TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

I want flexible queries, not RAG

237 pointsby jmount12 months ago

29 comments

jillesvangurp12 months ago
The retrieval part of RAG is usually vector search; but it doesn&#x27;t have to be. Or at least not exclusively.<p>I&#x27;ve worked with various search backends for about 20 years. People treat vector search like magic pixie dust but the reality is that it&#x27;s not that great unless you heavily tune your models to your use cases. A well tuned manually crafted query goes a long way.<p>Pretty much any system I&#x27;ve built over the last few years, the best way to think about search is about building a search context that includes anything relevant to answering the user&#x27;s question. The user&#x27;s direct input is only a small part of that. In the case of mobile systems, the user entered query is actually typically a very minor part of it. People type two or three letters and then expect magic to happen. Vector search is completely useless in situations like that. Why does search on mobile work anyway? Because of everything else we know to create a query (user location, time zone, locale, past searches, preferences, etc.)<p>RAG isn&#x27;t any different. It&#x27;s just search where the search results are post processed by an LLM with whatever the user typed. The better the query and retrieval, the better the result. The LLM can&#x27;t rescue a poorly tuned search. But it can dig through a massive result of search results and extract key points.
评论 #40439139 未加载
评论 #40443601 未加载
评论 #40439191 未加载
评论 #40438102 未加载
mpweiher12 months ago
This has been my thinking as well: the natural language interface is <i>amazing</i> and something we&#x27;ve been wanting for some time.<p>The generation is a showy gimmick.<p>So why aren&#x27;t we separating the useful bit out? My sneaking suspicion is that we can&#x27;t. It&#x27;s a package deal in that there are no two parts, it&#x27;s just one big soup of free-associating some text with other text, the stochastic parrot.<p>LLM do not understand. They generate associated text that looks like an answer to the question. In order to separate the two parts, we&#x27;d need LLMs that understand.<p>That, apparently, is a lot harder.
评论 #40435258 未加载
评论 #40434323 未加载
评论 #40435669 未加载
评论 #40433702 未加载
评论 #40436806 未加载
评论 #40433673 未加载
评论 #40431672 未加载
评论 #40443982 未加载
评论 #40444175 未加载
bambax12 months ago
&gt; <i>There is a lot of excitement around retrieval augmented generation or “RAG.” Roughly the idea is: some of the deficiencies in current generative AI or large language models (LLMs) can be papered over by augmenting their hallucinations with links, references, and extracts from definitive source documents. I.e.: knocking the LLM back into the lane.</i><p>This seems like a misunderstanding of what RAG is. RAG is not used to try to anchor to reality a general LLM by somehow making it come up with sources and links. RAG is a technology to augment search engines with vector search and, yes, a natural language interface. This concerns, typically, &quot;small&#x27; search engines indexing a specific corpus. It lets them retrieve documents or document fragments that do not contain the terms in the search query, but that are conceptually similar (according to the encoder used).<p>RAG isn&#x27;t a cure for ChatGPT&#x27;s hallucinations, at all. It&#x27;s a tool to improve and go past inverted indexes.
评论 #40434598 未加载
评论 #40436649 未加载
评论 #40434703 未加载
评论 #40436825 未加载
评论 #40436105 未加载
评论 #40435957 未加载
评论 #40434665 未加载
评论 #40435151 未加载
评论 #40435387 未加载
ritzaco12 months ago
I don&#x27;t think the author completely understands RAG, and this article is a bit disconnected and unclear to me. Google already provides a &quot;flexible natural language query interface&quot;.<p>I think ironically it would be fairly trivial to build what he wants _using_ RAG.<p>1. Accept a natural language query, like ChatGPT et al already do<p>2. Ask an LLM to rephrase it in N different ways, optimized for Google searches<p>3. Scrape the top M pages for each of the output of 2 in parallel. You now have dozens of search results<p>4. Clean and vectorize all of these<p>5. Use either vector similarity or an LLM to return the best matching snippets for the original query from 1, constrained to stuff contained in 4.<p>It would take a little longer than a ChatGPT or Google response, but I can see the appeal too.
评论 #40444626 未加载
评论 #40444766 未加载
评论 #40444120 未加载
ilaksh12 months ago
The article would be more convincing if they showed the recipe from the book so we could compare it with the one that ChatGPT output.<p>From a Google search, it looks like he&#x27;s right about the poor accuracy. It gets the basic idea of the ingredients, but is not really accurate. And is initially wrong about the region.<p>But actually, this is what RAG is for. You would typically do a vector search for something similar to the question about &quot;rice baked in an egg mixture&quot;. And assuming it found a match on the real recipe or on a few similar possibilities, feed those into the prompt for the LLM to incorporate.<p>So if you have a well indexed recipe database and large context window to include multiple possible matches, then RAG would probably work perfectly for this case.
