Live data from Hacker News

I just tested ChatGPT Search vs. Google – here's the results

tomsguide.com

21–30 of 34 posts

Re: I just tested ChatGPT Search vs. Google – here's the results

#21

Chatgpt search looks great from this, but if it does end up replacing Google, the biggest losers will be sites like tomsguide.com. It's hard to see how they could continue to make money when I can get the same info directly from the search engine and never need to open their site again. They even mention in the "best tvs" example that Google returns a link to one of their articles on tomsguide.com, while ChatGPT just…

Exactly. This is great for the user until the incentive for anyone to put out new content completely dries up, then not so much.

Re: I just tested ChatGPT Search vs. Google – here's the results

#22

This is an obvious Apples vs Oranges comparison. You should be comparing ChatGPT Search vs Gemini since they both are LLM that query a search engine for results that generates a text response based the the text it has been provided. Google Search is an actual search engine and querying it like it is an LLM is going to give you bad results. A good example of this is the query, What is Apple’s current stock price, and…

Yes, and no. User does not care how it is implemented, or how it works.

This is clearly same domain: search.

Re: I just tested ChatGPT Search vs. Google – here's the results

#23

> ChatGPT: Wow. Once again ChatGPT search provides a cleaner interface without promoted content. For searches this personal, the extra privacy-focused approach is much appreciated. It is far more appealing to me as a user who wants information without being targeted by ads during the search — or after. “Without promoted content” Suuure… as it is and forever shall be. > Winner: ChatGPT leads for an approach that takes…

Totally agree. Chatgpt is at the start of the enshittification circle.

Re: I just tested ChatGPT Search vs. Google – here's the results

#24
post #19
post #16

Earlier quoted context omitted.

> They probably realized they need to revamp their whole search It’s not the UI that should scare them, but rather the fact that their business model relies on showing answers in a format that users don’t want and someone is now presenting that. It’s like a running a bakery during the Atkins low carb phase, except this isn’t a phase that will disappear and Google doesn’t know how not to be a bakery.

The “format” is the UI, that is what allows them to show ads. If they are forced to change the “format” to one single explanation or answer they cannot show ads like they do now and instead change their “UI” to be like chaptgpt’s. Which also means they have to change their advertising model fundamentally

It is deeper than what you are saying, which is the point of my post. Change the UI and the business model they rely on breaks. It isn’t a UI issue but a fundamental business model issue. Even if they wanted to change their model, there is zero chance they could replace that revenue quickly enough with different revenue. They are addicted to ad sales to the point where it is a wholesale reliance. Like an alcoholic that has a heart attack shortly after kicking alcohol, they can’t change now. They should have disrupted their own business ten years ago when the writing was on the wall.

To summarize: the UI is a symptom of the larger issue.

Re: I just tested ChatGPT Search vs. Google – here's the results

#25

I might understand the excitement for a truly new competitor in the search space, bit this article sounds more like an ad from openAI than something critical and informative.

This was absolutely my take, as well. I was wondering if I was looking at some Tom's guide knockoff.

And I still don't trust ChatGPT with facts. I do use it as a starting point for research, but then I switched to [my preferred paid search engine] to verify and correct.

Re: I just tested ChatGPT Search vs. Google – here's the results

#26

Chatgpt search looks great from this, but if it does end up replacing Google, the biggest losers will be sites like tomsguide.com. It's hard to see how they could continue to make money when I can get the same info directly from the search engine and never need to open their site again. They even mention in the "best tvs" example that Google returns a link to one of their articles on tomsguide.com, while ChatGPT just…

It depends. When someone is doing a real serious search, they will check the linked references, in which case they will visit TomsGuide if it's a cited reference. In contrast, when someone is doing a non-serious casual search, they typically won't visit the linked site. Overall, it will cut the visits by ~80%, but that takes us back to the early 2000s level of page hits, when good sites like TomsGuide could then still thrive.

Re: I just tested ChatGPT Search vs. Google – here's the results

#27
post #24
post #19

Earlier quoted context omitted.

The “format” is the UI, that is what allows them to show ads. If they are forced to change the “format” to one single explanation or answer they cannot show ads like they do now and instead change their “UI” to be like chaptgpt’s. Which also means they have to change their advertising model fundamentally

It is deeper than what you are saying, which is the point of my post. Change the UI and the business model they rely on breaks. It isn’t a UI issue but a fundamental business model issue. Even if they wanted to change their model, there is zero chance they could replace that revenue quickly enough with different revenue. They are addicted to ad sales to the point where it is a wholesale reliance. Like an alcoholic th…

I tried to call explain the same thing to you that I was alluding and commenting on their business model to fundamentally change if the UI changes. It’s worse than the business model changing because from an end user perspective they will be chqnging to meet ChatGPT’s UI which will seem like much larger capitulation. I think we understand each other now though :)

Re: I just tested ChatGPT Search vs. Google – here's the results

#28

This is an obvious Apples vs Oranges comparison. You should be comparing ChatGPT Search vs Gemini since they both are LLM that query a search engine for results that generates a text response based the the text it has been provided. Google Search is an actual search engine and querying it like it is an LLM is going to give you bad results. A good example of this is the query, What is Apple’s current stock price, and…

> vs Gemini

The issue is that the free version of Gemini remains really bad. Perhaps the paid version of Gemini is better, but that's not available to most people.

Re: I just tested ChatGPT Search vs. Google – here's the results

#29
post #25

I might understand the excitement for a truly new competitor in the search space, bit this article sounds more like an ad from openAI than something critical and informative.

This was absolutely my take, as well. I was wondering if I was looking at some Tom's guide knockoff. And I still don't trust ChatGPT with facts. I do use it as a starting point for research, but then I switched to [my preferred paid search engine] to verify and correct.

Are you the same beej that made those network programming eBooks from forever ago? If so, thank you very much for those!

Re: I just tested ChatGPT Search vs. Google – here's the results

#30

I did a similar test and tried to pull up certain categories of individuals I am interested in, with their names and linkedin profile links. ChatGPT hallucinated the names and the links. I simply cannot move to a search, where there is random hallucination because having to check for each and every result for hallucination defeats the purpose of search itself.

[dead]
Post reply on HN