Ironic to use AI writing to complain about AI search results to advertise your product. But to the point, being an old school firefox user I still have separate address and search bars. My main search engine on my PC is 'google (web only)' which opens directly to the web tab which only has result links. I'm dreading the day they remove it as a feature.
We have been getting a ton of LLM generated essays talking about how bad LLMs are. It happens so much that it feels like a new kind of engagement bait.
There's a recognizable style to the overall piece.
The subheadings for example "The summary is not the answer" and "The verification problem is real and it compounds" are typical being punchy but not really making much sense.
Also consider
>This is slower. That's real. It's more work. And it is, in practice, more accurate — because accuracy in information retrieval is a function of reading, not of being read to.
The over punchy - It's this. It's that. It's the other - is typical or LLMs. The accuracy being a function of reading not being read to doesn't really make sense - with both normal results and AI you are reading.
I'd prefer more proof / context than GP gave, but I personally find it very useful to see people making judgements about AI-assistance of articles. Almost no such articles are worth my time, and the more HN people saying it, the more I know not to click past the HN headline.
Same as a NSFW tag. It's not about adding to the discussion; it's a very brief warning to users who don't want to see that kind of content, and its brevity makes it come at almost no cost to users who don't care about SFW/NSFW or AI/OC distinctions.
I agree I run the risk of being wrong and could at least provide some evidence, but I think at the very least it can be one additional piece of information someone could use in their consumption of this content.
If you can't be bothered to write even one specific thing from the article content you object to, then your account might as well be a pre-LLM bot that just posts "first" or "LLM Content" which is somehow even lazier than an LLM blog post.
I don't want either, if I'm indeed "searching." But I find that often times I am indeed just looking for a quick answer, and Gemini/Google's "new" search does it fine.
It's one of the few AI features, despite still being shoved in my face, that I actually find useful.
With that said, the worst thing is how search results have degraded significantly since the AI years, even before they added the actual "AI mode."
Google now (and quite a few search features on other services, e.g., Twitter) often returns results that have ZERO relationship to the search keywords I gave -- like an entirely different person when searching for a person's name, which I think should never happen and did not happen when search was still based on a "rigid" algorithm of indexed content. So, I can only assume it's because they have some AI thingy along the process.
Sometimes it's not very clear what I really want when typing.
Also, Google Search's AI answer doesn't prevent the actual search results from still showing below; so if I later realize I need those (search results) instead, they're already there.
Also, I use browsers all the time, so typing things into the location bar is quicker than deliberately opening the Gemini app/website or whatever.
I value one fewer click/action a lot when such an action will be performed hundreds of times a day.
I love AI summary and AI mode in Google search. I think it should be up to you (the human) to use your judgement to decide when to do further research by following the links and when to just rely on the AI summary. If I'm searching a TV show by name, I'm generally just looking for an overview, the IMDb rating and a review of that show. If AI summary gets that wrong, it's not the end of the world. I don't bother doing further research. There are a lot of such casual searches I do daily for which AI summary is good enough.
OTOH if I'm looking up an answer to a tax question, I don't just immediately trust the first answer from AI mode. I use it more as a knowledgeable friend who is not a tax attorney and so cannot be 100% trusted, but he/she is giving me useful pointers to go do deeper research and arrive at an answer.
Same. AI still makes a lot of mistakes. Like today I was looking for new smoke detectors, and it clarified and then forgot Washington state code multiple times, and then ignored alternatives that were better fit for my problem until I asked pointed questions, and kept recommending solutions that I absolutely didn’t want to consider (or couldn’t because of WA code). If it wasn’t a conversation, it would have gone from a great to a disastrous experience.
well the "AI" is smart because the search is "DUMB" if I google "how long do I fry an egg" I don't need an article with 600 words about the ladys grandmums special egg recipe, but thats how SEO got us here.
I left google for chagpt/claude, but have slowly been coming back because these results have become rather useful -- especially when google inspects actual video content to produce replies. This is one of the best outcomes of the AI wave for me: content which was locked away in YouTube videos and forced me to consume it sequentially (without power-reading as I do with articles) is now indexed in search results in some AI summary. It makes mistakes, but it lets me figure out the right videos to watch, and often gives me enough context to find the solution myself. This applies to work, but also to something as simple as game walkthroughs!
As with everything, you need to know when to trust and when to doubt, when to iterate on chat (or chat with it with a smarter assistant like ChatGPT/claude). Hasn't it always been this way (minus the chat)?
I don't understand why people seem to be under this idea that just because it's at the top of a search results page it must be entirely correct, when we have spent decades knowing that you can't trust what's on the internet and you need to have critical thinking. Were you blindly trusting your first-page google results? What about when Wikipedia was starting, were you also trusting it blindly? Do you trust it blindly now?
