That would be nice, but I’m not confident.
That’s interesting - I stand corrected.
Has Apple ever done that before - introducing a major feature in Beta status on a non-beta operating system release?
I want to say they have released new features, in beta, to existing programs in the past. But I can’t name one right now. In this week’s show, David and Stephen said they think Siri AI could be in beta for a year +/-.
And I agree.
Better that, IMO, than having Apple lose interest and forget Siri to death. Again.
Or maybe it fails entirely and they just give up on it
Remember this is on the heels this is on the heels of a prayerful start were they advertised serie capabilities that were not yet reality
This is not some minor feature - It is the flagship feature of both iOS 27 and Golden Gate. How do you introduce a major operating system upgrade with such fundamental flaws in both security and operation of the one feature you have advertised so aggressively?
This all seems so incredibly out of character for Apple.
WWDC 2024 surprised me. I still don’t understand how senior management was caught flat footed. But that’s the past. Apple has accepted that it needs help developing a digital assistant and is working with Google. And Google is working with a few billion users which should help them sort out their assistant.
In years past I read that as many as 4 million people participated in Apple beta programs. If that is correct, that would be something like 0.3% of all iPhone users.
It is entirely too soon to say that Siri 2.0 is a success or failure. IMO. I don’t know if any of us even know to use Siri AI yet, but what I’ve seen up to now is encouraging.
I plan to enjoy the next few months putting Siri through it’s paces.
Have you experienced what others have observed where it reports that it completed a task but actually has not done the task?
I am usually on the side very enthusiastically encouraging AI integration with software. My adoption of both off-shelf AI and custom Ai software is quite notable. Despite that, the description of the current status of Apple AI - both by Apple and by users - really concerns me. It is one thing to offer AI features and quite another to integrate it so tightly with software that it is so hard to avoid. And it is incalculably worse to invoke AI to this degree and with so much advertising fanfare when it remains a beta product.
What really surprises me about the iOS 27 issue is that problems with misleading advertising over the new Siri were identified over two years ago - with a $250 Million settlement due to false advertising. Now they do it again - before the first settlement payout is even complete?
It definitely feels like a beta. Like others, I have mixed experiences.
Some have been awe-inspiring. e.g. I asked it about the meaning of some obscure abbreviations in a niche app that I was reading about (not yet installed) and it gave me a very neat list with exactly what I needed (and it was accurate) and with a specific disclaimer (some were vaguely medically related). It occurred to me it might be useful to keep that list, so a couple of minutes later I asked it to make a new note in Bear with the list from the last answer, and it did so perfectly (including nicely formatted markdown). Fantastic and unimaginable from the last Siri.
Then I asked it to check off my 9pm medications in the health app and got a “I can’t do that”. Really frustrating when your hands are full. Annoyed, I asked it “why can’t you do that” and got a precise reply saying that I needed to do some set up first (give siri access to the health app and be very precise with instructions to that app - use keywords) which is probably reasonable.
It’s a huge jump ahead. So far it’s more benefit than cost.
“With Apple’s 2027 software releases, users can now access new Apple Intelligence capabilities across apps. Siri AI begins rolling out today in beta in English, and will expand to French, Japanese, Korean, Portuguese, and Spanish next month.”
The key word in this statement is beta. Siri AI is pre-release software. It wasn’t included in any OS 27, I had to request it. Nothing misleading about that.
If Siri AI fails me, I’ll use Gemini. My guess is on-device AI will be limited, for the foreseeable future, to consumers, hobbyists, etc. and/or businesses running on-premises servers.
Apple has plenty of time to get it right this time. IMO
Two points I have noticed, it seems really slow and that little spinny thing goes on for ages
Also the app seems unfinished. It’s very bare and just a load of white with text on it. I’m sure it’s only first iteration and will improve with time.
While I wouldn’t go so far as to say mistrust everything, hasn’t it always been academic rigour to check sources? To not accept things without checking for yourself as far as possible?
Of course, and I do. I was specifically responding to the idea that we should “not trust anything we read, hear, or see - in person or online [emphasis added].”
While we must not accept at face value everything we read, hear, or see, that is not the same as trusting nothing. No relationship, from marriage to social institutions, can function without some level of trust. I simply refuse to live my life paranoid, mistrusting everyone and everything, and fearful of the end of the world. Others may live their lives with a cloud of suspicion and doom hanging over them wherever they go; I will not. ![]()
In theory that has been the case.
In reality - in the days of paper libraries it was impossible to do that except perhaps if you were a full-time dissertation student.
Even currently it is hard to do given paywalls.
I’ve had a low opinion of Siri, but that is no longer the case. My favorite feature so far is that it can work with on-screen graphics. This allows me to easily convert handwritten notes taken on my e-ink devices into Apple Notes or Numbers spreadsheets. My handwritten notes sync to my Macs and iPads, and once opened are visible to Siri for conversion. For instance, Siri transformed my handwritten list of subscriptions with prices and renewal dates into a Numbers spreadsheet.
Well, the principle of not trusting images you see on the Internet has been solid advice for over 2,000 years.
As Virgil wrote: Timeō Danaōs et dōna ferentēs – “I fear the Geeks even if they bear Gifs.”
That made me laugh!
20 characters
Amazing things it has done for me:
- surfacing a friends new address in an old text, updating their contact, and starting directions to their home
- I knew I had a bunch of emails about something that I had to set up, but I only knew the organization. I asked Siri to tell me what I had to do (interview a candidate). It summarized what I was supposed to do, the deadline, and provided a brief bio about the candidate. He then drafted an email (when I asked it to) to set up a time to meet with the candidate.
The bad
- it now seems to take a minute to navigate to Bluetooth on my Apple Watch
- very inconsistent about what it seems to be able to do. Old Siri could log my daily medications; new Siri only seems to manage this intermittently
Overall, I’m pretty excited about the work possibilities of this, but I wish it wouldn’t make other things really sluggish.
From what I’ve read old Siri was a bunch of hard coded rules and templates, etc. Siri AI has to “think” about what you have asked it to do?
FWIW I use the Health app to remind me to take my AM & PM meds.
Here is an “amazing” example. I had been corresponding with staff, during which an individual’s phone number and email were provided in different emails. This individual was not in my contacts. Rather than enter the information manually, I decided to test if Siri could create a new contact from multiple emails with the correct information. I asked Siri to locate the email address and the phone number for this individual and to create a new contact card, which it did quickly and perfectly. That saved me a lot of clicking and time. That is amazing.
