Apple Will 'Watch Everything Burn' When AI Bubble Bursts - Ed Zitron

[rgwebb]

He's not saying that AI is going anywhere. He IS saying that the spending that Open AI is doing isn't sustainable and this bubble of trying to monetize AI is something that will burst. AI isn't going to go anywhere, but this grifter phase of AI where people are trying to figure out how to make a quick buck out of it is going to end. It's just like the Dot Com bubble. The internet didn't go anywhere, but the bubble did burst and it did end a lot of early web businesses, and eventually the internet evolved to something more useful. But to think that what is going on with the AI businesses is going to continue like we're currently seeing is insanely naive. These people pouring billions of dollars in these companies is not going to keep doing so if there's no return on their investments

Right, I think his analysis is spot on. I do think the over-investment into AI is a leading indicator of the dead-end US tech is facing. These are useful technologies but not quite as transformative as advertised. Zitron made a salient statement that US tech faces an issue where hyper-growth is really not possible because the technology is not there for transformative human interfaces.

This is why big tech has championed AI so aggressively with dubious claims (charitably) to outright lies directed at investors. The lending environment that allowed tech to explode over the 2010s culminating with COVID in 2020 has dried up. They needed something that would lure investment and nothing entices like the idea that businesses can just stop paying for skilled labor (or pay significantly less).

 

[diewish2007]

Honest question: the depicted scenario takes into account that Transformer LLM based economy crashes itself in the long (well, not so long) run. What if a new AI technology takes hold like, let’s say RNNs (Recurrent Neural Networks)? Apple’s paper shows that ParaRNN brings a lot of efficiency in training, making them, for many tasks, a better alternative.
So, to expand on the article: the world can only support one or two transformer LLMs which do dedicated general knowledge tasks, and don’t need no way near as much compute as they have now. The rest will be RNNs, running on device and decentralized. Apple is in an incredible advantage position in this regard. I hope they stick with their RNN guns.

I agree. I think that, for a while now, it has been very clear that the future of AI will be on local models that do not require to connect to the internet to function in order to do basic things.

 

[frogslapper]

that's a load-bearing when

Reading this was like seeing a glitch in the matrix.

 

[spaxxedout]

Ed Zitron is just another talking head in front of a webcam. He isn’t qualified in any way to talk about almost any subject he covers. He has no professional expertise in deep learning or building any tech companies or useful applications.

AI in all its forms is here to stay and will be embedded throughout many applications, devices and appliances. It’s not going to vanish. Talking heads with webcams said bitcoin was going to go to zero. If something silly like bitcoin is worth nearly 70,000 grand today then AI which is used by half of humanity is not going to “burst” and vanish.

Ed will be long gone one day. AI won’t be.

One really needs to ask if any of these influencers and talking heads manipulate markets or have short positions in the things they speak against.

Thank you. He’s a complete grifter. The same could almost be said about Doctorow, but at least Cory’s intelligent. I don’t mind people disagreeing with me about AI. I mind the lack of sincerity or the refusal to accept its capabilities. This is the most prima facia technology ever.

 

[DaveCummings]

How long was Amazon in the red? For how many years did they focus solely on growth?

I think only three AI providers will survive:

OpenAI, Anthropic, and Gemini.

How many active users do they have?
How many would be willing to pay 10–20€ per month if free access were increasingly restricted or eliminated?

All these people are hooked on AI for now, but what about later?

Who wants to go back to Google Search and wade through 10 pages of results?

Don’t forget:
People pay ridiculous prices at Apple for iPhone cases and stuff.

I have no doubt that Gemini will stay because it's Google and Google has more than just AI behind their business and I think Anthropic might survive because they somehow managed to find a way to actually make some kind of profit, I honestly don't think OpenAI will survive. They've burnt through way too much money and like I think that Tesla and SpaceX will eventually get severely damaged if not destroyed due to their tight association with Elon Musk (seeing how much those stocks tanked, it might already be happening) , I think that to a lesser extent keeping things so heavily associated with Sam Altman will damage the company more than help it.

 

[Troff]

Nice interview!

I think one thing that is still missing to me, and nobody talks about, is that whatever the fate of AI is, hyperscalers will own pretty much all the hardware out there.

On the other hand, I think customer market for computers will not recover after this, we will never see 64G of ram for $120 and prices will never go down.

All together, makes me fear a future where hyperscalers will try to sell us simply terminals under subscription to being able to access the computing power on the Cloud. I hope not, but all is going into that direction.

He's talked about this off and on on his podcast and in his newsletter. Forgive me, for I am a bit Ed-pilled, but he does reflect my thinking on a lot of this and even when I disagree with him, I seem to pretty much agree (sure, that makes little sense).