评论 #40434566 未加载
评论 #40434582 未加载
valstu12 months ago
We use the term &quot;pre-googling&quot; for this sort of &quot;information retrieval&quot;. You might have some concept in your head and you want to know the exact term for it, once you get the term you&#x27;re looking for from LLM you&#x27;ll move to Google and search the &quot;facts&quot;.<p>This might be a weird example for native english speakers but recently I just couldn&#x27;t remember the term for graph where you&#x27;re allowed to move in one direction and cannot do loops. LLM gave me the answer (directed acyclic graph or DAG)right away. Once I got the term I was looking for I moved on to Google search.<p>Same &quot;pre-googling&quot; works if you don&#x27;t know if some concept exits.
评论 #40438483 未加载
评论 #40441715 未加载
advisedwang12 months ago
Isn&#x27;t the point of RAGs to make (in this example) actual recipe databases accessible to the LLM? Wouldn&#x27;t it get <i>closer</i> to the articles stated goal of getting the actual recipie?
评论 #40436166 未加载
评论 #40435317 未加载
评论 #40434721 未加载
pseudosavant12 months ago
I think complaints like this show just how amazing AI is getting. This person really expected that ChatGPT would single-shot give them this obscure recipe that took them a ton of effort to find themselves. Current AI can do so much, that people lament that it can&#x27;t do everything. It is incredible to me when people bring up bad AI generated legal fillings... like people actually expect it to already to all of the work of an attorney, without error.
评论 #40434586 未加载
评论 #40434577 未加载
评论 #40442883 未加载
surfingdino12 months ago
&gt; I wanted the retrieval of a good recipe, not an amalgam of “things that plausibly look like recipes.”<p>And that&#x27;s the core issue with AI. It is not meant to give you answers, but to construct output that looks like an answer. How is that useful I fail to understand.
评论 #40443079 未加载
评论 #40444217 未加载
评论 #40442841 未加载
评论 #40444021 未加载
LASR12 months ago
I continue to hold the strong position that calling LLMs without injecting source truths is pointless.<p>LLMs are exceptionally powerful as a reasoning engine. It’s useless as a source of truths or facts.<p>We have chat bots, chat bots with automatic RAG etc. After the initial excitement wears off, you’re going to want a way to inspect and adjust the source queries yourself. In this case, being able to select what to search for in Google might be a good way for the cooking recipe usecase.
评论 #40434318 未加载
评论 #40436916 未加载
martin29312 months ago
I do not understand the point of this article.<p>You did not even tell us what the correct recipe was called. But let&#x27;s ignore that for now.<p>I did some googling around, and the Wikipedia article for sartu di riso [0] mentions the fact that &quot;it (the dish) found success in Sicily&quot;. Also, in [1] a commenter going by the name of passalamoda mentions that they also make this dish in Sicily. The comment they wrote is in Italy, but Frank Fariello has translated it for us, or if you don&#x27;t believe him for whatever reason, google translate does a fine job. All of this is to say that associating this dish with Sicily based on the short description you&#x27;ve given, is not far fetched <i>at all</i>.<p>&gt; I fail to see how an LLM summarizing the material would be an improvement.<p>I am fairly confident that typing the question you gave ChatGPT, waiting a few seconds for an answer, and then reading it can easily take under a minute. Lets be lenient, and say it takes 5 minutes to also ask a second question and receive the answer. That would still take way way <i>way</i> less time then to find a book, get the book and go through the book to find the correct recipe.<p>Also, you yourself have given a reason in your article as to why ChatGPT would be an improvement. I will quote it now:<p>&gt; Most of her time was dealing with the brittleness of the query interface (depending on word matching and source popularity), spam, and locked down sources.<p>I have already spent way too much time on debunking a random internet article, but I also decided to try to get an answer from ChatGPT. I did that by continuing to ask it questions that a person looking for an answer, not contradictions, would ask. If we make the assumption, that deanputney is correct, and the dish you were looking for is Arancini al Burru, we are able to get an answer from ChatGPT by asking the very simple and natural question shown in [2].<p>[0] <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Sart%C3%B9_di_riso" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Sart%C3%B9_di_riso</a><p>[1] <a href="https:&#x2F;&#x2F;memoriediangelina.com&#x2F;2013&#x2F;01&#x2F;21&#x2F;sartu-di-riso-neapolitan-rice-timbale&#x2F;" rel="nofollow">https:&#x2F;&#x2F;memoriediangelina.com&#x2F;2013&#x2F;01&#x2F;21&#x2F;sartu-di-riso-neapo...</a><p>[2] <a href="https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;PKqGXyK" rel="nofollow">https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;PKqGXyK</a>
PaulHoule12 months ago
… you really want to be able to cook documents down to facts, as in the old A.I., and then be able to make logical queries. Trouble is it is easy to ontologize some things (ingredients) but not so easy to ontologize the aspects of things that make things memorable.
antirez12 months ago
Suggested book about traditional Sicilian dishes: <a href="https:&#x2F;&#x2F;www.amazon.it&#x2F;Profumi-Sicilia-libro-cucina-siciliana&#x2F;dp&#x2F;8886803737" rel="nofollow">https:&#x2F;&#x2F;www.amazon.it&#x2F;Profumi-Sicilia-libro-cucina-siciliana...</a>
评论 #40434210 未加载
评论 #40434059 未加载
jszymborski12 months ago
Is that ChatGPT example representative of RAG? I thought ChatGPT was primarily generative.<p>I think of something like Brave Search&#x27;s AI feature when I think of RAG.
