Investing General Discussion

Jysin

Bronze Baronet of the Realm
6,852
5,035
Today's net SPY Gamma is slightly tipped to calls.

However, next week is hugely skewed to puts. (Guessing many are set up to fade the bounce next week)
1741972629851.png
 

Blazin

Creative Title
<Nazi Janitors>
7,421
37,339
Thinking of making a change to LT hold portfolio.

GOOGL

I'm thinking of selling GOOGL, not because it's down. I'm finding I'm not using it anymore, and I feel like I shouldn't ignore that. When I want to search or research something I tend find myself gravitating towards using grok and not google and I don't think I'm the only one. By no means am I suggesting that GOOGL is going to just collapse and they still have a juggernaut hold on online advertising. The moat however is being threatened in away that I don't think should be ignored. It could morph into a value trap in that you will continue to put up impressive numbers and it will trade at a discounted PE and this will keep sucking in new people looking for a deal, but the stock just can't get moving.

I don't want to make this decision too abruptly, I'll most likely start selling CCs much closer to the money as my first action. In the long run I think I'd rather apply that capital to an investment with better prospects towards growth.

AAPL isn't far off from similar thinking without some sort of innovation new products I might start reducing holdings, less interested in the company just as a financial engineering cash machine. It's been much kinder to my money than GOOGL so it has some room yet.
 
  • 2Like
Reactions: 1 users

Jysin

Bronze Baronet of the Realm
6,852
5,035
Interesting that premarket we opened around -0.5% on SPY. We are now at +0.4%, but notably this is while TSLA is in the shitter again -5% and NVDA is -1.7%
 
  • 1Like
Reactions: 1 user

Blazin

Creative Title
<Nazi Janitors>
7,421
37,339
Interesting that premarket we opened around -0.5% on SPY. We are now at +0.4%, but notably this is while TSLA is in the shitter again -5% and NVDA is -1.7%
Look at RSP and QQQE both up .8%+ big tech struggling some here but lots of names holding up well so far. Long way to close and it will be difficult to continue the overall bounce without the big boys
 
  • 1Like
Reactions: 1 user

Jysin

Bronze Baronet of the Realm
6,852
5,035
Breadth matters. It's a good thing that some froth is being taken out of tech, when there is strength everywhere else. Eg: IWM up 0.9% while QQQ only 0.1%

NYSE has 78% advancers. This is still nothing more than a textbook correction so far.

We are certainly still in a headline driven market though, so keep an eye on Trump tweets. Positive news of any kind can shoot the SPY right back to retest the 200D.
 
  • 1Solidarity
Reactions: 1 user

Jysin

Bronze Baronet of the Realm
6,852
5,035
SPY still near highs and NVDA now -3% and TSLA -6.5%


Mag7

1742230413249.png
 
Last edited:
  • 1Solidarity
  • 1Like
Reactions: 1 users

Haus

I am Big Balls!
<Gold Donor>
15,001
61,171
Thinking of making a change to LT hold portfolio.

GOOGL

I'm thinking of selling GOOGL, not because it's down. I'm finding I'm not using it anymore, and I feel like I shouldn't ignore that. When I want to search or research something I tend find myself gravitating towards using grok and not google and I don't think I'm the only one. By no means am I suggesting that GOOGL is going to just collapse and they still have a juggernaut hold on online advertising. The moat however is being threatened in away that I don't think should be ignored. It could morph into a value trap in that you will continue to put up impressive numbers and it will trade at a discounted PE and this will keep sucking in new people looking for a deal, but the stock just can't get moving.

I don't want to make this decision too abruptly, I'll most likely start selling CCs much closer to the money as my first action. In the long run I think I'd rather apply that capital to an investment with better prospects towards growth.

AAPL isn't far off from similar thinking without some sort of innovation new products I might start reducing holdings, less interested in the company just as a financial engineering cash machine. It's been much kinder to my money than GOOGL so it has some room yet.
You're far from the only one seeing this.

It's now "what AI do you ask for explanations (that probably have search on the back end)" and that's a new market/competition where Google is far from as dominant as it was in traditional search. With Grok, Microsoft, ChatGPT, others all in there.

