I’m confused. Is the OP using Google the search engine? Because I didn’t think Google answered questions; I thought Google searched for websites that contain the words and phrases you ask it to search for. Which is a different thing.
Nah, this pitting is deserved. Google should not be forcing its AI to the top of the results when it’s so bad. It shouldn’t be answering a completely different question. That’s what you did before LLMs.
Most people I talk to think the AI response is being shoved on to people to meet some sort of quota. That’s the only reason they won’t let you turn it off. Not even after they were roasted for all of the hallucinations it has, like telling people to put glue on pizza because a Reddit post jokingly said that once.
In general, I’m not a fan of using AI as a search engine, as any chance of hallucination means I have to fact check it–which is what I would have already do with a regular search. So I feel it just wastes time.
But Google’s is worse than that.
But a properly-constructed natural language question can, sometimes, if you’re lucky, function as properly-constructed search query. That’s what happens when your extra interrogative sentence fluff doesn’t misdirect the search. But more importantly, if the keywords of your search happen to be popular enough and current enough to come up in the first dozen pages of the search results.
For obscure or dated searches, you’re more likely to get page after page of the flavor of the month.
Then you clearly understand nothing about the economics of the internet which is not the same thing as the web.
You pay good money to connect to the internet plumbing that enables your computer to send and receive packets from other computers. That is what you are paying for. And nothing more.
Google is performing a service on the web for advertisers, and only for advertisers. They are the only people who pay Google anything and therefore they are the only people whose interests Google is paying attention to.
How does Google get advertisers to pay them? By creating a set of free-to-use features just attractive enough to consumers that consumers come back to Google over and over to use the free services, and thereby to be served more ads along the way. Your only role in this triangle is to be the victim. And evidently a clueless victim at that.
Until we are each paying a buck or two to perform a search (or an AI query) we are not the customer and should expect the service we are using to do the bare minimum to not piss us off too severely. Unless of course they are a de facto monopoly in which case, just Like Ernestine’s Telephone Company, they are free to consider themselves Omnipotent and all of the end users impotent. Because they are and we made them that way.
If I ask it to find “red tennis shoes size 12w” I occasionally get make something like “Answers do not include “red”. Did you want to include “red” in your query?”
Wutthefuk?? Yes, I want to include “red” in my query, dumbfuck, and where is the “Fuck you up the ass sideways” emoji when I really need it?
Most of the humans I know rarely answer the question I asked.
Given the extreme skill with which most people, including our OP, ask questions, I’m not surprised. Answerers in general have been taught to ignore what you (any you) say and attempt to divine what you meant. Or they’re just lazy and stupid. Not knowing who you interact with I can’t say which. But if your cross-section of humanity is anything like mine, one of those two answers is the right one.
Viewed from the opposite end of the telescope …
If you want to royally piss off a lot of people, simply answer the question they did ask, not the one they intended.
I’ve long thought that I’d pay extra for those times that I want an answer from a human being… maybe a retired librarian who’s a genius in many fields.
Eh, I’d say at least 90% of the time Google gives me a good answer if I know how to ask the question properly. I have indeed been frustrated at times like the OP when it’s clearly ignoring what I’m asking and giving me not only the wrong answers, but sometimes the opposite of what I’m asking for.
It’s still better than any other I tried.
I’ll give a real life example from a week ago. I was in the living room with my wife and she wanted to look up an old Almost Live! sketch. (For those unfamiliar, it was a sketch comedy show aired locally years ago from out of Seattle; people like Bill Nye and Joel McHale got their start there.) Since we were in the living room with a huge TV and an Xbox connected it was natural to try to look it up there so we could both watch it.
First we tried the YouTube app but it was obviously not there. So then we launched the console’s web browser (Edge) to look it up. We spent quite a few frustrated minutes trying to get the damn thing to find it. I then realized, hold on, this piece of shit is using Bing which is like the village idiot of search engines. Let me go to Google’s web site and search.