He has had Cory Doctorow on a couple times -- or they've spoken together a couple times on record, anyway -- and Doctorow sees the looming AI bubble a bit more hopefully, using the analogy of all the fiber that was left over after the first dotcom boom and bust that at least left everyone with a path to ubiquitous hi-speed bandwidth. And one the current bubble bursts well, at least there will be an infrastructure that can be used by the smart and savvy for...whatever. I think his new book covers this idea, I haven't read it yet. Ed always counters with asking what sort of infrastructure will be left behind, he thinks it will be pretty useless in general or aged out or just not usable for anything practical. The gobs of nvidia hardware will be worn out and/or not applicable to anything but LLMs or whatever. And all the phantom chip orders will disappear, leaving us with no real new stuff for market, and what does exist won;t be salvageable (I guess it would be a good time to learn microsoldering and desoldering).

I myself have no real idea except I do think that you're right, the market for computers will be beyond recovery. I feel like I'm typing this on hardware that is going to last me for 10 or 15 years, and will eventually be replaced by something less capable, more locked down, and vastly less engaging.

 

[mctrees02]

He's not saying that AI is going anywhere. He IS saying that the spending that Open AI is doing isn't sustainable and this bubble of trying to monetize AI is something that will burst. AI isn't going to go anywhere, but this grifter phase of AI where people are trying to figure out how to make a quick buck out of it is going to end. It's just like the Dot Com bubble. The internet didn't go anywhere, but the bubble did burst and it did end a lot of early web businesses, and eventually the internet evolved to something more useful. But to think that what is going on with the AI businesses is going to continue like we're currently seeing is insanely naive. These people pouring billions of dollars in these companies is not going to keep doing so if there's no return on their investments

Exactly.

The dot com bubble burst and a few companies (most notably Amazon) survived...but along the way, there was this quirky, little company called Google that neither rushed a mediocre product to market nor chased a quick IPO windfall for employees and investors.

Instead, they focused solely on making web search fast and accurate. Once they released AdWords in October 2000 and AdSense a few years later, they had created a high margin, cash printing machine long before their IPO in 2004.

Fittingly, Google appears to be taking a somewhat similar approach to the AI craze. They don't seem to particularly care who wins and loses in the software development because they're more focused on playing the long game by building the necessary infrastructure to service the winner(s) after the dust settles.

Likewise, Apple seem to have realized that consumer AI tools are a long ways away from being a worthwhile investment venture and so there's little need for them to rush to shove AI stuff onto their devices. Instead, they'll quietly partner with Google to support new Siri while continuing to sit on their mountain of cash until they find the right acquisition targets who can add value to Apple's offering.

IMO, Apple are under no real pressure to rush into the AI wave given that they've returned over $900b to their shareholders during the Tim Cook era.

 

[Taipan]

Oh, yes, please! We won‘t get rid of AI, but at least some of the unpleasant side effects.

 

[CarAnalogy]

Nice interview!

I think one thing that is still missing to me, and nobody talks about, is that whatever the fate of AI is, hyperscalers will own pretty much all the hardware out there.

On the other hand, I think customer market for computers will not recover after this, we will never see 64G of ram for $120 and prices will never go down.

All together, makes me fear a future where hyperscalers will try to sell us simply terminals under subscription to being able to access the computing power on the Cloud. I hope not, but all is going into that direction.

This is a great point a lot of people miss.

It seems to be a big part of this plan is to finally take any semblance of power away from the users.

No owned hardware, no local processing that can’t be monitored and controlled, everything done via something that decides whether you’re allowed to do it or not.

I hope everyone likes Claude’s attitude problem lately where it constantly argues with you and tells you to take a break. That’s their vision of all computing.

 

[kyjaotkb]

So every AI company will burn to the ground and the last thing standing will be Apple Intelligence and genmoji?

And Filemaker Agentic Development.

 

[DogGone]

Isn't this the way tech always goes. A new thing and everyone jumps on the bandwagon. The only difference is the scale of the investment in AI infrastructure.
As with a startup, a lot of funds or bets need to be made. The people making the money at the moment are those involved in building the structures and making the hardware needed.
The other people who are making money are the ones convincing others to invest.
Its always the case.
So I may have to find out how deep my retirement funds are in with AI.
But the AI bubble will carry on for a while unless there is some credit crunch event that causes everything to crash down.

 

[neotint]

Actual journalism! More of this please instead of copy/pasting corporate press releases.

I've long said Apple, if it's wise and stays grounded thanks to its massive capital reserves, can leap frog the whole sector, hold out and let this stochastic inherently error-filled transformer LLM era collapse as this bubble will since it's inflated up by speculation, misdirection, illiteracy, dumb money, and hype/PR.

 

[Macalway]

So I guess we should feel good about Siri?

 

[abowlby]

So Apple continues to invest in AI and paid Google a king's ransom for the Gemini models to try and fix their own AI product, which at this point also includes Siri (a once flagship feature), but have no interest in the game? They are just splashing the pot for the hell of it?

