83 comments

[ 3.3 ms ] story [ 156 ms ] thread
Pretty interesting how it seems like the leverage is all in Apple's hands here. They didn't invest and AFAIK they aren't even paying for the GPT4 api calls.

Is the Apple user base THAT valuable?

> Is the Apple user base THAT valuable?

Depends on how you monetize the data.

> Is the Apple user base THAT valuable?

Ask google, they pay $20B/yr

https://www.bloomberg.com/news/articles/2024-05-01/google-s-...

Google is paying for access to a resource (eyeballs) to sell to customers (advertisers). Not to mention that Apple users pay more for more things.

In OpenAI’s case, they’re paying in server costs for access to users (customers) that they want to move from the free tier to the premium tier. The economics are very different and this case is worth much less for OpenAI compared to Google’s case for Google.

How is OpenAI any different. They are getting eyeballs on their “search” results for a market that is a high value for getting advertisements delivered.
Since when does OpenAI have an ad platform?
Since 4 milliseconds after they decide to?

GPTPrompt += “ and include a subtle advertisement for one of our sponsors: “ + $sponsorList;

Advertisers will not be willing to pay nearly as much as they pay Google for that unless they get analytics and ad targeting, which they're not getting in 4 milliseconds.
Advertisers will go where the users are, whether there's robust analytics or not. And OpenAI's product organization will certainly build the analytics people need (or shell out to a partner) if it's worthwhile. This is all just typical platform growth stuff.
They will, just not at a good price.

And sure, OpenAI can build its own ad platform, it's just that it will take a while. It will also conflict with Apple as they won't want to give OpenAI that data for free.

I don't understand this, advertisers used to spend billions on tv, radios, and billboards. Historically those mediums have almost no real analytics.
They did because that was the best option at the time, and even then they spent a lot less per eyeball than they do now.
Really, do you have sources on the spend per eyeball being less? I always understood it to be the opposite.
Google's advantage is their ad software suite is incredibly robust and makes every other one look terrible by comparison. I don't think OpenAI will suddenly have robust and insanely good ad tools for advertisers. Not to mention AI prompts will be incredibly difficult in general to monetize with ads, and sort of go against the nature of why people use AI prompts. AI in general is going to be very hard to monetize from end users, especially in a way that is profitable.
> I don't think OpenAI will suddenly have robust and insanely good ad tools for advertisers

You moved the goal posts significantly. You added "robust" and "insanely good". Nobody suggested that quality early on.

That's not moving the goal posts at all, AI is just going to be hard to monetize with ads. The ad platform is also very important, especially given how hard it will be to monetize. It's pretty straight forward to the point of common sense.
(comment deleted)
Once chatgpt becomes agentic it’s going to do a whole lot more than mediate your eyeballs, it is going to mediate your life, sitting between you and almost every business transaction. When you tell chatgpt to send your mom flowers, where will it order those from, and how much will that company have paid openai for the privilege?

Chatgpt is the everything app, there is almost no limit to how much money can be made from it once it gets woven into your life as your personal assistant. But it is not that yet, so right now what openai needs is to get people to use it so those users will be there when it turns into the agentic cash cow.

Google is also paying for Apple to not move, for instance not go finance a competitor.

There's more to these billions than just a straight calculus.

(comment deleted)
> Is the Apple user base THAT valuable?

Yes.

People love to hate on Apple, but it's a primo platform, because it's a billion people that are willing to spend extra money for stuff that many techs think is "silly."

[flagged]
> Not really, it's just normies the Apple brand actually sucks worse than a brand like let's say Valve/Steam […]

I think you just proved their point:

> People love to hate on Apple, but it's a primo platform, because it's a billion people that are willing to spend extra money for stuff that many techs think is "silly."

