Hacker News Reader

Comment by scubadude | Hacker News Reader

I guess unfortunately Apple has no choice but to include bullshit AI in it's products and services or look like they have been left behind.

View on HN

TBF, if you exclude Generative AI, they’ve been ahead of the curve with including an NPU and on device ML features. They’ve pretty much included some ML feature in their product releases since the iPhone X was released.

Came to say the same. My gut tells me they wouldn't go anywhere near this unless they were concerned to be seen not getting on the bandwagon. It is worrying that "AI" (read silly LLM toys) has become so prevalent that they need to do this.

I just noticed the other day that my photos have a Siri tag which auto identifies plants. Voice memos doesn't identify bird songs.

I'm of the opinion that not every company needs to do its own AI thing. Let AI companies fight it out, and use the best.

Right now sure AI is a money vacuum, but will eventually be commodotized where only the best make much money.

Apple is probably smart to focus on their own hardware and software, and just rent out the best AI available.

Apple should consider an iPhone/iOS light, a version stripped of anything but the bare essentials. I get why they don't, the market isn't big enough, and they can barely sell their current best phone, the iPhone SE.

> they can barely sell their current best phone, the iPhone SE.

They need to stop making phones every year. Most people in the world already have phones and don’t need a new one every year.

They also need to make 4 inch phone, like iPhone SE 1. That was the perfect size for a phone.

But people need a new phone eventually. If you continuously manufacture phones anyways you might as well update the hardware while you're at it. (It also helps with keeping the price up at >1k)

I for one wouldn't buy a one year old new phone. Used sure but then the price is right.

> They also need to make 4 inch phone, like iPhone SE 1. That was the perfect size for a phone.

There are dozens of people who agree with you ;)

Apple came late to the big phone and were pulled there by the market. I'm surprised they still release a smaller phone at all given the rumors of such low sales.

They would be giving up massive amounts of profit if they did that. There would be a serious case for investors to sue if they made that decision.

I don’t think Siri but as an LLM would be “bullshit.” It’d be an absolutely immense upgrade.

Hallucinations would be terrible for brand appeal beyond being limited. Nobody has solved that problem yet. Just take a look at Google’s latest search results giving actively harmful information.

I suspect that current ML tech is good enough to constitute a big discrete step forward in terms of how much value Siri is providing as a digital assistant that's also in charge of your home. If that's the case, there's more harm in being reputed as the company that is one serious step behind in a space that isn't yet settled vs facing the same level of gaffs that Google and Microsoft are putting out.

Just look at Microsoft right now. They're experiencing all the heat that OpenAI brings, including unforced errors like Sky. Reputationally speaking, are they better or worse off having made such an early push for LLM product integration?

I would argue that Microsoft do not have the same brand that Apple does. So they’re much more free to pursue OpenAI partnerships regardless of how well it works.

I don’t mean brand as in purely recognition, but what it means to people. Apple are held to a higher standard because of the specific identity they’ve cultivated. (Arguing the specifics of whether that’s deserved or not will be its own essay so I’m leaving that out)

You see this all the time on HN or Reddit, where any negative subject about Apple gains significantly more traction than other companies.

Meanwhile Microsoft has an identify that is more “meh”. Users expect Microsoft products to not have the polish, to have ads and be weird/quirky. Microsoft did Windows phone and the zune, Apple did the iPhone and the iPod.

Does that mean one companies products are better or worse? Maybe? Maybe not.

But it does mean that the brand appeal of every company has different characteristics that need to be accounted for. Siri giving bad answers is much worse perceptually than Cortana/Bing giving the same answers.

It also means Microsoft is more free to experiment and do things that are half baked because their brand identity allows for that. Often that leads to really interesting use cases ahead of its time.

I’d also add that Apple being “behind the curve” perceptually has rarely been an issue. The brand is that they enter late and redefine the segment. Again, not always deserved, but that’s the association. They often ride out trends that way.

You can see this even with Google vs Microsoft/OpenAI. Google had their LLM chat-bot working internally (remember the ethics guy who got fired because he was so convinced that the chatbot had a soul?), but they clearly thought that the level of hallucinations at the time would be bad publicity as all the news would be "Google's chatbot spews nonsense!". But OpenAI and Bing had nothing to lose, they could launch them with no expectations of quality, taking Google by surprise, and now everyone just says Google was behind instead and had to scramble to catch up.

Same thing with Apple - if Apple had launched Siri with GPT-3.0 before ChatGPT, they would have been a laughing stock with how bad it was. But instead now they're behind, even though the competing products are STILL unsuitable.

Yeah the Google vs OpenAI comparison is really apt.

ChatGPT will give terrible hallucinations as well, but Google is getting lambasted right now for theirs. But even then, it’s written off as Google being Google.

If Apple did it, it would take over the tech news cycle for weeks. Personally I do think it’s better for Apple to be seen as behind in this trade off.

