPT can apply complicated mathematical concepts to other fields. I can ask chat GPT, how can I apply this mathematical concept in computer vision. It does come up with good idea. Ask it to code, it will give good stuff using opencv. Now request the same in Pytorch3d, it’s starts guessing based on what’s been learned, and comes with non existent libraries.
Yes it has its flaws, but to even compare to lossy jpeg shows clear lack of basic logic or understanding. Sure it uses some statistical similarities, but that’s like saying Ford Pinto, Yogo are same as Tesla or any other car.
For some reason, it does really well with optical flow capture of videos, depth maps and all complicated algorithms. Ask it to give a simple code to remove background using Apple/AMD GPU it comes up short.
I will happily pay for chatgpt to not deal with google search or others, if it keeps improving with more relevant training.
My two kids use my wife and my iPads to play a few games, chat with their friends and watch Netflix.
Everyone’s use is different - don’t try to say they’re useless.
Everyone’s use is different - don’t try to say they’re useless.