Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Most iPhone owners see little to no value in Apple Intelligence so far

9to5mac.com

Post reply on HN
Print view View original

351–360 of 407 posts

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#351

Post by xvector »

reilly3000 wrote: post #315
bangaroo wrote: post #218

Apple Intelligence is kind of uniquely bad. I don't understand how, but its AI message summaries often completely flip the meaning of texts I've received. I've gotten summaries that say "X person hates Y" and I read the message and it's them talking about how much they love it. It is impressively, wildly inaccurate. It confuses things like people's names and the messages they're sending and sometimes will paraphrase…

It’s bad because they are using very small local models on device, like on the order of 3 billion parameters as opposed to the trillion+ param models available in the cloud. They could keep the security model and have larger, slower models for summaries (what’s the urgency of a summary?) but that would be a battery killer. It’s clear they need to train on messaging content, but they have promised not to use their use…

There are ≤3B models far better than the absolute garbage Apple is running on the iPhone.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#352

Post by xvector »

Etheryte wrote: post #337
jdprgm wrote: post #175

I think this is much more on a botched rollout and a lack of vision. Apple should have been working on this for years and have some pretty big advantages. A deeply integrated version of the best o1 chat and vision models with context from all various apps and ability to do complex app interactions could be amazing. They could even have the ability to have AI basically create on demand personalized new applications fo…

This wouldn't align with what Apple is going for though. Their selling point in this category is first and foremost that your data stays private. They're looking to do as much as possible on-device, not in the cloud. This rules out models akin to o1 because they're just too large in every respect. I agree that Apple has the chance to do amazing things here, but I think it will take a while until small models are as c…

The problem is they compromised on both fronts (privacy, compute) and produced something that is ultimately a useless joke.

They could have produced something that was good and required accepting a privacy disclaimer to use. Or they could have waited until they managed to catch up with the rest of the industry wrt good 3B parameter models (the one they're using is incredibly terrible.)

Instead we got this half-assed compromise.

The execs in charge of Siri are demonstrably terrible at their job and should feel bad. It's incredible they get paid the millions they do while doing such a horrible job.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#353

Post by Smaug123 »

llm_nerd wrote: post #339
Smaug123 wrote: post #335

Earlier quoted context omitted.

You certainly strongly implied that NDAs don't apply with the phrase "their magically novel, super unique code is just too secret". You used the words "is just too secret" sarcastically, which led me to believe you think in fact the work is not too secret to exfiltrate to the LLMs on the Internet: that is, that NDAs don't apply. > In reality a programmer could yield immense value from such tools having a) given it ze…

I did not imply that. I implied that people who have projects for whom these tools are purportedly useless can never tell you anything about the project even remotely because it makes it impossible to note that it, like most projects is likely overwhelmingly well-trodden ground with a mild remix. Loads and loads of super proprietary projects working in very unique domain spaces have code that is overwhelmingly common…

If you want something concrete: I have never yet seen any even slightly competent code come out of an LLM when it's writing in F#, Mathematica, or POSIX sh, which are broadly the languages I write in. (It's possible Wolfram's very new LLM does better on Mathematica because it's been trained on it specifically, but I haven't tried it.) I've only had Sonnet for a day (having been burned too many times by other LLMs, but decided yesterday it was time to try again), so am not in a great position to comment on Sonnet's abilities with those languages specifically. My preliminary guess is that we've gone from 4o's "20% correct" to Sonnet's "50% correct".

> Their state in the past is irrelevant.

Apparently also their state in the present? I said "just about edged out ahead", whereas you say "immense value".

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#354

Post by layer8 »

PittleyDunkin wrote: post #277
layer8 wrote: post #223

Earlier quoted context omitted.

I agree, but based on their software track record over the last couple of years, I have little hope that we’ll get much beyond a disparate collection of mostly half-baked features.

> I have little hope that we’ll get much beyond a disparate collection of mostly half-baked features. Really? AI has driven most of the interesting features they've released in the last decade. Notably, organization of the photos app and indexing of people, pets, places, etc. (Nb I have not yet updated to ios 18 and I understand the photos app is somewhat controversial.)

