iZac
macrumors 68040
Excellent, a standalone app I can just delete
They can run into inconsistencies. Will I be able to use extensions to change Siri's voice to third-party ones like chatGPT or Anthropic (the EU sued them for it), but not use the Siri powered by Gemini because it's a local model? Can I use the chatGPT or Gemini app because it's cloud-based, but not Siri's?That's a great question. Apple will likely limit it to iPhone 15 Pro or newer, even though it could probably run on much older devices. We should find out soon enough - WWDC is just a few weeks away.
This will be nice for controlling the iPhone functions but I'll continue to use Grok day to day for chat and information.
I really hope you are right. Apple has missed the ball on more than a few occasions. Twice actually. I think the term is...3 strikes and you are out.This is going to change chatbots forever mark my words.
I had an mp3 player before the iPod and a smart phone before iPhone.
We are about to see history repeat itself!
Van down by the river Siri snaps back angrily.., then takes a few seconds to cool down and answers the question. Closing with the standard, “do you have any extra cigarettes?”In a van down by the river!
I really liked the glowing orb. It seemed really futuristic to me.I actually like the current Siri visual implementation where the edges of the screen glow. It makes for nice consistency across all of Apple's devices. If they use the "pill" for Siri, how will that work on a Mac or iPad that doesn't have a "pill" cutout, or even CarPlay for that matter? I think it might end up being more confusing for users.
Yeah…it’s not like the people from the future got burned by fake Apple promises in the past. Say it ain’t so Joe.Wow, this place has so many people from the future. It’s amazing.
I will have a hard time believing it even after it ships...I believe it when it ships.
Apple failed at LLM transformer models. Which is not the same as failing at AI. In April 2026 (recently) Apple held a presentation of ParaRNN which have significant advantages over transformer LLMs. Especially now that Apple demoed the parallel training of such models is possible. I wouldn't count them out just yet. ParaRNN performs MUCH better at many tasks and can be run locally lot less "expensively", in terms of resources.Apple failed at AI and has to rely on Google to bail them out!