"Hain't we got all the fools in town on our side? and ain't that a big enough majority in any town?"
-Mark Twain, The Adventures of Huckleberry Finn
Know your market!
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.
Apple Intelligence's Writing Tools Feature Highlighted in Humorous Ads
- Thread starter MacRumors
- Start date
- Sort by reaction score
"Hain't we got all the fools in town on our side? and ain't that a big enough majority in any town?"
-Mark Twain, The Adventures of Huckleberry Finn
Know your market!
Apple is certainly different. They're bad at selling targeted ads, especially here in Norway.
It's not about what they collect, but how they use it.
Probably because Spark and Grammarly are doing everything in the cloud? Whereas Apple Intelligence does pretty much everything directly on the device.
Our Intel Mac Pro is still going strong. I think it’s great that Apple keep releasing new MacOSs for it after all these years. Sure it doesn’t do everything that the M series chips do, for example no iPhone on the Intel Mac, or the AI stuff that Apple has created to run on M series chips. But as you’ve said, Spark & Grammarly work well on it, so for me it’s not a huge loss. Apple is writing MacOS primarily for M series chips, not intel. Apple is moving on. The Intel Macs still do everything they did when we bought them, and more. Why be upset that our Intel Macs are no longer the #1 priority for Apple? We’ve had our Intel Mac Pro for 7 years now. I expect we’ll get another 3 years and then upgrade to a new machine. 10 years for 1 computer is pretty good and what we expected... This may be the last OS for our computers, what a great run! If you envy the M series chips for their new features, maybe you’ll look forward to a new Mac when the time is right. 😉
Yes, it matters how they got there, because the point of doing anything is that the people actually understand what they're doing. People who don't understand anything and copy/paste out of a tool that did all the work for them are useless in a crisis where actual thought is necessary, and cannot provide any useful insight into understanding and improving processes.
Count me in on the debate. The way I see it is… what are you paying the person for?
Is it their speed at typing? Or is it their charisma to close a sale? To wordsmith an ad campaign? Or spin bad numbers on a presentation? Maybe it’s just to proofread a rough draft? Some are skills and some use tools.
When we allow a computer to replace the human, do we need the human anymore? That’s the debate. The endgame scenario may be 1 person leading a company and using said AI tools to do the work that the other humans did before being replaced.
I don’t think it will get that far, but the potential is there (especially small business). It just depends on how much humanity the business requires to do its business.
And why was mocking her daughter's "I had a great teacher" response necessary? I really don't like the humor where somebody has to be mean to someone else.
You quote statistical jargon but may have missed a class or two. Statistically under the Bell Curve, approximately 68% of people fall under the average intelligence classification (+-1 standard deviation from the mean aka high or low average barriers, 85-115 IQ) but that leaves 32% statistically above or below average intelligence, meaning that only 16% of the general population are actually below average intelligence. 37th percentile is in the average range. This was not written with Apple Intelligence 😆
Last edited:
"You're still running Intel?"
"Ohhhh my gawwwwwd!"
... in a dumb blonde bimbo tone. 🤣
So... just use Spark or Grammarly? That's the beauty of freedom. 😇
How do you know it would "work fine"? The new ML-based technology leverages the Neural Engine hardware directly. It's about the future. Why would Apple invest engineering resources into now-replaced hardware?
There are plenty of alternatives to Apple Intelligence that you can use. If you want Apple Intelligence, you need to buy the hardware that it's built for.
Interesting ads. Expecting Apple Intelligence to become better over time.
If you were in my classes you'd have picked up that there was no quantitative measure specified and got some marks and I'd be less disappointed with you for making assumptions about data. I suggest you go read Spielgelhalter's book...
The defacto "average" is the mean of the distribution. A sample can be below the mean thus is below average. A standard deviation either side is irrelevant to the model until it is well-defined. Anyway my point was not really an academic one but if you use less effective than unity tools, assuming that is measurable (which it is) and you are average then you're doing yourself a disservice by likely producing worse crap even faster than you usually do.
Last edited:
I think that's a bit of a hyperbolic example though. We're not talking about intricate skills here, we're talking about common, basic tasks being done more efficiently (and sometimes better). If I have 100 numbers to add up, I'm going to do it in a couple of minutes on a calculator or a computer rather than spend much longer doing it manually with more scope for error. That doesn't mean that I'm going to lose my ability to do it manually in a crisis. Likewise, if I have a task that I can do quicker and potentially better using AI, I'm going to use AI.
I recently wrote a long report and needed to finish it off by providing an executive summary, something that a year ago might have taken me an hour or so. Using Copilot, I generated one in seconds, then spent about 5 minutes checking it and making a couple of tweaks and ended up with a product that was arguably better than what I would have written from scratch, especially as I was having one of those 'thick head days'. There's lots of times I've been struggling to word something so have put my clumsy effort into ChatGPT and have thought 'yes, that's exactly what I was trying to say' at the output. None of those things erode my ability to write something myself should the need arise. Indeed, I'd argue that seeing how AI can re-parse what I've written can be a good learning opportunity.
It's also a great enabler. I manage a team of people across EMEA and despite English being the lingua franca of the company I work for, language skills vary between individuals and sometimes language weaknesses detract from the quality of what they produce. However, Copilot now means that work can be turned into good quality business English, which they have the language skills to review for accuracy even if they were unable to write it themselves in the first place. Plus, as several have mentioned, reviewing what AI produces from their input actually helps their language development. Even in a group of primary English speakers, you may have somebody who has great ideas but lacks the skills to communicate these well, but with AI that's something that can be overcome.
This both outlines and misses the principal risk. Which is that you need to know what you are doing and critically review the output. This really has limited utility when it is being sold as a way to outsource human labour to avoid paying for it. It just changes things. Not necessarily for the best.
I do, it just strikes me as a little absurd. My little intel rubber bands seem to be running without friction
Well, this one I'm grumbling about was born in 2019... it does actually connect with iphone mirroring. Amusingly I also have a Mac Pro 3,1 from 2008 which is still going strong!
AI can't do math.
The struggle of wording things correctly is a struggle because that's part of the learning process, which is literally reconfiguring the neurons in your brain to make you better for the next time. Avoiding the struggle means you're avoiding the learning process. The same goes for developing language skills.
LLM AIs are not particularly good at math. Other kinds of AI, like Apple Math Notes are written specifically for that purpose and can handle it quite well.
Neither are any good for mathematics. Mathematics has determinism and is rule driven. LLMs are stochastic garbage generators.
Math Notes is terrible as well. It's basically a trick handwriting parser that throws everything into an absolutely dire rule engine and then craps out a numeric answer while trying to mimic your handwriting. Ask it what the square root of 320 is (hint it's not 17.88854382)
We have literally had CAS engines for at least 40 years that are better. Hell even microsoft's equation writing recognition in word is better.