评论 #40431549 未加载
lmeyerov12 months ago
To keep control in the hands of the analyst, we&#x27;ve been working on UX&#x27;s over agentic neurosymbolic RAG in louie.ai --<p>Ex: &quot;search for login alerts from the morning, and if none, expand to the full day&quot;<p>That requires generating a one-shot query combining semantic search + symbolic filters, and an LLM-reasoned agentic loop recovering if it turns up not enough such as a poorly formed query around &#x27;login alerts&#x27; and the user&#x27;s trigger around &#x27;if none&#x27;<p>Likewise, unlike Disneyified consumer tools like chatgpt and perplexity that are designed to hide what is happening, we work with analysts who need visibility and control. That means designing search so subqueries and decisions flow back to the user in an understandable way: they need to inspect what is happening and be confident they missed nothing, and edit via natural language or their own queries when they want to proceed<p>Crazy days!
评论 #40436944 未加载
评论 #40438241 未加载
khaki5412 months ago
Clearly this author doesn&#x27;t know what RAG is. RAG would be if he first did a &#x27;retrieval&#x27; of all his mother&#x27;s cookbooks containing Italian recipes, then reviewed the index for rice, scanned and OCRd those pages. That data would be submitted to ChatGPT with the query to &#x27;augment&#x27; it within the constraints of the context window so that ChatGPT could generate a response with the highly relevant cookbook info.
WhatIsDukkha12 months ago
Your input was incredibly low effort and prompted a very low effort output.<p>I took part of your blog post (which you clear were willing to put a few more tokens into) -<p>&quot;My mother remembers growing up with a sicilian dish that was primarily rice baked in an egg mixture. Roughly a &quot;rice frittata&quot;. What are some distinctly Sicilian dishes that this could be referring to?&quot;<p>Notice there is not much extra context that you&#x27;ve offered any of us, either the LLM or us. You didn&#x27;t even tell us what the recipe was...<p>How was the dish served, what did it look like?<p>What are you expecting of the LLM here? It not a psychic AGI.
评论 #40434167 未加载
a_c12 months ago
Querying knowledge is not a nail, but the hammer is generation. It is written on the tin, &quot;generation&quot; AI. People want &quot;insight&quot;, &quot;summary&quot;, &quot;workflow automation&quot;, &quot;code completion&quot; from a guided proverbial monkey hammering on the keyboard hoping that our problem will become a nail. It is getting closer though
lukol12 months ago
Thanks for sharing your thoughts! I published a blog entry that describes our solution to this exact problem yesterday. We call it &quot;Generation Augmented Retrieval (GAR)&quot;, please find the details here: <a href="https:&#x2F;&#x2F;blog.luk.sh&#x2F;rag-vs-gar" rel="nofollow">https:&#x2F;&#x2F;blog.luk.sh&#x2F;rag-vs-gar</a>
ashu146112 months ago
I think the author is over generalising things. Every tech is good for few things and not good for others.<p>Vector Search, LLMs all are revolutionary technologies and have their own limitations but we should not form a bias on the basis of few edge cases.
budududuroiu12 months ago
An interface for vector search is 100x more helpful to me than an LLM spitting out the same content as slop.<p>The key to vector search is how you chunk your data, but I have some libraries to help with that
prionassembly12 months ago
Is anyone doing something like using LLMs to generate Prolog (or Cyc, or some appropriately complex, brittle knowledge representation GOFAI)?
slushy-chivalry12 months ago
&gt; To me, LLM query management and retrieval is much more valuable than response generation<p>Sure, for certain tasks. For other tasks retrieval is less useful.
amath12 months ago
This is just that authors opinion. It seems clear that there is a spectrum of what users want from basic retrieval to content generation.
评论 #40441920 未加载
immibis12 months ago
Users don&#x27;t want what they think they want. Therefore, we won&#x27;t give users what they think they want.
more_corn12 months ago
lol the thing they’re asking for is literally rag. “I want a smart person to parse relevant information and return a concise and relevant answer” Take the query, find the book, dump the book into the context window and the LLM’s answer will be exactly what you want.
mrbonner12 months ago
How do you delete or update entries in a vector database?
评论 #40441147 未加载
kunalgupta12 months ago
this person just wants perplexity
评论 #40444058 未加载
humanlity12 months ago
if the data is all structured and can be easily trained, maybe the ML can be even better