Apple Engineers Reportedly Working on ChatGPT-Like AI Despite Siri Design Flaws

[DocMultimedia]

Hey Siri, increase the volume.

"I'll do that after you get divorced and start drinking a six pack for lunch."

 

[xraydoc]

They need to rebuild Siri from the ground up if they want to be serious about AI

Agreed. Hopefully they can do this before too long. In another year, ChatGPT and spinoffs are going to make Apple look like they're a decade out of date.

 

[brofkand]

Siri is complete trash. They should abandon it and buy something better if they can’t manage it themselves. I am sure there are AI startups that are in big trouble right now due to the SVB collapse

The sad thing is that is how Siri started - they bought it. But they have no idea how to do this type of thing.

 

[JM]

The phrase "working on" in corporate double speak can have various meanings depending on the context, but it generally implies that the company is actively engaged in a project or initiative. However, it may also suggest that the project is in its early stages, and there are no concrete plans or timelines in place yet.

In some cases, "working on" can be a vague and non-committal way for a company to acknowledge a problem or issue without making any promises or guarantees about a solution. It can also indicate that the company is exploring different options or considering various approaches before making a decision.

Overall, the phrase "working on" can be interpreted as a way for a company to signal that it is taking action, without committing to any specific outcomes or timelines. It is important to note that this phrase should be evaluated in the context of the specific company and situation, as it can have different connotations depending on the context in which it is used.

From your AI overlord, ChatGPT.

Our forum posts will all be be replaced by ChatGPT one day.

 

[onemorechris]

In my personal opinion, Siri should be scrapped entirely. 🤷🏼‍♂️

Apple have rewritten things from the ground up before. Fingers crossed they do it with Siri

 

[citysnaps]

Why does Apple keep creating products that are seemingly great to start, then allowed to whither? It happens to so much of Apple's eco system both hardware and software. Things always seem half baked. Pick any of their software offerings, FCP, Pages, Keynote, Numbers, Calendar, Siri, AppleTV, HomeKit, etc. It feels like they start out with great expectations and then nothing. They sit without any true innovative updates. And the hardware, Lord, Mac Pro, iPhone, that mouse (OUCH!). It feels like over the years, they have done just enough to sell the same tweaked version as last years model, but nothing that stretches their creativity. I know it's a very complicated process, but where are all the billions going for R&D? I don't know, I guess I'm not really feeling excited about what Apple "may" have coming out with going forward like I used to be.

Yeah...Apple is doomed. Again.

 

[onemorechris]

At some point, you can't keep updating MacOS 8 (later known as Classic Mac). You need a group of engineers to evaluate the acquisition of the AI equivalent of BeOS or NextOS and simply start over...

...except this time, Apple isn't on the brink of bankruptcy.

Yes...I'm that old...I remember.


...in other words, rebuild from the ground up and run current Siri in emulation mode, if necessary

Agreed. Siri is ready for total rework from the ground up. Im sure there are people inside the Loop that have been trying to make this happen

 

[bartszyszka]

If he had that assessment in 2014, couldn't they have completely rebuilt from scratch since then? That was 9 years ago.

 

[falkon-engine]

If Siri "intelligence" is any indication, I would say don't bother.

 

[t0rqx]

From what I have seen the ChatGPT is very impressive yet also very dangerous. Terminator 1 and 2 style?

 

[BornAgainMac]

My Siri experiences have not been that bad.

 

[ipedro]

Tim Cook has been a stable hand at Apple, building on the success of the roadmap Steve had defined. But he's not Steve Jobs. It takes a visionary to look beyond all the noise of "what's next" and identify future technologies and how they'll meet a still unrequested need that'll intersect with that technology becoming available. Steve Jobs' gift was that he could see years and even decades in advance and would start skating towards that puck. Apple doesn't have that today.

Neglecting Siri will probably be Tim Cook's undoing, specially if they're now just reacting to generative AI. Apple isn't going to suddenly collapse just because it doesn't have an openAI competitor but it could gradually make Apple devices less competitive if say Google integrates Bard directly into Android at a time when every user wants apps with generative AI built in and importantly, every developer wants to build those apps. If they can't do it on iOS, they'll build them exclusively on Android and that's going to be a problem for Apple.

I started a thread in February where I go deeper into this subject and my fears of Apple having missed the boat are starting to be realized.

https://forums.macrumors.com/threads/the-ai-wars-are-here-whats-apples-response-to-chatgpt-google-bard.2379877/

 

[spinedoc77]

Maybe with AI Siri will finally be able to understand my poor wife's accent, lol. Don't get me started on my 6 year old, here is him telling Siri to message me "Miss you."

 

[NT1440]

Lots of comments here showing that reality doesn’t matter at all.