 

[DrJR]

I feel like this guy misses the point that the rich and all these companies all know its a farce. (much like many of their schemes and scams) Just like so much of our economic and business systems etc. They're all based on nonsense, hopes, dreams, farts. Capex this, capex that. As long as stonks go uppity and investors are fat that's all that matters. They will inflate any bubble. Sure it might burst here at some point, but don't act like that wasn't the plan all along. Oh and guess what, its not the end of the world! That bubble bursts? Money shifts around, industries shift around, and the rich continue to get richer. A new world and new opportunities will be born out of the burst. Just like the .com burst. Just like all the other bubbles that have popped. AI is here to stay, and its been here before these rich dorks drowned us in marketing about it.

Your quality of life also improves while you sit in the sidelines. Don’t forget that.

 

[nikhsub1]

I'm old enough to remember the dot com bubble and what happened when that burst. AI is 100% here to stay but the frenzy around it and the spend is reminiscent of what happened at the very end of last century.

 

[the future]

It’s really not. I just asked it a question and it completely whiffed the answer. ChatGPT and Grok both nailed it.

n=1

 

[DrJR]

I posted a long time ago on here that Apple is actually quietly in front on AI, not behind. Most people don’t realize “talking” computers have been around 40+ years. (Look up John Sowa). They are a hardware company that has been building their systems for ML for over a decade now. If anything they are adding too much AI into the OS in order to appease, but they fully recognize their value is to provide “just enough AI” with more than enough security. Information security, user privacy, is their schtick.

As far as bubbles the stock market has been in the precipice of collapse my whole life. I watched 2008 in real time advising people to not bail out. Never bet against the good ol USA! 🇺🇸

 

[MiamiFL]

This article is nothing but doomer-ism claptrap by brainwashed individuals seeking an infinite source of copium rather than accepting reality -- AI isn't going anywhere, it's only going to keep growing.

Once you discover all the anti-AI/anti-datacenter nonsense is nothing but a psyop being promulgated by our biggest global competitor in the world market in order to get us to kneecap ourselves so they can achieve a lead that we could never hope to catch, the game is given away. As with all disinformation campaigns by said international rival, it relies on uninformed sheeple to prop up and push the false narrative.

 

[Unregistered 4U]

I can't even use mine anymore because it needs an update that requires you to wear the thing the whole time. So much promise, released too early, shoved out the door by a CEO on his way out.

This is like reading a seemingly thoughtful insightful piece that appears to connect a lot of dots with what appears to be reasonable data and facts… then the very last line in that piece being “Also, no one understands how magnets work anyway.”

It’s like… what even did he actually mean? To see the images you have to wear the thing the whole time? That’s sort of how it works, sir. At that point, it feels like, “How can I string the right words together that will get me attention?”

 

[macbookj0e]

Ed Zitron is just another talking head in front of a webcam. He isn’t qualified in any way to talk about almost any subject he covers. He has no professional expertise in deep learning or building any tech companies or useful applications.

AI in all its forms is here to stay and will be embedded throughout many applications, devices and appliances. It’s not going to vanish. Talking heads with webcams said bitcoin was going to go to zero. If something silly like bitcoin is worth nearly 70,000 grand today then AI which is used by half of humanity is not going to “burst” and vanish.

Ed will be long gone one day. AI won’t be.

One really needs to ask if any of these influencers and talking heads manipulate markets or have short positions in the things they speak against.

I get your points but it doesn’t mean AI isn’t going to “burst”.

Dot coms are still around but that bubble burst. I’m sure AI is here to stay but it doesn’t mean it’ll be a smooth ride.

 

[bigboy29]

Thank you. He’s a complete grifter. The same could almost be said about Doctorow, but at least Cory’s intelligent. I don’t mind people disagreeing with me about AI. I mind the lack of sincerity or the refusal to accept its capabilities. This is the most prima facia technology ever.

AI can be the most transformational tech for a very long time and can be here to stay.

AI spending can be totally out of control, and not every company heavily investing will necessarily make it. Today's winners are not guaranteed to be long-time winners.

Both statements ^ can be true at the same time.

Over-investment into fiber eventually worked out; but the landscape of winners and losers has changed dramatically in the mean time. Most companies that over invested went under. With AI models being more and more of a commodity, it is very reasonable to ask yourself (if you invest in this space) just how much investment is too fast and too much.

 

[capamac]

I don't think they should call it "AI" when it wasn't too long ago this "AI" would confidently proclaim there are two "r"s in "strawberry. It's wrong so frequently.

 

[the future]

Isn't this the way tech always goes. A new thing and everyone jumps on the bandwagon. The only difference is the scale of the investment in AI infrastructure.

But the scale is the decisive difference. This „bet“ on AI is much, much bigger than all other bets before.

 

[gleepskip]

How long was Amazon in the red? For how many years did they focus solely on growth?

I think only three AI providers will survive:

OpenAI, Anthropic, and Gemini.

How many active users do they have?
How many would be willing to pay 10–20€ per month if free access were increasingly restricted or eliminated?

All these people are hooked on AI for now, but what about later?

Who wants to go back to Google Search and wade through 10 pages of results?

Don’t forget:
People pay ridiculous prices at Apple for iPhone cases and stuff.

Gemini? I think you misspelled xAI.