Tim Cook Defends Apple's AI Delay: 'We've Rarely Been First'

[Johnny907]

Meanwhile more AI talent continues to be poached from Apple. That, plus the fact Tim lied through his teeth a year ago regarding the state of Apple AI, tells me everything I need to know. I'm not about to jump ship for Samsung, but I won't be looking to Apple for any kind of AI solution anytime soon.

 

[Johnny907]

I still don't use Apple Maps.

I hope it's going to be better than that.

Right? Anytime an app or website tries to load Apple Maps by default, that's a hard nope. Copy, open Google Maps or Waze (google for local, Waze for roadtrip), paste.

 

[novagamer]

I dunno... If Apple with their $500 Billion investment will soon be manufacturing their own AI servers this year, I would guess any Srouji team AI chips are pretty far along and they have already been testing first silicon. Pure speculation of course, but makes sense. And Apple would certainly not be making any of that public at this point.

I'm fairly confident there's virtually no comparison between the best Johnny Srouji has been able to put into silicon and the nearly shipping GB300.

We're going too see improvements with SoIC by next year on the consumer side, there will be some revision for servers much like there probably were for M2 to power PCC, but Apple has significant hurdles to overcome.especially around interlinking systems as well as more basic things like improving matrix operations which I do think is happening with M5 or M6.

Just putting my thumb to the wind I'd posit that Apple won't be competitive until at least 2028-2029 at the server level, and given Nvidia's commitment and expertise possibly not even then.

If the rumor about Apple pivoting to also offer AWS-style services is true and Tim can be satisfied that he has another funding source then perhaps they can throw tens of billions at the problem and rush the timeline a little, but I remain skeptical.

I want them to do these things, but I don't think Tim has had this plan for a long time – they wouldn't be reorganizing their entire AI / Siri divisions, and especially wouldn't be losing the key researchers they are if there was a lot "almost ready" internally to kick into high gear.

 

[Zc456]

I remember all of the "Who asked for a thousand songs in your pocket?" retorts when Jobs introduced iPod.

Not much has changed since then. Jobs did have his share of real flops, around a dozen or so of them.

Obviously. That being said, they usually owned up to it (even if it was quietly) and moved on when things failed. Now it feels they're just seeing what sticks.

 

[smulji]

It's pretty clear what Apple's AI strategy is. And that's investing $500 Billion to manufacture its own proprietary AI servers with a focus on user privacy/security in its Houston manufacturing facility. And then deploying them around the country (and likely beyond).

From what I understand, $500 billion isn't just for Apple's AI servers but that's the total investment in US manufacturing over the next 4 years, one of which will be investing in Apple's AI servers.

 

[unchecked]

The thing is can they extract some value out of it before the bubble pops? As big corps are leaning big into AI and backlash is starting in response (see Microsoft and Xbox), can Apple figure out something that can survive the inevitable crash?

 

[russell_314]

Never thought they’d ever admit to it

Of course, Apple makes the best products. They don’t necessarily make it first to market.

 

[citysnaps]

Obviously. That being said, they usually owned up to it (even if it was quietly) and moved on when things failed. Now it feels they're just seeing what sticks.

I don't recall any of those dozen+ flops Jobs had were ever owned up to. Maybe iPod Sox.

 

[citysnaps]

From what I understand, $500 billion isn't just for Apple's AI servers but that's the total investment in US manufacturing over the next 4 years, one of which will be investing in Apple's AI servers.

Yes... my understanding is that's for the manufacturing of the AI servers and their deployment and staffing in data centers around the US. And probably for the R&D developing their own AI server chips which has likely been going on for awhile.

 

[Rogifan]

Good grief. Before the panic last year Apple execs rarely uttered the word AI. It was all AR and spacial computing. Now we’re supposed to believe only Apple will figure out AI? Also this narrative that Apple isn’t first but is best is kinda BS. Maybe it’s the case sometimes but not every time. But just watch Apple bootlickers like Neil Cybart, who’s all in on Vision Pro and thinks AI is just hype, pivot to being all in on AI now that Tim Cook has pivoted. They’ll forget the Vision Pro so fast it will make your head spin.