One of the next big pit fights I expect will be when Grok or ChatGPT wants to have it so that you can choose your voice assistant in Android. For instance, use Grok instead of Google Gemini on Android phones.
 

Il_Duce Lightning Lord Rule

Lightning Fast
<Charitable Administrator>
11,736
64,431
Thinking of making a change to LT hold portfolio.

GOOGL

I'm thinking of selling GOOGL, not because it's down. I'm finding I'm not using it anymore, and I feel like I shouldn't ignore that. When I want to search or research something I tend find myself gravitating towards using grok and not google and I don't think I'm the only one. By no means am I suggesting that GOOGL is going to just collapse and they still have a juggernaut hold on online advertising. The moat however is being threatened in away that I don't think should be ignored. It could morph into a value trap in that you will continue to put up impressive numbers and it will trade at a discounted PE and this will keep sucking in new people looking for a deal, but the stock just can't get moving.

I don't want to make this decision too abruptly, I'll most likely start selling CCs much closer to the money as my first action. In the long run I think I'd rather apply that capital to an investment with better prospects towards growth.

AAPL isn't far off from similar thinking without some sort of innovation new products I might start reducing holdings, less interested in the company just as a financial engineering cash machine. It's been much kinder to my money than GOOGL so it has some room yet.
This isn't specifically stonk related, but I will concur about how much better at search Grok is.

I was using Bing as an alternative in an effort to get out of Google's sandbox for most things up until maybe 2 years ago. Then the results you get with bing started to go in the shitter. It was never that good, but it began to become literally useless. And not for political or even political adjacent stuff, I'm talking searching for parts or searching for esoteric info like engine models or weights or what not. Bing was outputting completely useless results, whereas google was outputting at least some useful info a lot of the time, but not all of the time. Anecdotally, call it about 75% useful. Whereas bing results were 100% useless or even wrong or misleading.

Now though? Google's definitely getting worse. As an example, I had a vid on youtube come up about a VW SP2. It's a rare VW beetle chassis variant built in Brazil. Very pretty. So, I was curious if there was a kit or replica for sale or if there had been. I spend probably 30 mins googling, and I can't get ANY good results about my question, it's all just ads or unrelated things or just plain wrong info.

A year ago, I'd have been like, "Well, fuck, this is irritating, I guess that means it doesn't exist? But I can't know because maybe I'm just not searching on the right thing or it's buried in a specialist forum that google can't search somewhere, or maybe facebook groups." Now though, I remembered grok exists. 2 minutes later, I have EXACTLY what I wanted to know in chapter and verse including history of people making related kits and what my likely alternatives are should I want to have something like that made. Amazing.



That kind of capability absolutely has the potential to threaten google, assuming their core business still partially depends on search (and the related ads, of course) being a significant portion of their business.
 
  • 3Like
Reactions: 2 users

Captain Suave

Caesar si viveret, ad remum dareris.
5,694
9,787
I guess I'm going to have to try grok. I'm SO fucking tired of trying to query google for exact terms and getting "nearest 1,000 rebranded Chinese products for sale" as a result.
 
  • 1Solidarity
  • 1Like
Reactions: 1 users

Cad

scientia potentia est
<Bronze Donator>
27,021
56,464
Grok is surprisingly smart about such a wide variety of things, it is actually scary. I wouldn't bet my job on it but it has been basically right about so many things that it is a great go-to instead of googling.
 
  • 1Like
  • 1Solidarity
Reactions: 1 users

Sanrith Descartes

You have insufficient privileges to reply here.
<Gold Donor>
46,949
132,226
My 2 cents. Currently Grok and Open AI are ad free. I think you pay for the current versions of both. We could end up with the scenario of free but "ad supported" Google vs paid but "ad-free" Grok/Open AI.
 

Sanrith Descartes

You have insufficient privileges to reply here.
<Gold Donor>
46,949
132,226
Grok is surprisingly smart about such a wide variety of things, it is actually scary. I wouldn't bet my job on it but it has been basically right about so many things that it is a great go-to instead of googling.
I like that so far every problem that has had some math component I give to Grok it shows the calculations and its actually correct. Math has never been chatbot strength. I have had it do annuity type calculations and its been spot on.
 

