Investing Thread

[jficke13]

Yep, largely agree.  I think a lot of money has flowed into AI on the promise of cost savings and no one wants to be left out.  But I think that there is a lot more development that has to happen for it to be practical for worker replacement, etc.

I think in all honesty a decent amount of the current AI bubble, and yes I think it's a bubble, is the same buzz word chasing that lead every tech company to have a metaverse-nowait-blockchain-nowait-crypto-nowait type operation. Go back and all of the big CEOs are giving interviews about how transformative each of these technologies were going to be only to quietly kill or drastically scale back their programs once the Wall St hype cycle had passed by. To some extent, AI is like this. Every johnny come lately company talking about how they're going to integrate AI and blah blah blah is just trying to get a hit of that sweet hype cycle share price improvement before moving on.

On the other hand, the infrastructure investment spending in the data centers and chipsets is very real, so that represents a much more tangible investment than a few press releases and engineer hires that allowed companies to claim they were using the blockchain. At the end of this cycle someone is going to have a monumental amount of data processing technology and infrastructure, whether it's used for AI or something else remains to be seen. To that end, I think AWS/Google/MSFT come out with the physical tools necessary to do... something...

But Silicon Valley is building out and hyping these tools as though they will be ubiquitous for the average person, the way Facebook sold itself as a tool that every single person on earth needed and wanted (and for a while, they grew into something that approximated that sales pitch). But for generative AI to be as ubiquitous as Facebook at its peak, it has to be as useful for everyone on earth as something like Google search is (or, was, it's kinda crappy now)... but it's not. For the average person, right now at best it's a novelty, and at worst it's doing a graygoo to the internet by scraping nonsense reddit troll posts, regurgitating it as search results, and then scraping the trolling about those bad results creating a recursive feedback loop of bad data in that's poisoning the models.

(Ethical problems created from the inherent plagiarism of their data sets aside) I have seen interesting applications of image generation technology used by authors in self publishing. I've seen interesting applications of it as a small part of a creative workflow from graphic artists. I've heard of machine learning processes being used in data analytics, coding, scientific research, etc.

But that's not ubiquity. For the AI hype to pay off it needs ubiquity, and I frankly don't see it.

[JWags85]

Congrats to Elon Musk for successfully holding shareholders hostage. He apparently got the full ransom he demanded, and he crowed about it well ahead of any announcement.

With that success, he immediately is threatening to do it again. He wants 25% ownership of Tesla and has suggested on numerous occasions that if he doesn't get his way, he might push the best AI initiatives into other businesses he owns rather than Tesla.

The compensation fight is interesting.  Cause on one hand, he's really totally in his right to demand the fulfillment of the agreement.  He/TSLA set a crazy mark to reach and to his credit, regardless of what he did with with stock price, they hit it.  Its not like he's suddenly created a compensation package he deserved for past efforts.

That being said, things have materially changed financially with Tesla, so its not really prudent to be wanting to pull that sort of level of cash out of the company now.  And obviously his temper tantrums and strong-arming after the fact are a terrible look and beyond petulant.  As we've said plenty of times, it doesn't exactly portray stable and reasoned leadership.

[GOO]

I wasn't specifically hyping apple except that was the topic. It applies to a lot of tech companies.

Is AI over hyped, sure in some ways I'm sure it is. And in ways unforeseen probably not. But based upon Tesla scrapping a decade of self driving programming on decision making, and making it a lot better in 12 months using AI training for decision making, I am sold that it is a game changer.  In the fall I thought self driving was way off, once I tried v12 that was AI trained on humane drivers, completely changed my mind. Nothing short of amazing. For once it isn't hype it is happening.

Look at how difficult it is to train robots using programming. AI changes that. They can copy humane behavior. A company can apply their vision training and knowledge and combine it with AI training for decision  making and make real progress quickly.


I've found tech, say a phone or MS products frustrating. Repetitive tasks that are performed over and over but the software has no clue, doesn't recognize a pattern etc. it should train itself on similar matters or patters. AI will change that quickly. For some it might be a yawn who cares. For me that will be a great time saver. I want my software to learn what I do and start to help me do it and to apply history to the present. Context aware etc.

[Hards Alumni]

I'd love for that all to be the case, good comment!

[jficke13]

I wasn't specifically hyping apple except that was the topic. It applies to a lot of tech companies.

Is AI over hyped, sure in some ways I'm sure it is. And in ways unforeseen probably not. But based upon Tesla scrapping a decade of self driving programming on decision making, and making it a lot better in 12 months using AI training for decision making, I am sold that it is a game changer.  In the fall I thought self driving was way off, once I tried v12 that was AI trained on humane drivers, completely changed my mind. Nothing short of amazing. For once it isn't hype it is happening.

Look at how difficult it is to train robots using programming. AI changes that. They can copy humane behavior. A company can apply their vision training and knowledge and combine it with AI training for decision  making and make real progress quickly.


I've found tech, say a phone or MS products frustrating. Repetitive tasks that are performed over and over but the software has no clue, doesn't recognize a pattern etc. it should train itself on similar matters or patters. AI will change that quickly. For some it might be a yawn who cares. For me that will be a great time saver. I want my software to learn what I do and start to help me do it and to apply history to the present. Context aware etc.

