Apple Chip Chief Johny Srouji Could Be Next to Go as Exodus Continues

[Brother Cavil]

But I also don't really these AI features integrated into the operating system. What Microsoft does and plans for example is a reason to move away from Windows for me.

For you (and me as well, for the record). But not for most people.

And so far I have not seen any of these tech companies deliver much impressive stuff, mostly only "stochastic parrots". Especially MS, all their Copilot offerings are so bad that it hurts.

I don't think you're being fair. Google has done some very impressive work. Apple is nowhere near having a conversational AI agent that can book flights, make dinner reservations, etc. I would argue that a lot of this stuff is still more proof of concept at this point, but it's clear that we're rapidly approaching the tipping point. I have a coworker who almost never types on her Android phone. She's always talking to it. We're not very far away from another paradigm shift.

I could rewrite your comment, set it in the late 1990s, and complain about the internet. And yet look at where we are today. So you're not impressed yet? That's fine, but don't kid yourself. This is the future. You'll come around at some point, just like people who couldn't see the future promise of the internet in the late 90s did.

Apple is woefully behind when it comes to this next paradigm shift. That's why I think they should buy Anthropic. They will never catch up without a major acquisition. Many people, myself included, believe the AI space will consolidate in the next few years and there will only be a few major players when the dust settles. Someone is going to buy Anthropic and I would say they are most aligned with Apple's principles.

Right now it looks like Apple might try to rely on Google for AI and Siri which would be disastrous long-term. Instead of staying competitive, Apple would cede control of this transformative technology to a third party, ultimately guaranteeing Apple's own irrelevance once day. When someone, maybe Google, maybe OpenAI, ships a product that is fully conversational, how many customers will jump ship?

Apple is one product away from being Blackberry and, as the iPhone goes, so goes Apple. They're bleeding talent and, worse, they're not seen as a good place to work if you're trying to build the next big (AI) thing.

 

[akula1488]

maybe he just likes Samsung's tri-fold phone too much

 

[RedWeasel]

My point of view is probably influenced by my day job, which is in the AI field. I'm not in research (far from it), but basically in deployment and providing AI to users in a large enterprise.

So while I don't have full insights in the latest and hottest research, I think I have a pretty good overview of what and how AI actually is getting used in enterprise scenarios. And I don't mean what managers tell each other, but what is really used. And I can only repeat that I'm not very impressed.

There are some nice and useful areas, sure. Text recognition is one for example, where we often get much better results than with "classic" OCR. But all these efficiency gains etc. that are getting promised are vastly exaggerated from what I saw so far.

Yes, you can use AI to write your mail in a more flowery language or to generate images for a PowerPoint presentation. But none of that has any actually positive impact on the bottom-line.

Now don't get me wrong, I think AI is a fascinating field and we will see a lot of progress in the future. But the problem is that most of the stuff that is hyped today are LLM, and while LLMs might be a part of future AI, they are not enough. But they do produce results that look fascinating, even if they are wrong.

I don't think you're being fair. Google has done some very impressive work. Apple is nowhere near having a conversational AI agent that can book flights, make dinner reservations, etc.

None of the major AI providers are nowhere near to having reliable agents to do these things. That's why I don't get the hype about agentic use (well I get it from a sales perspective of course). If these tools are so unreliable in interactive use, why would you want to send them out and let them to stuff "own their own"?

I could rewrite your comment, set it in the late 1990s, and complain about the internet. And yet look at where we are today.

That's an excellent point actually, because I think that's exactly where we are, in 1999 or 2000, shortly before the dot com bubble burst. Was there an Internet after the dot com bubble? Of course, and one could argue that things only really got started afterwards. And I think this will happen with AI as well.

Many people, myself included, believe the AI space will consolidate in the next few years and there will only be a few major players when the dust settles.

Fully agree. And to make matters worse, none of these companies have any kind of economic moat at the moment, i.e. they are completely replaceable by competitors. ChatGPT is not the best anymore? Easy, just use Gemini. There are no network effects like in a social network, no ecosystem buy-in like for example with Apple, there is nothing.

 

[Uncle Shadoe]

apple is sinking

we need a complete renovation, a new CEO. get rid of Apple Intelligence, Vision Pro, iPhone air, iPhone fold.

focus on making good devices.

I agree with you about what their focus should be on. However, Apple just created the most awesome iPhones back to back — iPhone 16 and iPhone 17. Apple Intelligence is a waste of time until the technology matures. They jumped into the market too fast.