The very first result of the first search (same search terms we tried before) went right to the sketch and we were able to watch it from the browser.
I don’t even try to use other search engines, so when I accidentally use another one it reminds me how bad they are. I should probably do that now and then when Google pisses me off to remind me how good I have it.
Strikethrough is typed like either of these examples:
~~strikethrough~~ or <s>strikethrough</s>
Either of which results in being rendered like this:
strikethroughorstrikethrough
The <del> code is not correctly used for simple strikethrough. Its purpose is to be used along with <ins> for indicating revisions to existing text. e.g.
Version 1: The sly brown fox stupidly spilled his beer.
Version 2: The clumsy brown fox stupidly spilled his coffee.
Marked up version showing both the original and the modified text:
: The sly clumsy brown fox stupidly spilled his beer coffee.
The exact appearance of inserted and deleted text will vary by theme. Under the SD light theme I (and many other Dopers) use, the deleted text is highlighted pink and is struck through, while the inserted text is highlighted green and underlined. Different themes will have different treatment of the two sorts of text.
Conversely, all themes render the ~~ or <s></s> as simple strikethrough with no other decorations. So that’s the one to use for the rhetorical device you’re looking for.
I forgot to mention: I block the AI nonsense on Google, and have since it started being forced on me. I use an extension, though in theory a userscript could also do it. (Adblock by itself would have more trouble, as Google uses random IDs for elements of the page, which sucks.)
It used to that. Now it searches for the words and phrases it thinks you mean. So it’s better for getting the right results if you are making a very common query and misspell something, but far worse if you are searching for a very specific uncommon thing that sounds a bit like a very common thing
Why the comments saying that Google isn’t an AI? Of course it is, and it was long before ChatGPT and its ilk existed. It’s a different kind of AI from ChatGPT, but that’s not the only kind of AI out there.
And Google is trying very hard to answer your question. The problem is that it doesn’t know what your actual question is. It only knows what the question you asked is. People, as you may have noticed, are by and large idiots, and have a very strong and very annoying tendency to ask different questions than the ones they actually want answered. So when you’re helping people, the first step is always to figure out just what it is they actually need. Which is a really hard problem, and it’s frankly amazing that Google is able to correctly solve it as often as it does.
Then by telling me that (instead of arbitrarily substituting its own question and answering that), it would be training me to formulate my questions better.
Google is doing no such thing. It cares not in the slightest what you actually what you want to know. It’s figuring out how to balance their need to monetize search, vs. cost of performing that search, vs. giving you just enough utility to prevent you from switching to a different search engine.
No, the first step is to give them exactly what they asked for. Do not presume. If that doesn’t work, then you try to make helpful inferences. Do not “help” me when I’m explicitly directing you not to help me.
As a basic example, Google no longer reliably respects fine-tuned search directive. Consider this search:
hydrangea +neutronium
I have told it that I need a search result about hydrangeas, but it must also contain neutronium. We’d expect this search to return nothing, because these topics have no relationship to each other. A negative result would confirm the assumption, while a positive result would teach me novel and interesting facts about the relationship between hydrangeas and neutronium.
But Google doesn’t gain revenue from informing me, it gains revenue from search results, so it force-feeds me plausible-seeming search results that not only fail to answer my question, but fail to return any new insights at all.
When Google answers questions other than the one the user wants answered, the user becomes upset, and becomes less likely to use Google and to see the targeted ads that Google is paid for. When Google answers the questions users actually want answered, the users are happy, and use Google more, and see more of the targeted ads. And so, what’s best for Google’s bottom line, is to start by answering the question that it thinks is the most likely to be the actual question. Which very often isn’t the question actually asked, because see above about people being idiots.
You have to use quotes. Search for “hydrangea” +neutronium and it indeed comes up with zero results.
The last residual effect of Google Plus is this annoyance.
Does no one else use Google Advanced Search?
I do. That’s my default.