Can the machine cut through all the noise?

Community
1

I just watched a video by Arun Rupesh Maini where he explores why Google Search might be getting worse.

Arun brings up some interesting points about how keywords and SEO optimized intros are distorting search results. Sites are designed to rank higher, not necessarily to give us the best answers. It makes me wonder… is AI already sharp enough to cut through this kinda noise?

(Edit: I understand that SEO tactics aren’t a new thing. The problem appears to be the extent, and how it warps patterns, or even dates and time.)

He also worries that Google Search might eventually “replace websites entirely”, showing only Gemini responses and ads at the top. If that happens, search could turn into a space filled with ads linking to more ads, and the real content gets buried.

Personally, I doubt Google will go that far, but I’ve never used Google Search much anyway. I’ve been using Ecosia for as long as I can remember. :deciduous_tree:

That said, I’ve been turning to my GPTs more often these days. For me the natural, conversational responses are a much better way to get real answers without wading through the SEO swamp.

I hope ChatGPT can cut through the noise and become a serious challenger to Google Search.

There’s money in ads and data, I get that, but maybe there’s another way to monetize search, like, honestly, even a paywall would be better. Though I’m realistic enough to know that subs would never compete with the revenue stream from ads and data mining… oh, well.

Still, it’s nice to think there could be a future where search is driven by clarity and truth, not just profit. Maybe this sounds naive, but I want to believe this is what the “Open” in OpenAI could stand for. (ngl, I’m getting really tired of people tweeting that OpenAI should be renamed.)

What do you think?

1 Like
2

Majority of the content is pure (marketing) trash and that is the content that is used to educate a GPT. And at same time fact based sites makes everything to stop data harvesting. The strong bias is already there and I don’t see that an AI can or would get media reading ability.

Trash in, trash out, and because GPTs are just google on steroids with better makeup, the results are never any better.

Yes, I use GPTs a lot, because those saves my time when I wouldnt need to struggle thru 20 pages of fairytales and ads before real hits. But hallucinations aren’t the real issue here, it is how GPT is trained.

So I’m not that thrilled. Perhaps it gets better at some point years to come. But… an american business would do that… :joy: Chatting is just a feature was masses without real benefits. Real work, good and bad, happens somewhere else and doesn’t help ordinary searching.

1 Like
3

AI isnt trained on “websites”. AI is - as far as it is humanly possible - trained on meaningful data. This is also why AI will not will take over the world anytime soon. There simply isnt the amount of useful data available to make AI significantly more knowledgable.

I almost never Google anymore.
AI is a much better way to find information, and to guide you to the websites you was looking for. At least for now. These things cant work without monetising, and it is a very scary thought that AI will start to become biased, based on advertising.

Google has many great products, but their search engine today, has the last decade declined from being a source of information, to a source of advertising.

1 Like
4

People trying to game a system is inevitable. I wouldn’t say that SEO is the ruination of Google Search. Google is. They have continuously prioritized short-term profits over longevity.

They most certainly know how to bring the most useful content to the first.

Seems like a classic cannibalistic dependency (yum). If it regurgitates the content of the website it has consumed and leads to less people clicking on the website, what do? Why make website?

In my opinion this is all a clear indicator of a transition in the WWW. Static websites will mostly be RAG databases that AI can communicate with. Most websites will be really “web apps” - highly interactive platforms that can deliver information faster and easier than one can communicate.

I have to note that there was a “leak” of slide deck. OpenAI was pitching to the big publishers that if they joined up they would have

priority placement and “richer brand expression” in chat conversations, and their content benefits from more prominent link treatments.

Worse than SEO - which at it’s peak was simply just the best practices to reach the people who are looking for you.

BUT OpenAI did respond to this leak saying:

A comment from OpenAI said Adweek’s report “contains a number of mischaracterizations and outdated information.”

So who knows. But I would be very surprised if ChatGPT doesn’t eventually have it’s own sponsored content and SEO - it’s inevitable.

100%. OpenAI has shifted from churning the internet to capture the general meaning of semantics to now having very smart people in their specific fields create the training data, leading to highly more intelligent models.

Each day the internet becomes more artificial from AI-generated “slop” but the good news is that AI like ChatGPT is “smart” enough to know that maybe you shouldn’t add glue to your pizza :rofl:.

It’s very easy to just pull whatever content was returned and feed it through your own local LLM & framework to determine how credible it really is. I think this is where the power is.

I mean, really. Just only 100 years ago if Uncle Roger said “The moon is made of cheese” it’s like, dang… Well, I ain’t going to a library to disprove that, I guess the moon is made of cheese.

Being able to not only confirm information, but now also attack it from multiple subjective angles is easier than ever. I think the question should be “Will HUMANS be able to cut through all the noise??” cue dramatic music

2 Likes
5

can’t wait for searchGPT

another example would be free youtube… it has gotten worse over the years, much like google search has… the pattern I see is that when the server costs of these websites/apps turn into billions in profits the priority goes profit over quality

it depends, on real time use cases, it might not due to training data cut off and hallucination on search (e.g. non existent links/references)

1 Like
6

Well, that ain’t true by OpenAI. And they aren’t curated the content in any means.