I’m not particularly a fan of Steve Jobs, but he wasn’t a follow-the-hype hustler like Sam Altman is. The situation of the two companies are also so different that the analogy makes no sense in any way.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Reignites Talks With OpenAI About Generative AI for iOS 18
- Thread starter MacRumors
- Start date
- Sort by reaction score
ChatGPT stormed onto the scene on Nov 30, 2022, but was in development for 6 years before that. Apple has been working with ML (machine learning) technology for far longer. They are not suddenly trying to catch up with "AI" in general.
It's just this new "generative AI" that suddenly overtook the world that is giving the impression that Apple is somehow "behind" on the times.
It continues to puzzle me why people think that Apple even needs to be part of the generative AI craze. You want ChatGPT, Baidu or Gemini? They all run beautifully on your iPhone and your Mac. Apple already made the sale of hardware upon which engineers built these tools and consumers consume them. Having Apple offer up their own is just adding another burger to the plate, not a new plate.
EDIT: Re-reading the above comment ... maybe the drive is that Apple will do generative AI with privacy and responsible data sourcing as part of their effort, compared to these others?
You are very wrong. You've been using AI (the subset called ML - machine learning) for decades with Apple tech.
This new era is a separate field called "generative" AI and it's only just beginning.
- Did Apple miss the boat on personal computing? This existing before the Mac.
- Did Apple miss the boat on MP3 players? They existed long before the iPod.
- Did Apple miss the boat on "smart" cellphones? They existed long before iPhone.
Experts in the AI field often cringe when people use "artificial intelligence" too specifically. It's a very broad umbrella term with MANY different sub-fields under it. When you type on your iPhone keyboard, that's employing a form of AI called Machine Learning (ML). When you take a photo with your iPhone, that's applying machine learning. And the list goes on...
Machine learning research dates back to the 60's and 70's. It's just that "AI" sounds so cool, it gets used to describe it when in fact many types of machine learning are not "intelligent" at all.
What is true intelligence? The ability to make new decisions based on old decisions, reinforcing those old decisions even further.
The definition of insanity: "Making the same choices over and over, but expecting a different outcome". Each decision made is adding reinforcement to neural pathways.
Could you imagine if our computers became truly insane? It's only a matter of time...!
Incomplete leaks, patchy reporting, and knee-jerk assumptions make it seem like Apple is off their game. Maybe they are, maybe not. Maybe Apple might be counting on uninformed ****** bags to keep public expectations in check. We all know how Siri turned out, and how Maps turned out And porn scanning. And AirPower. and dare I bring it up, VisionPro...
Apple has almost certainly been developing their own LLMs and heavily testing all the others, on device, on local servers and in every mainstream cloud infrastructure service. Probably since they were research thesis, alphas, betas, dot-zero, and ongoing. They certainly have the budget for it.
And with that budget, they almost certainly built alpha hardware and software for scenarios where Apple's own and each mainstream AI engines were integrated into the OS - playing what-if.
Apple would be (has been?) wise to let the initial wave of consumer-grade AI mania die down. Apple should invest in AI methodologies that incorporate, or intrinsically apply, intellectual property protections, indelible watermarking, and suppress "hallucination". Maybe Apple is attempting to "AI Different" - to break away by emphasizing INTEGRITY and TRUSTWORTHINESS.
Or maybe Apple sucks and is doomed. Again. There. Covered all the bases. Moving on...
Apple has almost certainly been developing their own LLMs and heavily testing all the others, on device, on local servers and in every mainstream cloud infrastructure service. Probably since they were research thesis, alphas, betas, dot-zero, and ongoing. They certainly have the budget for it.
And with that budget, they almost certainly built alpha hardware and software for scenarios where Apple's own and each mainstream AI engines were integrated into the OS - playing what-if.