The GP comment was calling for complex context-aware cross-app functionality. The new Photos app features do have their use, but they are limited to the particular compartmentalized use cases envisaged by Apple. For example, they support cats and dogs specifically, but not pets in general.

There’s also the tendency that once a feature has shipped, it’s apparently considered “done”, and the lead developers likely move on to other things. It doesn’t feel like there is a continuous ownership and coherent vision behind the features and apps, or behind the overall OS trajectory.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#355

Post by weikju »

reilly3000 wrote: post #332
reilly3000 wrote: post #322

Here is what I want from Apple Intelligence: Given a user story “I chipped my tooth” I say “hey siri, I just chipped my tooth, can you help me?” I would want it to: - help me schedule an appointment with my dentist (it could create a reminder, ideally know who my dentist is, even ideally queue up or place a call) - put that appointment on my home and work calendar - remind me to tell my manager I need the time away -…

Ohh and I would like to be able to tell it not to keep featuring photos of my ex wife years after the divorce.

> and I would like to be able to tell it not to keep featuring photos of my ex wife years after the divorce.

You can already do that by pressing the three dots menu on a featured photo or memory and selecting the Feature this person less, with then options to not feature at all or just less.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#356

Post by billychuck21st »

nozzlegear wrote: post #231
billychuck21st wrote: post #8

Earlier quoted context omitted.

I think the data would likely to show most people are spending time on short videos.

Why do you think that?

because there are plenty of research showing majority of screen time is spent on video watching, followed by social media and games.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#357

Post by nozzlegear »

billychuck21st wrote: post #356
nozzlegear wrote: post #231

Earlier quoted context omitted.

Why do you think that?

because there are plenty of research showing majority of screen time is spent on video watching, followed by social media and games.

Ah, I didn't know that. I try to avoid video watching myself, especially shorts/tiktok-style content. My wife, on the other hand, primarily uses her phone for that stuff, so my gut tells me you're right.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#358

Post by drivingmenuts »

Aside from the image cleanup, the usages on the iPhone seem like solutions in search of a problem. Apple Intelligence on the desktop is pretty limited, so it's pretty underwhelming, too. There may be some valid usages in certain business areas, but all the hype outside of that is starting to smell like a con job.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#359

Post by llm_nerd »

Smaug123 wrote: post #353
llm_nerd wrote: post #339

Earlier quoted context omitted.

I did not imply that. I implied that people who have projects for whom these tools are purportedly useless can never tell you anything about the project even remotely because it makes it impossible to note that it, like most projects is likely overwhelmingly well-trodden ground with a mild remix. Loads and loads of super proprietary projects working in very unique domain spaces have code that is overwhelmingly common…

If you want something concrete: I have never yet seen any even slightly competent code come out of an LLM when it's writing in F#, Mathematica, or POSIX sh, which are broadly the languages I write in. (It's possible Wolfram's very new LLM does better on Mathematica because it's been trained on it specifically, but I haven't tried it.) I've only had Sonnet for a day (having been burned too many times by other LLMs, bu…

>Apparently also their state in the present?

I find a number of models fantastically useful in the present. I'm not sure why your one day evaluation of them is relevant to how I utilize them.

It was inevitable that someone was going to do the "but I insist upon writing in some fringe language" bit (when the "my project and domain is too unique" bit faltered), but many models actually have excellent F# abilities. Eh.

Re: Most iPhone owners see little to no value in Apple Intelligence so far

#360

Post by ToucanLoucan »

llm_nerd wrote: post #347
mrtranscendence wrote: post #345

Earlier quoted context omitted.

You seem to readily agree that my use case is inappropriate for LLMs, but not ToucanLoucan?

Zero LLMs have been trained on doing sales forecasting to my knowledge (and it isn't the right use regardless). In contrast, many LLMs have been trained on enormous quantities of code, coding languages and platforms and uses. Billions and billions of lines of code covering just about every sort of project. Millions of projects. If someone says "Well my software dev project is too unique and novel and they are therefo…

Would it help if I said most “normal folk” applications of LLMs are a waste of time and money too then? Because I’m also absolutely a believer that a huge bubble burst is coming for OpenAI and company.
Post reply on HN