In fact, when google started stealing the content of websites to display it, wasn't it already producing largely incorrect results because so many hits were SEO-garbage or outdated? (Not that we don't need to address the fact that this steals traffic from websites -- but it's a problem that PREDATES AI and absolutely still needs to be addressed somehow)
I can understand it if they believe the average joe will be more likely to blindly trust this, but that's an education problem -- and one we already had and has perhaps been compounded.
The visceral hate for a a life-changing technology that enables us to do more, faster, never ceases to amaze me. I'm growing so tired of it around my inner circles.
I get that there is a real bubble with real overhype and that some companies are profiting from it and shoving it down everyone's throats. I get that your actual life is worse because of this (you are forced to use shit models to produce more features while being paid the same). I get that maybe your job might change, and that for some people they will actually need to job-hop. This all absolutely sucks and it is a problem for many people, I get that! But that doesn't make the technology as a whole terrible, much less a bunch of its amazing applications. It's this constant wave of negativity, like nothing good will ever come out of AI and we're all doomed to be cornered by the big evil corporations that will run us dry.
It's hard to be rational when you're overcome with emotion and fear, I know, but it's so aggravating to have to interact with someone who can only see the technology itself as negative.
Maybe it’s misplaced nostalgia but google search before google plus ruined the +, and the image search was top tier (not shopping ads) - truly incredible the signal to noise you could find.
Now all the junk comes to the top and the sites you get all have ads and modal popups or sales funnel flows
Of course - it's SEO's fault that Google search frequently ignores words in your queries.
It's SEO's fault that Google frequently prioritizes giving you results that are semantically or thematically related but ultimately irrelevant to what you searched for.
It's SEO's fault that Google refuses to place ads on websites not full of meandering irrelevant "content".
It's SEO's fault Google search all-but-ignores most single-purpose tool websites.
It's not the result of decisions made by Google's programmers and project managers - it's allllll the fault of SEO!
The SEO industry definitely isn't an all-purpose scapegoat for horribly unpopular decisions made by human beings at Google.
Couldn't disagree with the article more. Not wading knee deep through SEO chaff is probably my favorite thing about LLMs. In the rare circumstance that I feel the need to wade through chaff that option is of course still available to me.
> When you search for something, you're usually not looking for a sentence. You're looking for evidence.
There is a long and storied history of Google offering more than just a list of links to go search for, since at least 2012, because a massive amount of people literally are looking for the single answer to a question, whether stated explicitly or implicit in the search term.
When LLM results started being shoved into standard result pages, I immediately added them to my ever-growing Userscript, CSS mods and uBO filters for https://www.google.com/search*.
That said, I DO sometimes prompt explicit web searches via LLM because for certain searches, an LLM can be the best tool for the job. It's almost always cases where it's hard to target terms with inclusions, exclusions and booleans (even if Google hadn't partially nerfed them). This is usually due to English language ambiguities in word meaning and LLMs are actually ideal for that because they already have discrete tokens for "Wear" (as in clothes) and "Wear" (as in tires).
My conflict with Google product managers arises because I'm an advanced search user and I know when an LLM is the best tool for a search and when it's not. And when it's not, LLM results are always cognitive noise that just gets in the way and eats screen-space. I understand that most users may not be able to reliably make that distinction but Google refuses to offer any option for users like me to enable a functional experience, even buried in an advanced menu.
I don't think this gap is just monetization driven. I've worked as a senior group PM in a FAANG-ish tech company and I suspect the problem is related to a deeper issue I saw in many UX designers and product managers - an inability to accurately model the frequency and quantity of ways "automatic" or default options can fail to serve users. It's an insidious issue because it tends to appear as the absence of signal in normal interface usage analytics. And it's hard to detect in focus group and observational studies, unless you aggressively look for it. And they usually don't because there's an innate bias pushing them toward believing the automatic "it just works" solution serves everyone as well as they've designed it to. They insist it must be right because it's exactly the sort of elegant, simple solution their D-School profs graded highly.
There will be a learning curve for some, like the person behind this article. Searching in AI mode will mean your query is going to need to be a bit more pointed.
The example given is you want evidence from an original source in Stack Overflow. So instead of just typing a few keywords and digging through pages, your query needs to ask Google exactly what you want and the format you want it in. If that’s a list of Stack Overflow pages, then that’s what you ask for. You can test this out now with AI summaries and a well-written query.
The quality of your results is going to depend on what you put into it. It will probably be annoying for some at first but for those that get it it’s going to be a step up.
No, those search engines will simply stop being used by people that care. Google's "AI results" have already wrongly accused people of being sex offenders and make up gibberish constantly. They are a blight and a scourge. There are already many lawsuits against the garbage it invents.