Right, I think this is spot on and completely in line with what I said above. Machine learning applications will have plenty of real, impressive, *niche* applications. The AI hype cycle however is predicated on pushing it out into the hands of everyone for all tasks, when at best it's making writing emails a tiny bit faster (but also flattening them all into the same predictive-text diction and syntax).

[rocket surgeon]

Do you use Morningstar or Value Line or something to filter down what you're looking for?

Note: I think that's what I've used the public library system for the most in the past 10+ years. Usually you can get an online library card and your local library will have a subscription to one or the other for free. Really valuable insights into companies from industry analysts. Attached an example for VZ (pdf warning)

sorry, just saw this-

   first I Google dividend stocks, then run them thru mrket watch and Yahoo financial analysis then run them past my Edward Jones guy

  Eddy Jones uses morningstar a lot

[MU82]

Leaders at firms such as T. Rowe Price and JPMorgan expect a "broadening" of the market in coming years - in other words, it won't just be a few companies (such as the so-called Magnificent 7) leading the earnings gains. Such a shift would generally be viewed as a positive thing.

Here's a chart by JPMorgan Asset Management to illustrate the company's expectations:

[SoCalEagle]

sorry, just saw this-

   first I Google dividend stocks, then run them thru mrket watch and Yahoo financial analysis then run them past my Edward Jones guy

  Eddy Jones uses morningstar a lot

Rocket, I hope you don't really have an EJ guy.  Why give away your hard earned money for something you can easily do yourself? 

[SoCalEagle]

Leaders at firms such as T. Rowe Price and JPMorgan expect a "broadening" of the market in coming years - in other words, it won't just be a few companies (such as the so-called Magnificent 7) leading the earnings gains. Such a shift would generally be viewed as a positive thing.

Here's a chart by JPMorgan Asset Management to illustrate the company's expectations:

82, that is a very interesting chart.  I can see this happening as AI matures over time.  In my opinion, we won't see who the real winners of the AI wars are for another decade or so. 

[Skatastrophy]

AI has a long, long way to go before it's broadly impactful. Right now it's very Web 2.0, or Big Data. Both interesting and sustained impact but not market shifting economy changing impact.

[MU82]

82, that is a very interesting chart.  I can see this happening as AI matures over time.  In my opinion, we won't see who the real winners of the AI wars are for another decade or so.

Agree.

[MU82]

Interesting graphic on Yahoo Finance showing that the market could have plenty more room to run before hitting a "danger" zone. It obviously doesn't mean it will go up that much more.

[MU82]

As recently as about a month ago, there were concerns that Apple might be the odd company out of the AI derby. Now, many analysts are starting to view the company as an AI innovator that will be a huge beneficiary.

This analyst says Apple will benefit from an "AI supercycle" that alone could be worth $30-40 to the AAPL share price:

https://seekingalpha.com/news/4118267-apple-likely-to-see-ai-driven-super-cycle-analysts-say?

[TSmith34, Inc.]

If they can pull it off yet again, Apple will be the practical beneficiary of new technology without being the ones that had to burn billions and billions of dollars to create it.

[Skatastrophy]

For the interesting bits Apple is just licensing OpenAI, so the profits funnel directly to Microsoft?

[jficke13]

For the interesting bits Apple is just licensing OpenAI, so the profits funnel directly to Microsoft?

But not really using it for anything. OpenAI/Microsoft are losing money on the deal. It's... a weird deal.

https://www.wheresyoured.at/untitled/

"One might imagine this integration was a huge coup for Altman and OpenAI, except for one wrinkle — as reported by Bloomberg's Mark Gurman, Apple didn't give OpenAI any money to integrate ChatGPT, paying instead in "distribution" of a tool that loses OpenAI money on literally every transaction.

Gurman also reports that users will also be able to upgrade their ChatGPT accounts to ChatGPT Plus through the integration — although these upgrades will likely operate on the same terms of every other digital goods sale on iOS, with OpenAI forced to share 30% of the take with Apple, unless the user upgrades directly on the ChatGPT website.

To be clear, in most cases companies integrating ChatGPT pay on a per-thousand-tokens basis, meaning that OpenAI, while unprofitable, still generates revenue of some sort on each query. Yet Apple's deal doesn't even appear to pay OpenAI for using ChatGPT, meaning that OpenAI's deal will immediately lose them money from the second iOS 18 launches."

[Skatastrophy]

So Apple is distributing OpenAI product like a normal app but integrating it natively so that they can say they're doing AI stuff

OpenAI is obviously gunning for market share first before raising prices (as is tradition). I don't think that's such a weird deal. Get those iOS uses vendor-locked, right?

If I'm wrong, tell me. I really haven't looked into it so much. I have a macbook pro but I don't really participate in the apple ecosystem so I'm behind the times.

[GOO]

If they can pull it off yet again, Apple will be the practical beneficiary of new technology without being the ones that had to burn billions and billions of dollars to create it.