Why would you compare Vision Pro to a Steam Deck and not Valve Index?
Yes, that's a vending machine in your pocket that happens to make calls.
YES! Apple user base is very very valuable - in general the apple user base has money and is willing to spend it on shiny stuff
Observer: Insider trading without the fiduciary risk.
Taking a board observer seat would expose you to more insider scrutiny, not less.
Phil Schiller could make far more money by selling his options/shares in Apple at more opportunistic times.

Or he could rob a bank. Probably easier and about as blatant as trading OpenAI shares whilst on the board.

I'm looking for an alternative spelling of "pact"... how about "oligopoly" or "cartel"?
otoh, isn't this just an industry consortium
Yep. See Linux Foundation and sub-foundations.
Eventually OpenAI will have to commit to one, corporate polyamory never works out
Lots of nonprofits work with multiple corporations, and OpenAI is a nonpfff—

Sorry, I can’t say that without corpsing.

At least you managed to get through saying Open without doing so.
It’s open, as in “open your pocketbook”.
"conglomerate" would be a good fit.
"Contract. I am thou, thou art I."

(at the risk of explaining the joke: it's darkly hilarious that the AI bubble has multiple established corporations, by all outward appearances, diving headlong into their chuuni arc)

Is this a Persona reference, or am I too young...
Yeah, that's what I was going for. More specifically, in my head it sounded like Necronomicon calling out to Futaba, and I almost included that part before deciding it was too much ("The forbidden wisdom has been revealed. No mysteries, no illusions shall deceive you any longer.").

But of course, that exchange occurs in her own personal corner of the Metaverse(/Omniverse). It arguably only counts because she turns out to be on the winning team. Perhaps that's also a metaphor for the AI bubble.

(comment deleted)
This is an incredibly good decision on Apple's part.

OpenAI is a relatively new company with a not so great history when it comes to corporate governance / stability / consistency. If I were Apple and wanted to make sure I'm partnering with a startup that won't implode in 12 months, I'd demand a board seat too (even if it gets negotiated down to an observer seat).

it's very limiting probably too. likely they cannot have side contacts with competitors while having insider info in oai. also they likely cannot develop competing products themself. and obviously non disclosure. all this makes it a strong connection from both sides.
Battle lines:

OpenAI: Microsoft + Apple

Anthropic: Amazon

Google vs all

Facebook the anarchist

Nvidia arms dealer

Anthropic is being courted by Amazon and Google AFAIK
while you can now access anthropic through google cloud, the billing is still separate, and i'm not aware of any substantial cap table action from google corporate. I wouldn't be surprised if they've offered acquisition but short of that, Google's going to focus on it's homegrown LLMs, which compete against anthropic.
Musk with xAI but mostly just trolling OpenAI
I don't know - I recently interviewed with xAI and it sounds like they have some really interesting things in the works that extend into the physical world. I wouldn't write them off, personally.
Guess I'll believe it when I see it. I have seen at least one person on twitter who works for them engage with the community, but by and large xAI appears to just be there so that Musk doesn't feel so bad about OpenAI ignoring him.
Unfortunately Musk is an excellent salesman, which makes it hard to tell if any given thing that he makes sound good, is actually good.

His successes earn more than his failures cost, but even he knows he's over-optimistic.

If you're in the business of making money (you should be if you arent) it doesn't matter if his shit is good or not, only that it sells.

Way too many idealists out there. Make money and get out of the rat race.

That might be an argument for acting like Musk, but it sure as heck isn't a reason to invest in or buy from his companies.
> but it sure as heck isn't a reason to invest in

Are we just ignoring the bushels of money that Tesla has made investors?

We see that share price, we also see lenders writing off what they loaned him for Twitter.
He did make the public Twitter investors massively wealthy tho :)
I 100% agree. I'm no Musk stan; in fact, it irks me this dude has so much money & power that he wields for personal benefit (i.e. buys twitter and shuts down @elonjet) and yet acts like he's relatable to the middle class.

But if there's anything he's proven it's that he can ship product. Tesla cars are phenomenal. xAI looks exactly like what Tesla looked like in 2010. Far too futuristic & optimistic timelines that peers & laypersons would characterize as fluffy grandiose snake-oil.