Captain Suave

Caesar si viveret, ad remum dareris.
5,694
9,787
I like that so far every problem that has had some math component I give to Grok it shows the calculations and its actually correct. Math has never been chatbot strength. I have had it do annuity type calculations and its been spot on.

I just ran some spot checks and get inaccuracy at high precision. For example, in the course of doing compound growth math Grok swears e^(.192) = 1.211563846831682, when it's really 1.21167051696. This was repeatable across multiple questions requiring the same computation.

Nitpicky, but odd. Be careful if you're doing complicated operations that require precision.
 
  • 2Like
Reactions: 1 users

Cad

scientia potentia est
<Bronze Donator>
27,021
56,464
I just ran some spot checks and get inaccuracy at high precision. For example, in the course of doing compound growth math Grok swears e^(.192) = 1.211563846831682, when it's really 1.21167051696. This was repeatable across multiple questions requiring the same computation.

Nitpicky, but odd. Be careful if you're doing complicated operations that require precision.
I don't see grok and the others replacing actual work flows yet, but it seems like a great way to get a ballpark answer or get specifics on things that you don't have time to look up. You can go check the citations it gives and get the ball rolling on your own analysis, but I definitely would not just trust its output without checking.
 
  • 1Like
  • 1Solidarity
Reactions: 1 users

Captain Suave

Caesar si viveret, ad remum dareris.
5,694
9,787
I don't see grok and the others replacing actual work flows yet, but it seems like a great way to get a ballpark answer or get specifics on things that you don't have time to look up. You can go check the citations it gives and get the ball rolling on your own analysis, but I definitely would not just trust its output without checking.

Yeah. Sadly the AI builders do a terrible job at publicizing the limitations and putting up guardrails. My clients keep asking me why I don't use AI for data ingestion to parameterize the optimization models I build. I have to turn out examples like this about twice a month.

I know enough about LLM technology under the hood to understand why they behave this way, but it still seems like a failure of product development.
 

Cad

scientia potentia est
<Bronze Donator>
27,021
56,464
Yeah. Sadly the AI builders do a terrible job at publicizing the limitations and putting up guardrails. My clients keep asking me why I don't use AI for data ingestion to parameterize the optimization models I build. I have to turn out examples like this about twice a month.

I know enough about LLM technology under the hood to understand why LLMs behave this way, but it still seems like a failure of product development.
I don't know how LLM's do math under the hood, but its weird that it can't outsource the math so to speak to Matlab or Mathmatica type programs which do do math correctly.
 
  • 1Solidarity
Reactions: 1 user

Captain Suave

Caesar si viveret, ad remum dareris.
5,694
9,787
I don't know how LLM's do math under the hood

Essentially they don't, which is the core of the problem. They just produce language tokens likely to follow the question you asked. I know there are calculation oversight modules that are supposed to bridge the gap, but they don't seem to work perfectly.
 
  • 1Like
  • 1Solidarity
Reactions: 1 users

Sanrith Descartes

You have insufficient privileges to reply here.
<Gold Donor>
46,949
132,226
Wasn't able to get around to funding our IRAs/SEPs last week but got it done today. This money was all fire and forget straight into SPLG.
 

Arden

Blackwing Lair Raider
2,888
2,169
I just ran some spot checks and get inaccuracy at high precision. For example, in the course of doing compound growth math Grok swears e^(.192) = 1.211563846831682, when it's really 1.21167051696. This was repeatable across multiple questions requiring the same computation.

Nitpicky, but odd. Be careful if you're doing complicated operations that require precision.

I pound the ChatGPT drum a lot, but only because I've been so fucking impressed with it. Every other AI I've used I've had issues with. Granted, I don't do use it for complex math, but for language stuff and general info stuff, it's frankly nothing short of astounding- especially the new 4.5 version when it comes to writing. Caveat: I have a sub, so I'm not sure if the free version is as good. I've heard OpenAI limits resources for the free version.

Here's how GPT calculated the answer when I asked it your question:


Screenshot 2025-03-17 131321 - Copy.png