You might have stumbled onto a practical and useful application of LLMs. Yahoo (or whoever) could crawl the web, categorise each page, and provide a genuine "Index".
Like in the back of a book, or an old subject based card catalogue ... Dewy-decimalise the web :)
I've always felt that an index by subject would be more useful than string-match based searching. Of course, the index might rank links within each sub-sub--sub-sub...category with something like the original page-rank.
Now if Yahoo (or whoever) could avoid the enshitification trap ... imagine what a fabulous resource that could be.
Until ChatGPT came along, I figured it was inevitable that human curated search came back into ascendancy as the crawler model has become such a failure.
Now we can use ChatGPT to filter through Google's infested mess, but this double edged Sword of Damocles will be able to create infinite attempts to bury genuine content with ad spam.
I might pay like max. 3 EUR a year for this to get a search engine that gives the good results without ads, SEO spam and bogus clone sites.
That amount of money is probably more than Google now makes from my online presence because I adblock, block 3rd party cookies, tend to click "block" to everything including the idiotic "legitimate concern" and never ever click on ads.
Can't remember exactly where I saw it, but last number I saw said that Google makes ~$12 a year per user. Which begs the question... why have they not atleast tried a "Google Premium".
Fuck it, I'd pay $15 a year to have a Google search that puts as much effort into finding me the shit I actually want as Google does today in serving me BS ads I never pay attention to.
I always wanted to see just a "price transparency" aspect.
Tell me exactly how much the advertiser paid for his placement, and that's a hugely important signal here.
If I'm searching for weird hobby parts, even though it's a high purchase intent query, they're probably paying pennies per click.
But if you start searching, say, financial stuff and the ad placement figures start showing multiple dollars per click, it's a warning "these people are willing to spend THIS MUCH MONEY to present a message to you, this probably means there's something sketchy involved."
I know, for example, anything pertaining to insurance and financial products is highly likely to turn into a farm of cross-selling and personal-information harvesting, because the cost per acquisition is so high and the tendency for everyone to sell the information to everyone else is so great.
I'd argue most people do not even realize they click on ads. My mother! The ad call-outs got so inconspicuous that they're almost indistinguishable from ordinary search results. And if you don't realize that and just click on the top results, you're amongst the top profitable users.
This is an interesting point. User managed or curated indices offer unique advantages, especially when 'depth of coverage' is more important than 'breadth of coverage'. I believe that we are witnessing people shift away from demanding 'search breadth' as we speak, so someone might possibly decide to do this.
Everything else is effectively the influencer scene. Which is increasingly deplorable as well.
Anything with wide enough reach becomes cost-effective for gaming.
So one would have to return to a highly fragmented world to make gaming the system cost prohibitive.
And that would get us to a pre-Internet world. But then again, it’s not entirely unthinkable that we’re headed towards increasing Internet fragmentation if various governments get their way.
Like in the back of a book, or an old subject based card catalogue ... Dewy-decimalise the web :)
I've always felt that an index by subject would be more useful than string-match based searching. Of course, the index might rank links within each sub-sub--sub-sub...category with something like the original page-rank.
Now if Yahoo (or whoever) could avoid the enshitification trap ... imagine what a fabulous resource that could be.