Investing General Discussion
- Thread starter Furious
- Start date
-
Guest, it's time once again for the massively important and exciting FoH Asshat Tournament!
Go here and give us your nominations!Who's been the biggest Asshat in the last year? Give us your worst ones!
Big Phoenix
Pronouns: zie/zhem/zer
- 46,808
- 99,998
It bothers me the head turns and looks down before the bar goes through the desk.
- 1
Palum
what Suineg set it to
- 27,537
- 43,864
Jysin
Ahn'Qiraj Raider
- 6,512
- 4,525
- 11,117
- 59,398
Sanrith Descartes
You have insufficient privileges to reply here.
- 45,473
- 123,817
Investing comes in many forms. A pretty amazing story.
finance.yahoo.com
How This Retired NBA Player Invested His Way To A Net Worth Of $600 Million: Hint, It's Not Magic Johnson, Shaq, or Jordan
If you look at a list of the richest former NBA players, you will find one name much less recognizable than LeBron James, Magic Johnson, and Michael Jordan: Junior Bridgeman. Bridgeman never made an all-star team or was paid more than $350,000 ...
- 3
- 13,216
- 51,602
I think I talked about this some in another thread but...
ChatGPT coming up with it's own search engine will end up being futile unless they also start their own web crawling/spidering and not piggy back on google/bing/whoever for the underlying "searches" data. This being because all Goog/MS/whoever has to do it block access from the GPT client leveraging their underlying search results and they're hosed. And in terms of cost model, a GPT driven search, PLUS the overhead to crawl the whole internet, is going to make being profitable as a venture hard for them...
I saw this play out some and be the reason that Brave started it's own search engine, and over time weened itself off relying on any other company for the underlying database. (if you haven't already, try it here.)
Also, all the major search players are integrating "AL/LLM" style response actions. The first thing in any brave search response now is an AI generated summary response.
All this taken together means what OpenAI is proposing here is nothing "new under the sun" and they'll need to displace much bigger players with their own technologies and years of head start in search. I would be curious as to how MS sees this in relation to a challenger to Bing with MS being an investor in OpenAI on the level they are.
Palum
what Suineg set it to
- 27,537
- 43,864
The problem is:
1) AI is useless if it is built for 'modern audiences'
2) AI can't be monetized using traditional methods if it is intended to provide better service
3) AI still isn't that great at a lot of domains and tasks, and I just have to use manual search to find stuff all the time still
So given all those I'm not sure how any of these will be able to do anything against Google with the constraints they put on themselves. Presumably they will have to recoup costs somehow and make money (is Open AI for profit again now?).
Let's take an example of an error or something broken. If you look it up on Google, you'll get ads for service, replacements, parts, etc. and eventually if you dig around enough you'll find some forum or something where some guy fixed it by doing xyz. Google hopes you'll see the shiny new thing and click. Does anyone really believe that Open AI is going to provide some sort of concise answer with no economic transaction associated instantly for free? Does anyone think that you'd care about an unsourced AI text snippet without some sort of context or video to understand the method?
If you just need to know the capital of North Dakota real quick, search engines with augmented AI summaries are already great at this. I don't see how OpenAI adds anything to this, since if I wanted an AI answer I'd just use ChatGPT anyway.
1) AI is useless if it is built for 'modern audiences'
2) AI can't be monetized using traditional methods if it is intended to provide better service
3) AI still isn't that great at a lot of domains and tasks, and I just have to use manual search to find stuff all the time still
So given all those I'm not sure how any of these will be able to do anything against Google with the constraints they put on themselves. Presumably they will have to recoup costs somehow and make money (is Open AI for profit again now?).
Let's take an example of an error or something broken. If you look it up on Google, you'll get ads for service, replacements, parts, etc. and eventually if you dig around enough you'll find some forum or something where some guy fixed it by doing xyz. Google hopes you'll see the shiny new thing and click. Does anyone really believe that Open AI is going to provide some sort of concise answer with no economic transaction associated instantly for free? Does anyone think that you'd care about an unsourced AI text snippet without some sort of context or video to understand the method?
If you just need to know the capital of North Dakota real quick, search engines with augmented AI summaries are already great at this. I don't see how OpenAI adds anything to this, since if I wanted an AI answer I'd just use ChatGPT anyway.
ToeMissile
Pronouns: zie/zhem/zer
- 3,317
- 2,170
Lex had the CEO of Perplexity on recently, they talk a good bit search/LLM and related stuff. It’s a good listen.
Loser Araysar
Chief Russia Reporter. Stock Pals CEO. Head of AI.
- 81,654
- 163,309
They launched their own crawl bot with this release. They will be crawling the web too it seems.
It makes sense because they cant continue to evolve without generating an independent data source themselves to continue to train and augment their Gen AI.
I do SEO for a living so I'm in the weeds on this stuff all the time. This is largely a gimmick by OpenAI but for people who don't understand how search marketing works, this stuff will deceptively look like a threat
- 2
- 1
Loser Araysar
Chief Russia Reporter. Stock Pals CEO. Head of AI.
- 81,654
- 163,309
I agree.
AI can be very effective for top of the funnel, 5W1H queries. The real problem it runs into is that as you move from top of the sales funnel to the bottom where a sale (aka profit) is made, AI utterly fails there. The most basic example is buying some appliance. Let's say you need a new washing machine. You Google "best washing machines" and you get top 10 lists, reviews, pricing that changes regularly, etc. AI simply can't provide an effective answer to that type of query. But it can tell you what you should consider when buying a washing machine, which is a 5WH1 query.
- 1
- 1
Palum
what Suineg set it to
- 27,537
- 43,864
LLMs don't really do H that well either though. Best case is there's a common explanation that is repeated enough to strongly weight answers. More often it abstracts or generalizes to ensure "higher quality", but the reality is this is actually higher quantity masquerading.
Blazin
Creative Title
- 7,142
- 36,640
Just partial now own $100k of spyDid you pull the trigger on this? Just curious
- 1
- 1
- 1