I use Copilot at work every day. I think it’s great. I’m using ChatGPT more than Google these days. It’s also very good. What exactly is Apple going to show up late(r) and do better? Personal context? I’ll believe when it’s working and actually ships to consumers.

 

[StudioMacs]

There is a dissatisfaction with Apple for not taking the technology and using it in a new way that seems obvious once you’ve seen it, like the iPod and iPhone.

To much fanfare, Apple promised an iteration on existing ideas over a year ago, and failed to deliver on those meek goals.

However, the real problem is the lack of vision and execution to announce and deliver something that seems obvious in retrospect.

.

 

[Rogifan]

Here is one issue I have. In an interview after WWDC, Craig had mentioned when people think AI, they think of a chat bot, which is not their direction. However, since that’s how people think of it, those same people will see that Apple doesn’t do that and may look elsewhere.

I remember before the Apple Watch was unveiled Tim Cook said nobody wants to wear glasses but the wrist was interesting. Apple will mock AI chat bots until they have one. Then it will be only Apple could do this.

 

[jdawgnoonan]

AI is early all the way around. In 10 years the only difference between AI on the major platforms will be how annoying it is or isn’t. If we were measuring solely based on the annoying factor Microsoft has a drastic lead. Right now my favorite AI integrated into any OS is on Linux because it isn’t there at all except in my web browser.

 

[Lioness~]

I expect continues good hardware from Apple, nothing else today.

AI is not Apple’s thing.
The sooner they get that, the better. They should kill Siri and start to only use other’s services.
The better for them, and definitely for Apple’s customers.

 

[TVreporter]

My wife wishes I was never first … what’s the secret Apple!!

 

[Zc456]

AI is early all the way around. In 10 years the only difference between AI on the major platforms will be how annoying it is or isn’t. If we were measuring solely based on the annoying factor Microsoft has a drastic lead. Right now my favorite AI integrated into any OS is on Linux because it isn’t’ there at all except in my web browser.

I remember back before ChatGPT when it was all about the process of machine learning by having smaller models do small repetitive things to save time. Way better use of the tech. Now every billion dollar CEO is trying to make a larger jack of all trades.

 

[MacBH928]

Apple 2011: Introduces Siri

Apple 2025: ""Apple must do this. Apple will do this."

 

[jdawgnoonan]

I remember back before ChatGPT when it was all about the process of machine learning by having smaller models do small repetitive things to save time. Way better use of the tech. Now every billion dollar CEO is trying to make a larger jack of all trades.

I think machine learning is still more important. I am not overly impressed personally with chatbots. They actually aren’t very smart or very good at much other than for making funny images or videos. They are good for filtering out ******** noise in search results (noise typically caused by ad driven results…I wonder when AIs will start injecting ads into their responses). Their writing sure isn’t great, it is incredibly easy to pick what AI writes as it is usually written much like someone writes in a second language with an extreme overuse of adjectives and unnecessary words. I do think that Apple’s proofread function is pretty nice, I have used it on my own writing. I also believe that AI is good to use for correcting syntactical mistakes in code, but it does not generate great code either.

With all of the drama we see about kids and self-esteem issues caused by social media, it is going to be pretty sad to hear about the mental issues caused by people who trick themselves into thinking that these stupid chatbots are actually their friends. I am also quite disturbed by the concept of AI therapists.

