Comment by ghm2199 | Hacker News Reader

Comment by ghm2199 | Hacker News Reader

My sister, a journalist, mentioned to me that she only uses google search because she had learned how to get information typically only Google indexed in the country she lives in, in a way it was not exposed on chat bots. She often has to search for information like Old govt forms released as public record with a fixed a certain format photo scanned into a pdf and indexed by Google were often on the second page of the search and beyond. But they are there. She knew how the forms looked and what bigrans and trigrams matching a certain part of form for a certain piece of information to search for and Google search has it. Like an official order on a tender notice for some government department which is no longer in the .gov.* website gave her the official's name and then she could track down who to contact in an office... ChatGPT and other bots don't have it. Some how all these government documents became part of the government record and are the key for her to do her job.

I sincerely hope google wont stop indexing that stuff just because of a PM in search "de/re-prioritizing" ranking in a way that makes this impossible.

This is a conundrum I always found interesting. If you know how to use a search engine (i.e. knowing how to use operators and structure a search query), you're almost always able to find what you're looking for very quickly, and in most cases (well, before SEO), the results are high-quality. You'll spend the same amount of time trying to fact-check an LLM (since you'll likely skim the articles it used in generating its response _which you would have done anyway if you used the search engine directly_).

I actually took a (required) class in middle school that taught us how to use a library. Amongst other things, the librarian taught us how to use Google effectively. Everything I learned then (this was in the early 2000s) still works today, since the process of using a search engine hasn't changed very much since its inception.

So many people never learned (or never cared about learning) how to use a search engine, thus why we're here today.

> you're almost always able to find what you're looking for very quickly[...] You'll spend the same amount of time trying to fact-check an LLM

I've been an expert Google user for over a decade and I can only partially agree with the first statement, and not at all with the second. Yes, a search engine alone is fantastic at finding things based on keywords if you know how to invoke it properly. However, there are lots of things one may want to find out which can't be reduced to a keyword search, because you can't have the vocabulary to search for it directly unless you already know the answer. Indirect questions such as "framework options to do x and y in z situation in this language". The best you could hope for pre-LLM was to find forum posts asking the same or a vaguely similar question and comparing a lot of options, finding out you picked a dud after spending an hour on it because it's fundamentally incompatible due to reasons, searching again, etc. It's hard to overstate what a massive improvement LLM's are for this kind of search to find and compare options for exactly what you're asking for given the context of your situation.

I guess it really comes down to how much faith you're willing to put into the answers LLMs are putting in front of you.

If you trust them blindly, then the vast experiment improvements are obvious and apparent.

If you don't, or if you're the kind of person that likes to research your sources, LLMs are a speed bump.

> because you can't have the vocabulary to search for it directly unless you already know the answer. Indirect questions such as "framework options to do x and y in z situation in this language". The best you could hope for pre-LLM was to find forum posts asking the same or a vaguely similar question and comparing a lot of options, finding out you picked a dud after spending an hour on it because it's fundamentally incompatible due to reasons, searching again, etc.

I disagree with this. Stack Overflow (pre-moderation insanity) forums and the like was and is great at finding answers to questions like this. Reddit threads were also useful for this sort of discussion. Much learning was had while reading through the comments on my way to the answer. Sometimes, doing that refined or re-aligned what I was looking for, as is common when doing research.

Again, it comes back to faith in LLMs. Sure, I can ask an LLM to give me a comprehensive overview of web serving frameworks for $LANGUAGE. It's up to the user to determine how valid the information being put in front of them is.

LLMs can also be extremely confidently incorrect. Example: I used an LLM recently in "research" mode to outline how a solution I sell stacks up to the next biggest competitor in pricing structures. It gave me a lot of (too much) information in a readily-digestible format, including, surprisingly, the "agreed-upon" price of the competitor's products per SKU.

Pricing for enterprise sales contracts is very dark arts, so I went to the sources attached to the result to confirm those numbers. Lo and behold, the prices I was given were nowhere to be found in any of those articles.