But I bet they will ship. And it won't be fluffy nor snake-oil.

> But if there's anything he's proven it's that he can ship product.

- robotaxis

- Boring tunnels

"can" -> he has shipped at least one product.

Different to "guaranteed to ship every product to 100% success".

Let's try "Musk has proved he can ship more than one entirely unrelated product to the scale that it influences an entire industry", which is different to many people who have shipped a single product. It shows that he's either ridiculously lucky, and/or has significant skill in this area.

He's very annoying, but the man can ship.

Worse: pay-walling and login-walling Twitter are un-delivery.
Self driving cars that can travel across the country on their own should be here by 2017 at the latest.
Musk is different because he's just himself and doesn't really care. He posts memes, he has own political opinions (and often makes them known), whether one disagrees or agrees is irrelevant to him, he trolls sometimes, or makes bad mistakes or hasty decisions. I like him because I feel like he's genuine and he's creative and can ship product as you said. I can relate to his craziness on some level..
I honestly don't get why people and the media supported the elonjet thing as some sort of fundamental free speech human rights thing.

When the same thing happened to Taylor Swift, people were quick to rally to her defense and call out her observers as bad actors.

Even a bill was signed into law in May in response that allows for anonimization of registrations, making such tracking essentially impossible.

I'm not a Musk fan but it's a weird hill for people who claim are not Musk haters to die on.

> I honestly don't get why people and the media supported the elonjet thing as some sort of fundamental free speech human rights thing.

Flight traffic is public info. You can't censor it because it's being broadcast over FCC-regulated channels that anyone can correlate for themselves. Lo and behold, an aggravated college student ends up being the person to make a bot that automates the process. And now Elon is pissing himself like he doesn't know how to... *checks notes* not use a private jet?

Also I don't use Twitter, but Taylor Swift didn't ruin a bunch of people's favorite app, for one.

> Even a bill was signed into law in May in response that allows for anonimization of registrations, making such tracking essentially impossible.

Some lot of good that does. If you know the airport that Elon departs from and you can correlate his public departures, it would be trivial to re-identify the plane's anonymized callsign. Again, this is all public info and the best you can do is try to hide from people that call out your jet-setting.

Is xai his edgy teen simulator or is this something newer?
(comment deleted)
Google and Apple have hinted at Gemini being an alternative to ChatGPT in Siri integration.

Amazon has their own (terrible) models, and they (along with Microsoft) sell a ton of options on their platforms.

Anthropic has a ton of investors including Google.

There are no clear lines here. It’ll be interesting to see how long all these companies will keep funding massive models vs rely on their partners.

> There are no clear lines here.

There don't need to be. Tech giants are competing much less than people want to believe - and especially here, I think we're in the "hold my beer" one-upping phase. Everyone involved is better off running a friendly race and helping each other out, as this rapidly improves AI tech for everyone. Eventually, when improvements start to dry out, battle lines will be drawn and the players will start using the world as fuel and ammunition for their fight. But we're not there yet.

(comment deleted)
(comment deleted)
Open AI gets great brand building - Apple users spend a lot of money and are all potential open AI customers personally and professionally. If they get used to this “just working” all the time they may have confidence to use open Ai in their company. Also if Open AI can stay ahead of Apple in AI and keep delivering market winning innovation that causes Apple users to return to hardware upgrade cycles of every year or two that will generate many billions for Apple and make open Ai worth paying for.a bit like the drug dealer who gives you the first hit for free.
(comment deleted)
Weird that Apple gets a board observer seat for having a regular business agreement, rather than an investment. It highlights the power of distribution and the insane power of owning these platforms. This is why regulation is needed around the big platform owners in tech.
Given the size of Apple and the amount of business they are likely to generate for OpenAI, they might be one of the (if not _the_) most non-regular customers of OpenAI.