ChatGPT is an extremely fancy language generator. It is NOT a repository of knowledge. It is not intelligent, and just because it’s output seem authoritative does not mean what it’s “saying” is *correct*.

It’s a massive problem if people don’t understand that key point.

 

[klrobinson999]

Apple should go full-speed ahead with Siri development, especially now that commercial AI is here. It's unforgivable the lack of advancement in this space over the past decade.

 

[Reason077]

John Burkey, who worked on ‌Siri‌ and was made responsible for improving it in 2014, explained that the voice assistant is built on "clunky code that took weeks to update with basic features."

Well, this explains a lot about why improvements to Siri have been slow-to-nonexistent in recent years. Siri was best-in-class in 2011 but seems hopelessly dated by today's standards.

So are we going to see a total rewrite of Siri's back end, built on a modern codebase utilizing GPT or another LLM AI?

(Edit: Apparently Microsoft has an exclusive license to use GPT-3/GPT-4 technology, so I guess Siri won't use GPT)

 

[jimbobb24]

There is a wide gap between conversational texting and speech recognition. And then another wide gap between getting the speech understood and translating that into code. Which is why most the digital assistants still fail to impress. Siri is just as bad as the rest.


I would like to see them integrate stable diffusion. They could automate the training and potentially make some very fun things possible. Right now it’s quite the process on mac.

 

[ThisBougieLife]

I’m interested in improvements to Siri. I’m not interested in an “essay writing” AI.

 

[bollman]

Siri has been broken for such a long time and it’s clear it’s not going to get substantially better anytime soon. I actually wish we would get to point again where big tech companies just don’t all have to create their own software and hardware inside their walled gardens but rather integrate technologies from other specialized companies. Since when is it a law of nature that Apple has to be good at everything? Why not just license ChatGPT and integrate it in a privacy-protecting way? Then they could focus more on their core competency: Making amazing hardware and some really nice apps. It’s annoying that these tech behemoths just think about how to outdo one another with their proprietary tech while investing billions in in stupid patent fights.

Very easy: that would mean: licensing money out, no "competitive edge" or "unique selling point".
Not that Siri is a very good selling point, still, Apple can tout being the only platform with Siri.

 

[spinedoc77]

Lots of comments here showing that reality doesn’t matter at all.

ChatGPT is an extremely fancy language generator. It is NOT a repository of knowledge. It is not intelligent, and just because it’s output seem authoritative does not mean what it’s “saying” is *correct*.

It’s a massive problem if people don’t understand that key point.

It's been absolutely incredible for me writing insurance denial appeals. I thoroughly check the citations and make sure everything is factual because I don't trust it. But what used to take me 2 hours now takes about 2 minutes, with another maybe 20 minutes of fact checking. But based on what I've read, they need to really get better at plagiarism and factual content.

 

[turbineseaplane]

“There was a problem, please try again”

 

[NT1440]

Well, this explains a lot about why improvements to Siri have been slow-to-nonexistent in recent years. Siri was best-in-class in 2011 but seems hopelessly dated by today's standards.

So are we going to see a total rewrite of Siri's back end, built on a modern codebase utilizing GPT or another LLM AI?

(Edit: Apparently Microsoft has an exclusive license to use GPT-3/GPT-4 technology, so I guess Siri won't use GPT)

As pointed out previously, the NYT article is poorly written to make it appear that the state of siri in 2014 is how the codebase is in 2023. The guy quoted left Apple in 2016 to run his own AI chatbot company.

Tech journalism is a joke, and media literacy is in an even more dire state.

 

[Mr. Dee]

For the basics I use Siri, which is setting a timer and reminders. For complex stuff, and I mean some deep $***, I use the new Bing. I recently used the new Bing to plan a trip and I swear, its someone from that place I plan to visit typing out the text.

 

[erikkfi]

Am I supposed to believe Apple can catch up in this race? They could've been the ones chucking a relatively paltry $13 billion at OpenAI instead of Microsoft if they were serious about this.

 

[NT1440]

It's been absolutely incredible for me writing insurance denial appeals. I thoroughly check the citations and make sure everything is factual because I don't trust it. But what used to take me 2 hours now takes about 2 minutes, with another maybe 20 minutes of fact checking. But based on what I've read, they need to really get better at plagiarism and factual content.

Oh no, it definitely has extremely useful use cases.

My fear is that when not used for work, but inevitably gets used for things it’s not designed for (academics, NEWS REPORTING, etc) it’s going to be a massive problem.

ChatGPT is not a search engine, and has no way to verify its own “factual” output.

When used in a confined space, to do what it’s designed for (generate language based on criteria) it’s great.

But it is NOT a *knowledge* source, at all.