Meanwhile, I used a search engine manually to see if I could find a leaked price book using "filetype:pdf" operators, mostly for grins. Found it in 15 seconds. That still wasn't applicable to what I was looking for, as it was for an industry different from mine, but it was there.

I could have told the LLM to deep search PDFs (despite telling it to "ultrathink"), but at that point, again, what is the point of using LLMs if I can do the work myself?

What LLM was this? I'm simply not having this experience at any regular frequency these days with GPT Sol as my go-to. If I know I'll find something with a Google query, then I still go straight to Google because there would be no point waiting for the AI. For more vague cases where I don't quite know what I want to search or how, or I want an overview like you mentioned, I haven't had a significant amount of trouble, and I do verify when it matters. It's done well with broad comparisons like comparing options across eBay and Amazon for some odd category of used hardware for example.

In essence there are a few critical and vulnerable(imo) sources of info on search engines. forums, llms, and and just plain search by query matches and web crawling/indexing. They are all very useful but the last one is being shadowed and destroyed. Forums are also being filled with bots.

Google as a search engine got much worse, especially in the last year or so. The index also got noticeably smaller. My 20+ years of Google Search experience are now failing me completely.

They've disabled an bunch of useful operators and invested more into natural language search and Gemini answers. Another reason why I use Kagi.

What are these ways of using search that yield nom-garbage?

- Terms in "quotes" still work for finding results with a certain phrase in them. (the +plus Operator used to be great for this also but Google killed that)

- AND and OR still work, as do parentheses. Very useful when combined with the above.

- before: and after: are super useful for narrowing down a result to within a certain timeframe, like before AI. (Sadly I use this all of the time)

- filetype:$EXTENSION for finding files. I used to use this and "index of" to find interesting PDFs on misconfigured web servers, but the Internet has caught on.

- -minus Operator to exclude results that contain a term. Great for excluding subreddits

- inurl:$SUBSTRING or intitle:$SUBSTRING to scope results to pages with $SUBSTRING in their URLs or page titles. Indispensable for finding people on LinkedIn, like recruiters.

I've been searching online content since the 1980s (computerised library catalogues), the Web since the mid/late '90s, Google since 1999, DDG since 2013.

My search-fu is pretty good.

Correction: my search-fu was pretty good. Tactics which used to work, regularly, don't. And I'm often intentionally not searching keywords but framing questions (generally to DDG's Search Assistant) to suss out stuff.

And that works pretty well, for general online stuff.

That's a subset of my searches. Often I'll go directly to a site via DDG bangs, such as HN (!hn by:dredmorbius <something I said once>), Wikipedia (!w <topic>), or a few other specific sites.

Discussion sites are ... dead. Reddit is absolute shit. Far too many discussions have gone into closed channels (Birdsite, FB, Discord, etc.), or ... are simply no longer online.

I do fairly well on the Fediverse, though that's still a pretty small world. (Getting bigger, and punches above its weight or so I find.)

Search engines themselves seem 1) no longer comprehensively index everthing, 2) don't retain previously-indexed content, 3) have fucked with how search queries are processed to a degree that even if the content is online and indexed it's not found, and 4) have polluted SERP rankings such that top results are increasingly irrelevant.

(DDG, on that last point, does permit blocking sites from results, which is useful, though a pain in how I use the site, which doesn't persist such preferences across sessions.)

"Knowing how to use a search engine" no longer affords utility in a way it did, say, five years ago. And things were already starting to rot then.

Would be interesting to know if those same long tail results come up in Alt-Power [0], which also uses Google's index. So far I get what I ask for, but would be reassuring to know the whole long tail index is indeed shared and I'm not missing relevant results.

[0]: https://altpower.app

hum interesting never heard of this!

> just because of a PM in search "de/re-prioritizing" ranking in a way that makes this impossible

I could be wrong, but I'm more inclined to say these directives are coming from leadership/advertising dollars rather than a seemingly rogue PM

Reply on news.ycombinator.com