This might be right. Apple has and is spending a lot on what they used to call machine learning including their neural processors, on device abilities, apps,  etc.

But for large language models, and associated uses, etc, for now they will rely a lot on others. So they get the benefit of selling devices that are AI devices, that need to be upgraded more quickly, without spending as much as they otherwise would have. And the user will be able to choose which outside AI companies they want to use, a benefit to apple users at least in the short term.

Apple will use their own AI on their apps and processes.  But when a query goes outside apple parameters, it will go to a users preferred LLM or a specialty one.
So ChatGPT and other LLMs will be integrated, like a google search almost, but optional to use and users will have choice on which company to use for different types of matters.

I envision something like Siri saying I can't handle X, but would you like LLM Y to handle that query or solve that problem etc. But there will be a choice of outside providers. Some of the providers may get subscriptions from apple users if they find them helpful enough.

So could be a win win for now.  But I  think apple will need to develop a backup plan other than using others to get 30% of subscriptions.

But ya, an AI device without paying for all the development seems to be in the offering. At least until it all shakes out and there are a few dominate well established LLMs that can dictate better terms. Or maybe LLMs become more like a commodity and no one dominates.

[jficke13]

So Apple is distributing OpenAI product like a normal app but integrating it natively so that they can say they're doing AI stuff

OpenAI is obviously gunning for market share first before raising prices (as is tradition). I don't think that's such a weird deal. Get those iOS uses vendor-locked, right?

If I'm wrong, tell me. I really haven't looked into it so much. I have a macbook pro but I don't really participate in the apple ecosystem so I'm behind the times.

I think part of the issue is not that OpenAI is gunning for market share before raising prices, but gunning for market share without any revenue at all. To the extent that they lose money via burn rate on every query (and they do... maybe they can claim Microsoft Azure credits make it revenue neutral), then this isn't Amazon exploiting a market by crushing competition, runnig Borders out of business, and then raising the price on books all the while deriving revenue from the market that is plowed back into the growth model, this is more like "I can keep allowing people to run queries because VC money is propping me up, also the more popular Apple's drive of traffic to me is, the *more* money I lose without any revenue to offset the activity and no mechanism to turn around and squeeze Apple (or Apple's users) via higher prices to create profit."

[Skatastrophy]

I think part of the issue is not that OpenAI is gunning for market share before raising prices, but gunning for market share without any revenue at all. To the extent that they lose money via burn rate on every query (and they do... maybe they can claim Microsoft Azure credits make it revenue neutral), then this isn't Amazon exploiting a market by crushing competition, runnig Borders out of business, and then raising the price on books all the while deriving revenue from the market that is plowed back into the growth model, this is more like "I can keep allowing people to run queries because VC money is propping me up, also the more popular Apple's drive of traffic to me is, the *more* money I lose without any revenue to offset the activity and no mechanism to turn around and squeeze Apple (or Apple's users) via higher prices to create profit."

OpenAI had (arguably) ~$2B in revenue in 2023 and on pace for $4B in 2024, but I think you're saying revenue when you mean income?

And you're right, they don't have VC cash to prop them up through their operational loss years. They have the $10B investment from Microsoft. And with Microsoft owning significant amount of OpenAI IP, the Microsoft investment won't dry up.

This business model is the same as lots of tech. Youtube has never been profitable, losing money every year since 2006, because it's so expensive to save and stream all of those videos. Twitch has never been profitable, because it's so expensive to store and stream video, but it was acquired by Amazon for $1B.

OpenAI + Microsoft can absolutely run at a loss indefinitely. Owning the platform and dominating an entire market is more important than profitability, depending on the platform.

[jficke13]

No I'm using the term revenue intentionally. All queries referred by Apple will generate OpenAI no revenue, because neither Apple nor the Apple user will be paying OpenAI under the terms of this agreement.

[Jay Bee]

Do you use Morningstar or Value Line or something to filter down what you're looking for?

Note: I think that's what I've used the public library system for the most in the past 10+ years. Usually you can get an online library card and your local library will have a subscription to one or the other for free. Really valuable insights into companies from industry analysts. Attached an example for VZ (pdf warning)

Have used Value Line via the library for decades. Great, free resource for screening

[rocky_warrior]

gunning for market share without any revenue at all

I present some history for your evaluation:
https://www.businessinsider.com/in-2000-businessweek-wondered-how-will-google-ever-make-money-2009-3

[Skatastrophy]

No I'm using the term revenue intentionally. All queries referred by Apple will generate OpenAI no revenue, because neither Apple nor the Apple user will be paying OpenAI under the terms of this agreement.

Got it, my bad. So when you said they aren't making revenue at all you're talking about the iOS users. My bad. In the keynote Apple said that iOS users will be able to upgrade to GPT Plus. iOS users are widely known to be easy to part with their $$ compared to mobile users on other platforms. I bet this works out ok for OpenAI.

ChatGPT has a free version too. They'll make plenty of revenue from Pro users

[jficke13]

I present some history for your evaluation:
https://www.businessinsider.com/in-2000-businessweek-wondered-how-will-google-ever-make-money-2009-3

And in turn I present some history for your evaluation: