Apple's Big AI Push Implied By Startup Acquisitions and Job Postings

[Tagbert]

Siri understands what I'm saying (I have "Always Show Speech" enabled, so I can see what it thinks I said), but it's what it does with the info that's a problem. In the last few days:

1) I said "remind me that Ronnie will be away when I get to work". It added a reminder that simply said "Ronnie" and didn't set my workplace as a location.

2) I asked Siri to play a song by a band that I frequently listen to. It gave me a cover version of said song by some obscure band. Why isn't the OS smart enough to see that I listen to the band regularly and give me the version of the song that I most likely want to hear.

3) I asked for directions to a business that I was near (a road was closed and I had to re-route - not that familiar with the area so I asked for help), and instead Siri gave me directions to a town that shared the same name as the business. The business shows up in Maps, so why is that not the first choice instead of a town that's seven hours away?

4) On Apple TV, I asked it to "find the movie Maestro on Netflix" and it took me to the preplay screen for the movie "My Octopus Teacher". Not even close!

It's pretty good at setting a timer though.

that’s why running the requests through an LLM-style AI would help Siri understand your request and do a better job connecting that to the right service.

 

[Tagbert]

I hope that Apple is not putting all its money in trying to make Siri work... but putting some of it in nice photo editing gimmicks... like the ones found on Android devices... deleting background people, intelligent cropping, etc...

Likely both a better Siri and continue to use AI to enhance photos.

 

[Tagbert]

Monetary return on Alexa is hard to gauge. How many people tell Alexa to order XYZ every day? Each of those sales is monetary return. Plus Amazon uses all the data it collects to more effectively market to its customers. As far as the future goes, Amazon doesn't have a platform/OS like Microsoft, Google, and Apple. I don't think they view Alexa as a full-blown assistant/AI. It's mostly a shopping/marketing tool and I think it probably pays off for them quite nicely.

Amazon reports that Alexa and Echos lost $10B last year. It doesn’t appear that very many people ask Alexa to buy something.

 

[svish]

Hoping to see a much improved Siri in iOS 18

 

[BlueCreek]

Apple is fine with being late to the game every time, hopefully their offering is very polished.

There’s being late, then there’s completely missing the boat. Which is what Apple has done with Siri.

 

[neuropsychguy]

Surprised this article has no mention of Apple’s Ferret open source project which is a multi-modal LLM.

Ferret is pretty cool: GitHub - apple/ml-ferret.

 

[brijazz]

that’s why running the requests through an LLM-style AI would help Siri understand your request and do a better job connecting that to the right service.

That's great, but wouldn't it have been easy enough for Apple to prioritize the artists I most often listen to in search results? They already know what's in my play history.

 

[brijazz]

Yes, Siri could be way better. I don't know if there are many people on earth that disagree with that.

But I still don't understand what that has to do with Siri being "one of the biggest missed opportunities in Apple's history".

Even if Siri was every bit as good - or even slightly better - as Alexa or Google Assistant, would it materially change anything?

I just don't see it.

Having something that's markedly better than your competitor tends to be good for sales, while having something that your customers complain about tends to be bad for sales. If one of your key features is crappy, people (especially first-time customers) might decide to look at other options next time around.

 

[Razorpit]

Turn on (or off) the lights in room a AND room b. Siri gets it right 10% of the time. From this user's perspective Siri and the surrounding functionality is an unbelievable failure. What makes it doubly awful is that it appears to be deployed with top notch hardware. But this hardware seems saddled with Siri related software that has zero quality control or even simple rational thought built in.

Siri gets my "Hey Siri, good night" correct about 80% of the time. Usually leaving one of my lights on somewhere in the house. A followup "Goodnight" fixes things almost 99.9% of the time. Usually by the third "Goodnight" my lights are set. Nothing has changed in my configuration for probably 3-4 years if not longer. Why it's gotten worse over the years instead of better is any ones guess.

Oh and I love when, "Hey Siri, set the window light to 25%" gets me the response, "Sorry I cannot accept two requests at once." 🙄

 

[Fraserpatty]

yes. Siri would be a lot better if she could just understand the conjunction “and“ as two separate commands.

 

[deepspacecowboy]

Amazon reports that Alexa and Echos lost $10B last year. It doesn’t appear that very many people ask Alexa to buy something.

That's a huge loss for sure, so I would have to agree with you!

Personally I think all of these assistants are trash. They seem exciting at first, but one quickly realizes how limited and error-prone they are after a few days of use. Siri was released ~13 years ago and none of these assistants has really improved in any significant way since then. I think the novelty is wearing off for most people.

 

[Tagbert]

That's a huge loss for sure, so I would have to agree with you!

Personally I think all of these assistants are trash. They seem exciting at first, but one quickly realizes how limited and error-prone they are after a few days of use. Siri was released ~13 years ago and none of these assistants has really improved in any significant way since then. I think the novelty is wearing off for most people.

That’s why we are hopeful that Apple can inject some next generation AI into Siri. It’s time for a change.

 

[deepspacecowboy]

That’s why we are hopeful that Apple can inject some next generation AI into Siri. It’s time for a change.

I'm not holding my breath. Apple hasn't improved Siri much in over a decade. It's very early in the AI game, but it's also quite clear that Google and Microsoft/OpenAI are making huge leaps. Apple is ultimately a very conservative company these days. They aren't bold anymore and their focus on user privacy is also going to make it harder (not impossible) to develop competitive AI. Just like Siri is the laughing stock of digital assistants, I think AppleAI will be the industry's joke AI.

 

[Tagbert]

I'm not holding my breath. Apple hasn't improved Siri much in over a decade. It's very early in the AI game, but it's also quite clear that Google and Microsoft/OpenAI are making huge leaps. Apple is ultimately a very conservative company these days. They aren't bold anymore and their focus on user privacy is also going to make it harder (not impossible) to develop competitive AI. Just like Siri is the laughing stock of digital assistants, I think AppleAI will be the industry's joke AI.

None of the voice assistants have improved. That doesn’t rule out this next generation being significantly different.

We’ll see.