mpantone

Two-thirds of the time is horrid. Even 90% is useless.

Put it in perspective using an actual real world comparable scenario: a human personal assistant.

Let's say you give your human P.A. three tasks:

1. pick up dry cleaning (via TaskRabbit for the AI assistant),
2. e-mail vendor that their account will be past due tomorrow thus incurring a 1.5% service charge, and
3. book round trip flight on April 17th from Los Angeles to San Jose (SJC in California)

Your AI assistant only correctly accomplished two of the three tasks. Now if it's the dry cleaning, that's maybe not a big deal. But the other two are. And there are plenty of ways the AI assistant can screw up. Maybe they told the vendor they would be fired tomorrow. Maybe the AI assistant quotes a 2.5% service charge. Maybe the AI assistant books you to SJO (San Jose, Costa Rica) instead of SJC.

The problem is you don't get to choose which task the AI assistant fails at.

Now if you had a human personal assistant, you'd fire them for effing up #2 or #3.

Realistically a useful AI assistant (or human assistant) really needs to be about 99.8% accurate. Assistants need to be reliable, accurate, and private. And not just two of those three attributes.

What if your cellular provider didn't deliver 40% of your text messages? Your transit card fails at 40% of fare gates. Your car won't start three days a week? Your credit card fails to authorize a couple times a day?

Hell, what if the Tokyo Metro subway payment system screwed up 0.02% of transactions every day? That's literally thousands of rides. Or if ATMs gave the wrong amount of cash withdrawals that many times. If you had a Pasmo subway transit card that only worked 40% of the time, you'd probably give up and just buy paper tickets from the ticket vending machine.

Apple knows this. An AI-assisted assistant needs to be way better than current Siri. It needs to at least be as good as a really, Really, REALLY good human assistant because going back to clean up someone else's mess (AI or human) takes too much time. And you lose trust in that assistant very quickly.

"Fake it until you make it" is not a credible business plan in the real world. That's something Elizabeth Holmes would do.

Apple cannot afford to put out an AI-assisted Siri that only gets things right two-thirds of the time and promise that it'll get better. We already have way too many LLM-powered AI chatbot assistants that dole out garbage on a regular basis. The world is not going to be any better with Yet Another Lame Assistant.

I never said that there shouldn't be native apps on all of Apple's platforms. But clearly someone at Apple decided to deprioritize native Mac and Apple TV apps. And guess what? We are not going to learn the reasoning for this delay. Who knows? Maybe the software engineer who would be working on these is on sabbatical, maternity leave, whatever.

Most likely they will get around to it at some point. 

Be grateful the classical music service is accessible by web browser.

Note that the same cannot be said for Apple Maps after all these years which will always stunt adoption by third parties.

I don't really think the hardware is the roadblock here. It's the software (as usual). If Joe Consumer plugs in a brand new Philips Hue smartbulb into a dumb socket, the smarthome network needs to be able to configure flawlessly. HomeKit adoption isn't going to grow if the pairing procedure is fussy or inconvenient, even if it's a one-time only thing.

Hell, look at Zoom video conferencing. It took off because it was far less fussy than other video conferencing competitors.

Remember that no one wants to get off the couch, walk to some wall-mounted terminal and fidget through a couple of menus just to dim the lights. That makes no sense. Heck, there's a dimmer switch that controls the two main lights in my living room. I can simply reach over with one hand and adjust.

Everything must go through the phone (or maybe smartwatch) because those are the two devices that people gladly carry on their person. There's no logic trying to wean people from their smartphones. PEOPLE LOVE BEING ATTACHED TO THEIR PHONE.

AI isn't a selling point for smartphone hardware (yet) so it's not like any delay will result in less marketshare.

However irresponsibly inept software implementations will erode customer trust. That's the big issue. TRUST IS EARNED. They can't just slice a hundred bucks of the price of the entry level iPhone and get it back.

Apple Intelligence has to be reliable, trustworthy and private. Yes, all three, not just two of them. Just like a personal assistant. If you had a human assistant that was missing one of those three attributes, you'd probably fire them. I certainly would.

And yeah, I'd rather not have a personal assistant until I could hire someone who checked all the boxes.

"Fake it until you make it" is not a real world business plan. That's something Elizabeth Holmes would do.

 

AppleZulu said:

Well, this is a profoundly odd choice on Apple's part. Why would they go to the trouble of making Apple Classical Music available via a web browser, but not natively on a mac or on AppleTV?

Because web browsers run on lots of things like Windows and Linux PCs.

If they addressed Macs and Apple TVs with native clients, they wouldn't be capturing any more marketshare. That's preaching to the choir.

Remember that the Mac marketshare hovers around ~12% of the total PC market depending which ANALyst report you're reading. That's still a small percentage.

Remember that there's already an Apple Classical Music app for iOS and the vast majority of Mac and Apple TV owners also have an iPhone.

Besides people have been streaming classical radio via web browsers for a very long time. This is a tradition that any classical music enthusiast should be well aware of since it predates widespread smartphone adoption.

This move is about broadening the audience for Apple Classical Music.