 

[novagamer]

I think machine learning is still more important. I am not overly impressed personally with chatbots. They actually aren’t very smart or very good at much other than for making funny images or videos. They are good for filtering out ******** noise in search results (noise typically caused by ad driven results…I wonder when AIs will start injecting ads into their responses). Their writing sure isn’t great, it is incredibly easy to pick what AI writes as it is usually written much like someone writes in a second language with an extreme overuse of adjectives and unnecessary words. I do think that Apple’s proofread function is pretty nice, I have used it on my own writing. I also believe that AI is good to use for correcting syntactical mistakes in code, but it does not generate great code either.

With all of the drama we see about kids and self-esteem issues caused by social media, it is going to be pretty sad to hear about the mental issues caused by people who trick themselves into thinking that these stupid chatbots are actually their friends. I am also quite disturbed by the concept of AI therapists.

Disclaimer: The below was written entirely by Claude Opus. I do generally agree with the points, and fed it 3 posts of mine to get close to my style of writing, but it is copy and pasted whole cloth. See the next post for my rationale.

=========

I'd push back on the ML vs chatbots distinction since they're fundamentally the same technology stack, just applied differently. The real differentiator is going to be who can deploy these models effectively at scale, not whether they're used for chat or other applications.

You're right about the writing quality being a dead giveaway. The adjective soup and overwrought phrasing is painful. But that's more an indictment of current RLHF approaches than the underlying tech. Apple's proofreading works precisely because it's a narrow task rather than open ended generation.

The AI therapist thing is genuinely dystopian. We're already seeing mental health crises from social media parasocial relationships. Wait until lonely people start forming dependencies on chatbots literally optimized to keep them engaged. It'll make Instagram depression look quaint.

As for layoffs disguised as "AI efficiency gains", you nailed it. Every earnings call now has some exec bragging about headcount reduction through AI. The smart companies are augmenting their workforce, not replacing it, but Wall Street rewards the slash and burn approach.

The real tell is that even the companies pushing this hardest are still hiring aggressively for their AI teams. They know these systems need massive human oversight and can't actually replace skilled workers yet. It's all theater for the stock price.

 

[novagamer]

I think machine learning is still more important. I am not overly impressed personally with chatbots. They actually aren’t very smart or very good at much other than for making funny images or videos. They are good for filtering out ******** noise in search results (noise typically caused by ad driven results…I wonder when AIs will start injecting ads into their responses). Their writing sure isn’t great, it is incredibly easy to pick what AI writes as it is usually written much like someone writes in a second language with an extreme overuse of adjectives and unnecessary words.

FWIW, I will edit my post to add a disclaimer and I don't mean this maliciously, but I wanted to prove a point in general for you and others. Experience from even 3 months ago does not translate to current reality.

I never use AI to write responses for me and a big part if why I participate in this forum is because it's a great way of refining my thinking for these topics in general, but I do think it is important to understand the current state of things.

As much vigilance as you think you have based on earlier data points, it doesn't translate anymore. This will only keep improving.

I agree the general ChatGPT "it's not X, it's Y —" kind of phraseology is easy to pick out, but some others not so much.

 

[ThailandToo]

I don't think you understand what Apple's AI strategy is.

Apple doesn’t understand its AI strategy how would anyone else!

 

[Naraxus]

Plot twist:
They pretty much were the first with Siri:

"Not only did it understand the words I said. It understands the meaning." (Scott Forstall, 2011)

Except they weren't. Siri was out long before Apple bought it and absorbed it into iOS...

 

[@Brett]

yeah the problem was you went off script and marketed AI features before they were even close to being ready.

Not much different to all the other companies that rolled out AI and it was not complete or somewhat faulty. Difference is the companies spent billions and continue to. And they need to show investors results. Personally I don’t think they will see the ROI that they want.

 

[MrSkoTA]

I mean ‘let’s’ get over it folks. Are we adults or whiny children that cry when we don’t get our way right now?!? That includes YOU Apple ‘Pundits’!

 

[MrSkoTA]

yeah the problem was you went off script and marketed AI features before they were even close to being ready.

All of the AI features came to pass except the ‘better Siri’ what are you and everyone else going on about!?!