Apple would be (has been?) wise to let the initial wave of consumer-grade AI mania die down. Apple should invest in AI methodologies that incorporate, or intrinsically apply, intellectual property protections, indelible watermarking, and suppress "hallucination". Maybe Apple is attempting to "AI Different" - to break away by emphasizing INTEGRITY and TRUSTWORTHINESS.
Or maybe Apple sucks and is doomed. Again. There. Covered all the bases. Moving on...
Uh buddy! Microsoft is a 3 Trillion dollar company too and they use Open AI almost exclusively!
They may be discussing how much to charge the public for their API usage and how much of a cut out of it Apple would receive. Maybe the discussions are about giving as little a cut as possible for OpenAI/Google out of whatever Apple charges. Not sure what else could Apple be discussing. Certainly not something that benefits its customers, that ship sailed a long time ago. It must be about something that benefits its shareholders.
And why do you think MSFT use OpenAI? Nothing to do with Satya had the foresight and invested 10B into OpenAI right?
What did Tim Cook see? Too busy living in metaverse of Vision Pro.
Not all AI features will require third party APIs and not all upcoming iOS LLM powered features are going to be released this year. Right now companies like OpenAI are very good at very specific AI use cases so Apple's AI push is not contingent on a single partner. Apple recently released a public version of an in house LLM called OpenELM, we can assume they've been working on this for some time and maybe some iOS 18 features using this tech are already in their final stages for the June beta.
Microsoft invested in Open AI in 2019 and again 2023. So, not really comparable.
Sometimes with new tech you outsource to the people that have already established a solid lead because we don't know how this field is going to develop over the next couple of years. You could spend a few hundred million building an in house model like Facebook only for that model to be outmoded by a new breakthrough next month (1 bit LLMs anyone?). Intel got themselves into this mess by building their own fabs which are lagging behind the best from TSMC.
This way they can utilize the best LLMs out there without getting themselves into deep technical debt. In a year or two they may decide (if they haven't already) what direction they want to focus on with in house development. Many in the industry are designing their own custom chips for AI/LLM instead of relying on Nvidia. Apple already have a chip design advantage so rushing to build an LLM with Nvidia like everyone else might not be the best strategy long term.
Apple are doing the wise thing here IMO. Apple have demonstrated their ability to think long term is a cut above a lot of their competitors. They don't rush into things out of desperation to impress the market, they already have an outstanding reputation so there's no need. Whether or not customers are satisfied with this approach is another thing.
The words “never” and “Apple” get paired a lot.
Everyone has a lot to say based on rumours, for all we know this could be but a fraction of Apple’s AI strategy.
It seems to me like they’re going all in with on device processing, and will supplement their local LLM with a third-party service to fill in the gaps
I wish Apple just released their own models instead of using other companies. It’d be interesting to see what they come up with and I don’t get why’d they want to increase dependency on other companies.
The whole “don’t do it first, do it right” concept.
I get baking in generative AI on device for specific tasks like mail, calendar, music and productivity apps. I don’t understand why everyone wants it in every app.
I have no need for Meta AI. I’m in WhatsApp to mute the group chats that blows up my phone, leave me alone 😂
If Siri is the leading indicator for Apple AI, prepare for underwhelment to give way to frustration and ridicule.
Maybe the roadmap is the problem as it seems to have blind spots that result in playing catchup on emerging trends.
Kinda puzzles me too. It’s not like I’d use anything AI related from apple. I think people just want a better Siri and that is what they think of when AI is mentioned.
It’s a complete disgrace that Apple doesn’t have their own tech to use here.
And now they’re choosing between the personal information exploiter and the copyright thief?
And now they’re choosing between the personal information exploiter and the copyright thief?
Most of Microsoft investment in Open AI is Cloud credits on Azure, basically renting GPUs for control of the technology. It’s not like Microsoft wrote a 10 B check to Open AI.
I hope not. I don’t want my iCloud contents to be used for training GPT.
Businessmen more interested in $$ than in human kind, although they may have said the opposite many times.