It can't be any worse than how bad Siri and Alexa are right now. Amazon gave me one of those Pop things recently and it was set up with the wrong zip code. I kept asking it for the weather in Chicago, every way that I could think of, but it had no idea what I was trying to ask for. The speaker is great in it, though.

It absolutely can be worse.

Right now, their limited subset of answers would be wrong in a way that is easily correctable. They fail safe.

An LLM doesn’t fail safe. It fails confidently. Correcting bad responses can be a very expensive and time consuming effort. In the meantime you get actively harmful suggestions: https://arstechnica.com/information-technology/2024/05/googl...

Alexa telling you the wrong weather is a lot less harmful than it telling you to do something that would harm you (staring at the sun, running with scissors, smoking when pregnant) or spreading political misinformation (telling you conspiracy theories).

> Hallucinations would be terrible

Tighten the temperature.

Everyone is trying to sell AI as AGI. Apple’s genius will be in breaking from that. Simply, dependable, tightly controlled LLM that isn’t even branded as AI. Just a thing that does certain tasks for you better.

They could use OpenAI only as the natural language processing part, and constrain it to only using the vetted services/actions that Siri already can do rather than answering open-ended questions. That would still make Siri massively more useful than it is today since it would remove the need to word things in a specific way.

It wouldn't be bullet-proof (there are always "jailbreaks" to these kinds of prompts) but it might be good enough if they slap a "beta" tag on it.

What would you be using it for? More and more I fail to see the value in LLMs. Something like Copilot, if used correctly, does have a bit of usefulness, but that's really the main use case right now. You can't trust the LLMs, so you can't use them for anything where you aren't going to go through the output meticulously afterwards.

Then again, I don't see the point in Siri either, it's useful in only a handful of cases.

I went back to iOS recently, and with AirPods or CarPlay, I found Siri to be a really pleasant surprise. If I need something done, like a text message sent or a quick calculation done, I can just ask without taking my hands or eyes off what I’m doing. It’s also gotten a lot better at taking the hot word + command + arguments without breaks. “hey siri, send a message to dan I’m here” gets you “to dan: I’m here, send it?” For CarPlay, “hey siri, navigate to xyz restaurant” works great to start the navigation if you change your mind about where you’re going once you’re already on the road.

Combine that with “announce notifications” and it really feels like the future, “dan says, are you on your way” and it listens for a few seconds, where you can just say “reply yep I’m 15 minutes away”.

They’re already bringing AI into it a bit, if you get a picture it will try to describe what’s happening in the picture “dan sent a photo of a piece of paper with text on it” which is interesting but it’s not getting enough detail in the descriptions to be of much use just yet. A good multimodal model could definitely improve that, though it must be noted I prefer a worse on-device model to a server based model that can collect a bunch of data.

I think for really off the wall questions/brainstorming, going out to an LLM would yield better results than the current behavior of pointing you towards Google in the browser. “hey siri, what are the typical speeds people achieve on a SPI bus” would be nice if it could give some answer even if it’s not always 100% perfect, it’s kinda like bouncing questions off a slightly unreliable coworker, but you know, better than nothing when you can do it hands/eyes free.

I also have a hope that AI (although maybe not specifically LLMs) would help these assistants handle more complex tasks, like “add eggs to my grocery list” where your grocery list is just in a note and not in some special place. It doesn’t feel out of the question given current technology, I just think nobody’s done it yet.

I have the opposite experience with Siri. Maybe it’s my Australian accent. I have HomePods around, and I use them for lots of stuff. But the voice assistant is barely useful.

“Hey siri, turn on the amp.” “Ok, your media will play louder.” “Siri Set alarm for 4:18” “I’ve set an alarm for 8 o’clock” “hey siri cancel alarm” - radio silence. (Listening to podcast, and my alarm goes off) “hey siri stop” (the alarm stops, so does the podcast). “Hey siri play” “ok, here is some media” (random music plays) … uuurrrrrgghhhh.

I can’t wait for chatgpt to replace Siri.

I just disable Siri. Other than driving I can't imagine talking to Siri, it feels inefficient compared to just pushing a button.

> You can't trust the LLMs, so you can't use them for anything where you aren't going to go through the output meticulously afterwards.

People keep saying things like this, and I’m confused because I go to Google for facts, and chatgpt is still super useful. Things I’ve used chatgpt for recently:

- Here’s some text I’m writing for a paper. Help me reword it to make it sound more natural.

- I want to put coconut crème into a stew. When should I add it? What else would go well?

- I’m making a Makefile which has some unusual requirements. How do I do that? Here’s the actual commands I want it to run. Write the makefile. (The makefile wasn’t great, but it was a good start and let me hit the ground running.)

- Generate 100 quirky characteristics for an npc in D&D

And so on. It’s a tool for creativity. I absolutely can’t wait for my HomePods to be replaced with chatgpt.

Reply on news.ycombinator.com