Rumored iOS 18 Siri boost will be driven by massive acquisitions over years

AppleInsider

Apple's investment in more AI firms than any of its major rivals, and a marked increased in related job ads, backs up claims that Siri will see significant improvements at WWDC 2024.

A Siri icon superimposed on Apple Park

Again belying the claim that Apple is lagging behind the industry in its use of AI, the company has been hiring generative AI engineers, and more.

It's also been acquiring companies such as AI-powered video compression firm WaveOne, and recently publishing AI research papers. Apple actually won an award for an AI research paper -- back in 2017.

Compiling reports of job ads, research papers and AI investments, the Financial Times says that Apple is now boosting its efforts to bring ChatGPT-style generative AI to iPhone.

"[Apple is] getting ready to do some significant M&A," Daniel Ives at Wedbush Securities told the publication. "I'd be shocked if they don't do a sizeable AI deal this year, because there's an AI arms race going on, and Apple is not going to be on the outside looking in."

Listing both equity and add-on investments, the Financial Times says that Apple has made 21 such investments in firms since 2017. Over the same period, Accenture made 19, while better-known AI rivals Microsoft and Meta made 12 and 11 investments respectively.

Intel reportedly invested in or acquired 9 companies, while IBM and Google's parent Alphabet worked with 8 each.

The Financial Times does not give a reason why its count starts in 2017. However, significantly, Apple hired AI head John Giannandrea away from Google in 2018.

It also cites Morgan Stanley as reporting that almost half of Apple's current job ads include some reference to AI, Machine Learning, or Deep Learning. That's a marked increase from Morgan Stanley's November 2023 report, when it said that since 2015 "AI-related job postings at Apple have risen from 5% of total Apple job postings to ~20% today."

Rumor Score: Possible

Read on AppleInsider

cincytee

Not really germane to the story, but, for the record, I like the Siri-over-Apple HQ illustration.

gatorguy

It's not at all important what they call it. Appls say Spatial Computing which is AR/VR to others.

Apple said Retina Display which was qHD or QVGA.

Apple uses a catch-all Machine Learning while others refer to AI or Generative AI as well as Machine Learning. 

All this special naming is simply attempting to make one's services sound unique and, thus, more special than someone else's feature/services.
What matters is how useful it is and whether it will work consistently. That's where they all have work to do. 

EDIT: After writing this I realized the third one is not like the first two, where Apple implies uniqueness. When it suits them they can do just the opposite and claim that their Machine Learning is just the same as what anyone else calls Generative AI. 

Apple has the strongest and most focused marketing machine on the planet.

macxpress

Hope so....tired of Siri being the laughing stock of assistants. It's pretty good some things but there are other areas it needs a lot of work. 

hexclock

Me: “Siri, get me directions to X”

Siri: “Take the I dash 90 East for 8 miles…”

Respite

It’ll need to be one hell of an improvement.

Clarus

gatorguy said:

Apple said Retina Display which was qHD or QVGA.

Part of your unhappiness may be from over-simplifying at least one of those terms.
Because you definitely over-simplified Retina.

You mentioned Retina being the equivalent of this or that standard fixed resolution. That is plainly wrong. Retina has never referred to a specific hard resolution, that is not how things work now.

Retina is more about the Device Independent Pixel (DIP), which is an industry-wide concept driving standards like CSS.

The DPI resolution of the DIP is not a single fixed number. This is where you ran into trouble. The pixel size of the DIP is corrected for display pixel density and viewing distance, and the way they quantify this consistently is defining the DIP as a specific angle of view. This allows an object to be displayed at a constant apparent size to the eye, because the DIP corrects for screens being viewed at phone distance vs computer distance vs billboard distance.

That is why Retina for a phone is a different DPI resolution that Retina for a laptop screen.

The other way your analogy falls down is that if you look through all the desktop and mobile Retina screens Apple has made since 2012, the DPIs are all over the place! And many do not exactly match your QHD or QVGA!

The uninformed view is that Retina doesn’t mean anything, but the informed view is that all those DPI values make total sense after you learn that they are consistent with what the DIP size should be at each device’s viewing distance.

mattinoz

hexclock said:

Me: “Siri, get me directions to X”

Siri: “Take the I dash 90 East for 8 miles…”

Hit a fun Siri quirk the other day. 

Asked how far away is the moon.
Siri showed the answer on screen in kilometres the correct units for my locale. Spoke the answer in miles.

when I repeated the question and asked in kilometres the response was “here’s something I found on the web”

given Apple are so keen for app developers to make sure they use localised conversations it is strange they don’t hold themselves to their own standards 

hoping for some big improvements.

gatorguy

Clarus said:

gatorguy said:

Apple said Retina Display which was qHD or QVGA.

Part of your unhappiness may be from over-simplifying at least one of those terms.
Because you definitely over-simplified Retina.

You mentioned Retina being the equivalent of this or that standard fixed resolution. That is plainly wrong. Retina has never referred to a specific hard resolution, that is not how things work now.

Retina is more about the Device Independent Pixel (DIP), which is an industry-wide concept driving standards like CSS.

The DPI resolution of the DIP is not a single fixed number. This is where you ran into trouble. The pixel size of the DIP is corrected for display pixel density and viewing distance, and the way they quantify this consistently is defining the DIP as a specific angle of view. This allows an object to be displayed at a constant apparent size to the eye, because the DIP corrects for screens being viewed at phone distance vs computer distance vs billboard distance.

That is why Retina for a phone is a different DPI resolution that Retina for a laptop screen.

The other way your analogy falls down is that if you look through all the desktop and mobile Retina screens Apple has made since 2012, the DPIs are all over the place! And many do not exactly match your QHD or QVGA!

The uninformed view is that Retina doesn’t mean anything, but the informed view is that all those DPI values make total sense after you learn that they are consistent with what the DIP size should be at each device’s viewing distance.

A nicely detailed post, but you could have saved the effort if you had paid attention to what I said: Was, when it was first employed, not is. I agree with your "is". 

Clarus

gatorguy said: A nicely detailed post, but you could have saved the effort if you had paid attention to what I said: Was, when it was first employed, not is. I agree with your "is". 

I don't think that is true. Retina was probably designed as not resolution specific from the beginning, following emerging industry trends. Retina and the parallel HiDPI standard in Windows and Android were the same response to the fact that if something was not done to compensate, the increasing resolution of monitors would have made text and UI tinier and tinier until they were impossible to read. The concept of distance-adjusted scale factors had to be implemented if that problem was to be avoided.

It seems very unlikely that one year Apple decided to say "We have xxx dpi." That has no long term road map. The actual Retina explanation fits into a much bigger picture and road map that the entire industry was following.