The "quality" of your prompt is irrelevant when you're feeding it to something that just imagines things.
As he said, its just a learning curve and you are behind on it. Dont use AI mode use perplexity or gpt for your search. its far superior to traditional search, just slower. The quality of your prompt also matters. Im pretty annoyed by those like you who hide and mask their ignorance behind anything they can grasp to cover themselves. It's transparent to most but yourself. Ignorance i could forgive, but not the dishonesty.
> There will be a learning curve for some, like the person behind this article.
That's ridiculous. There isn't a learning curve, we don't want AI. No amount of "learning" is going to make these misbegotten features into something that actually improves my experience with the product. You should listen to people, not assume that you know their desires better than they do.
Well tech companies have long been hijacking our brains thinking muscles. Remember how people used to be able to navigate on their own without a Maps app, now those same people can't get around their own town without their phone.
I'm genuinely scared for a generation of people who've offshored their thinking, planning and creativity muscles to a few tech companies.
We think we're gaining an edge but we're really participating in a mind control experiment thats optimized to benefit those companies, not us as individual.
Miss me with AI, it will break your brain and start to control more and more of your behavior if you let it. Don't become a drone. You're not going to become some crazy productive SaaS founder becauae you have AI, you'll become a drone who's competency is 1:1 correlated to the quality and quantity of tokens you have access too/
"Search" is a ridiculous thing to be doing post-2022. Imagine going to a doctor and asking them a question, and they give you 5 printouts for your to read through to synthesize your own answer. Imagine you asked your spouse a question and they responded "Here's 10 links for you to check out!"
We have AI now and it's doing a mostly incredible job getting us ANSWERS, not SEARCH LINKS. Trying to pretend that links are better is just trying to copy with rapid change.
Quite honestly I'm shocked that Google keeps making more money with search ads because I don't search anymore, I get answers directly from it or ChatGPT without clicking on any links.
> We have AI now and it's doing a mostly incredible job getting us ANSWERS, not SEARCH LINKS. Trying to pretend that links are better is just trying to copy with rapid change.
If you feel that it's better, you are free to use it. Leave the rest of us, who think AI is bad at finding information and want to search, in peace so we can work the way we like. I don't think that's an unreasonable expectation: you can work the way you like, and I can work the way I like.
I can't empathize with this mindset at all. I'm in the polar opposite camp. I don't want a list of search articles; just a direct answer to the information I was looking for. I see the problems pointed out by OP but they are being solved away.
I use SearchTool in the Agent program, but when using search engines, I usually use simple keywords and the search results may contain too much noise, but I still use the AI search results as a reference, which is not an either A or B thing for me
Google AI responses are generally crap and annoying. I've gone back to DDG or if I need some context - very specific guidance for claude/chatgpt. Goodle's AI is generally inadequate or wrong without significant clarification.
Totally get where you're coming from. Sometimes I just want straight answers without all the algorithmic guessing. Do you find yourself using quotes a lot to get more precise results?
i heard that a lot of anti-ai stuff is written by AI because the people reading it aren't prepared to recognize if it was AI-gen because they themselves don't use ai
I hate those AI summaries. Because I dont necessarily agree with suitability and credibility ranks assigned by AI to make those summaries. As author says, there are so many nuances and I usually scan the results page and click that appears more credible first. Not what appears first. I also know which site fills copy with verbiage and which ones give more useful advise - such as in health matters.
If you're like me, basically everything other than search results in a search engine is noise. To that end, I direct my Google searches directly to the "web" tab of the results. This chops off news, shopping, video, images, etc from the initial results you see but you can still change to those tabs just as easily as ever. In your browser's settings just find the option to manage search engines, add a new one, and set the url to use to https://www.google.com/search?q=%s&udm=14 and you land on results. No AI misinterpretation of what you're looking for. No confident nonsense that directly contradicts reality sitting at the top of the page. No judgement or chiding (Google's results AI has flat out told me my opinion sucks mid-search more times than I'd like to admit). Just results.
I've switched back to the pre-llm / AI summaries period when I extensively search for the results and also discover tons of great web resources along the way. The novelty of instant answers from the chatbot is gone. Now I just deliberately ignore the top AI summaries and go straight for the ten blue links.
If I need a chatbot I'll launch the app. Mixing the AI answers in SERP is too much noise for me.
Your search engine doesn't think for you. It never did. It still doesn't, even with AI summaries. What it does do, is summarize information.
It's up to you, the human, to decide what path you get the information and how to digest it and use it. I use LLMs for 99% of my "querying" these days because 99% of the information I'm looking for is either verifiable (coding) or low-stakes (how to do low-risk DIY house thing).
If I need actual search, I use Kagi. Otherwise, Claude has taken 99.8% of what Google used to do for me and it feels better than ever.