 

[Brother Cavil]

My point of view is probably influenced by my day job, which is in the AI field. I'm not in research (far from it), but basically in deployment and providing AI to users in a large enterprise.

So while I don't have full insights in the latest and hottest research, I think I have a pretty good overview of what and how AI actually is getting used in enterprise scenarios. And I don't mean what managers tell each other, but what is really used. And I can only repeat that I'm not very impressed.

There are some nice and useful areas, sure. Text recognition is one for example, where we often get much better results than with "classic" OCR. But all these efficiency gains etc. that are getting promised are vastly exaggerated from what I saw so far.

Yes, you can use AI to write your mail in a more flowery language or to generate images for a PowerPoint presentation. But none of that has any actually positive impact on the bottom-line.

I think we agree on a lot and I appreciate your perspective from a deployment point of view. I've read a lot of articles about AI not living up to the hype in the enterprise space, so I'm not surprised by what you wrote. That said (and I think you'd agree), I don't think that says a thing about AI in the long run.

Now don't get me wrong, I think AI is a fascinating field and we will see a lot of progress in the future. But the problem is that most of the stuff that is hyped today are LLM, and while LLMs might be a part of future AI, they are not enough. But they do produce results that look fascinating, even if they are wrong.

Well, you have to start somewhere, right?

None of the major AI providers are nowhere near to having reliable agents to do these things. That's why I don't get the hype about agentic use (well I get it from a sales perspective of course). If these tools are so unreliable in interactive use, why would you want to send them out and let them to stuff "own their own"?

100% reliable? No. 50% reliable? Likely. Ten years ago it was 0%. I completely get the hype. Based on the trajectory we've witnessed so far, we should get to 95% reliable in a few years. This is a transformative technology even if it's far from perfect today. Look at how many people now have "relationships" with AI chat bots.

That's an excellent point actually, because I think that's exactly where we are, in 1999 or 2000, shortly before the dot com bubble burst. Was there an Internet after the dot com bubble? Of course, and one could argue that things only really got started afterwards. And I think this will happen with AI as well.

100% agree. When we talk about a bubble, it's really more from a financial perspective than anything else. It's not like all of the internet tech of the late 90s got obliterated when the bubble burst. Tech survives bubbles.

Fully agree. And to make matters worse, none of these companies have any kind of economic moat at the moment, i.e. they are completely replaceable by competitors. ChatGPT is not the best anymore? Easy, just use Gemini. There are no network effects like in a social network, no ecosystem buy-in like for example with Apple, there is nothing.

This is an excellent point and I've heard it discussed a few times. I'm putting my money on Google when it comes to the UI/UX of the future. They're making a lot of progress and they're building high quality hardware these days. It's going to take a few more years, but when Gemini gets fully baked into Android, I think it will create something of a moat. We'll see, but for now I completely agree with you on this point.

 

[baconcat_8964]

MacRumours, don't make your website as low as YouTubers like Max Tech, making random rumours every day and hoping one of them works.

 

[RedWeasel]

Based on the trajectory we've witnessed so far, we should get to 95% reliable in a few years.

That’s the million dollar question right now (or rather, the I-don’t-know-how-many-billions dollar question).

How long will the improvements we saw during the last few years continue with the same rate? Will there be a plateau? Are we close to that plateau already maybe and is the problem inherent in LLMs and not only a matter of throwing more compute power at it?

And I agree, if there is a company that is well positioned at the moment, it’s probably Google.

 

[thekeyring]

Hmm. Vision Pro is a flop lets be honest even a cheaper lighter future iteration is unlikely to be the next big thing.

AR glasses are interesting but all we have from Apple are rumours, Meta is actually shipping a product. Sure Meta's product isn't earth shattering at the moment but they are further along than Apple are.

Last i heard about Apple's AR glasses effort was that they were hoping to leverage Siri (which doesn't really bode well) and they would be more of an iPhone accessory.

Without banging the 'what Steve would've done' schtick too much but Jobs was willing to kill the iPod so that they could launch the iPhone. There is absolutely nothing about the current leadership at Apple that suggests that they would be willing to cannibalise iPhone sales, we are seeing all of the articles now about senior Apple execs describing them as risk averse. Srouji will be seeing all this at close quarters.

It's classic innovators dilemma.

Yeah, all we have is rumours, but inside Apple they have products but they don't ship until they feel they are ready. Meta gets more credit because it's showing their working. Like Google being considered innovative for working on driverless cars 15 years ago.