TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

AI Restaurant Menu with RAG

61 点作者 byyoung3超过 1 年前

13 条评论

chadash超过 1 年前
&gt; <i>Physical menus are fine. They do the job. But if we&#x27;re looking to actually improve upon them, current approaches miss the mark. They offer limited interactivity and flexibility.</i><p>Sigh. Am I the only one who likes to hold a physical menu and browse it? I never go into a restaurant knowing what I’m gonna get… I like to see how the menu is themed and decide from there. The UI on a physical menu is a vast improvement over any digital menu I’ve seen.
评论 #39415330 未加载
评论 #39415538 未加载
评论 #39415454 未加载
评论 #39415284 未加载
评论 #39415413 未加载
评论 #39415480 未加载
评论 #39415342 未加载
评论 #39415308 未加载
评论 #39415713 未加载
评论 #39420398 未加载
评论 #39415372 未加载
评论 #39415517 未加载
minimaxir超过 1 年前
The rather complicated prompt to force the recipes to be in JSON w&#x2F; constraints can be substantially simplified using ChatGPT&#x27;s function calling&#x2F;structured data capabilities: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38782678">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=38782678</a><p>Additionally, a full vector store is overkill for a menu with &lt;100 items. You can just keep a numpy array in memory with vectors for that and use a matmul for RAG.
评论 #39415467 未加载
评论 #39415306 未加载
sebastiennight超过 1 年前
Aha interesting. I would guess it is missing a bit of nuance, eg a real word query would not be &quot;appetizers with chicken&quot; but something closer to &quot;I&#x27;m waiting for my friends, not too hungry, what do you have that&#x27;s spicy but not too filling? Oh, and I don&#x27;t eat red meat&quot;.<p>This would be much harder with basic vector database+RAG. You would at least need HyDE (hypothetical document embeddings) to act like a real waiter.<p>I find your project very interesting though, and the walkthrough is nice!
评论 #39415187 未加载
评论 #39415177 未加载
Animats超过 1 年前
For a single menu, this is silly. As a search engine for Doordash or Uber Eats, though, it might lead to this guy being acquired.
评论 #39416474 未加载
评论 #39415020 未加载
jumploops超过 1 年前
Semantic search with vector embeddings is great... as long as you maintain an inverted index for traditional full-text search.<p>As a user I sometimes want a &quot;buffalo chicken quesadilla&quot; not disparate buffalo wings and quesadilla recommendations.<p>I believe Elastic has a mechanism that supports both? Haven&#x27;t tested it yet.
评论 #39415304 未加载
steveBK123超过 1 年前
Using a chatbot to answer customer questions about menu items seems like a great way to kill someone with food allergies due to LLM hallucination, but hey.. AI!
评论 #39415503 未加载
cadence-超过 1 年前
RAG is probably the only thing you can do with LLMs nowadays that is useful, accurate, and cheap enough to deploy in a real business scenario. Every other use of LLMs so far is either too expensive to scale or produces too many hallucinations.
morbicer超过 1 年前
Different stokes for different folks but this is dystopian to me.<p>I prefer the approach of standard french restaurant: 2-3 options for each category of appetizer, main course and desert. Don&#x27;t paralyze me with choices. Too many options also mean the establishment has no focus - they won&#x27;t be all good so it&#x27;s a lottery.<p>If I have question I will ask the waiter. And I am as asocial and introverted as any common nerd.
charcircuit超过 1 年前
This is way over complicated for querying a single menu. This article seems like a solution in search of a problem. You don&#x27;t need a search engine when a menu is only going to have a handful of dishes.<p>Just use an picture of the menu as part of the prompt and you will make everyone&#x27;s lives easier. I&#x27;m skeptical that asking questions about the menu has much real utility.
solarized超过 1 年前
If we&#x27;ve fixed the 24&#x2F;7 &#x27;free&#x27; internet with zero latency issue, 24&#x2F;7 electricity, and cheap mobile devices – in short, ZERO FRICTION – then this kind of tech could all work. Otherwise, we&#x27;re just adding new problems instead of fixing them.
selimnairb超过 1 年前
There are already too many restaurants owned by small business tyrants who feel entitled to run a successful small business. If you don&#x27;t want to make your restaurant menu by yourself, maybe you shouldn&#x27;t be in the restaurant business?
评论 #39415275 未加载
cadence-超过 1 年前
Is there any specific reason for using LLAMA embedding models instead of the new models OpenAI just released? Are they better?
评论 #39415328 未加载
评论 #39415520 未加载
m3kw9超过 1 年前
Nice idea