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.

Show HN: ChatGPT UI for rabbit holes

766 pointsby maxkrieger12 months ago
I was inspired by the way ChatGPT writes bullet lists, then invites you to &quot;delve&quot; deeper.<p>This is an interface that reifies that rabbit-holing process into a tiling layout. The model is instructed to output hyperlink-prompts when it mentions something you might want to delve into.<p>Lots of features to add (sessions, sharing, navigation, highlight-to-delve, images, ...). Would love to hear other usecases and ideas!

89 comments

gfiorav12 months ago
This is very good. I can&#x27;t put my finger on it, but it seems more important than a mere &quot;gimmick.&quot; I noticed that if you click on a topic already explored, it won&#x27;t open again. That&#x27;s cool, I&#x27;d make it snap back to the pane where it&#x27;s open.<p>Kudos! This is an interesting perspective on how we really need to put a little more effort into the UX of LLMs.
评论 #40531713 未加载
评论 #40525089 未加载
评论 #40524209 未加载
wiremine12 months ago
Taking a step back: The UX&#x2F;UI for LLMs in general are very immature. We&#x27;re in the very early days of to best interact with these tools. We need more experimentation like this to help figure out what works, and doesn&#x27;t work.<p>Kudos!
评论 #40538226 未加载
评论 #40526982 未加载
评论 #40528307 未加载
评论 #40526844 未加载
spdustin12 months ago
If &quot;delve&quot; was meant to be an in-joke, I just wanted you to know: I got it.<p>I also have a Custom GPT &quot;AutoExpert (Chat)&quot; [0] that several reviewers have called &quot;the perfect Rabbit Hole GPT&quot; due to the way it leads users through learning a topic. You might dig it, especially since free tier users have access to these now.<p>[0]: <a href="https:&#x2F;&#x2F;chatgpt.com&#x2F;g&#x2F;g-LQHhJCXhW-autoexpert-chat" rel="nofollow">https:&#x2F;&#x2F;chatgpt.com&#x2F;g&#x2F;g-LQHhJCXhW-autoexpert-chat</a>
评论 #40523788 未加载
评论 #40536490 未加载
评论 #40526859 未加载
评论 #40526081 未加载
mariocesar12 months ago
Love it! I like that the site is straight to action, but I think it could really benefit from a walkthrough. Here’s an idea!<p>It would be great if we had an introduction to the site right in the prompt! to help understand its main purpose right from the start.<p>It&#x27;ll be great if the first thing you see is [Explain what &quot;delve into&quot; is] as a prompt suggestion. Next, it will reply with, &quot;It’s for exploring topics deeply, similar to going down &quot;rabbit holes&quot; where one interesting thing leads to another. Here are some examples ...&quot;<p>Then, you guide the user through the functions step-by-step. Something like, &quot;Click on option X to start a new thread, then choose from the suggested prompts or create your own. Follow the flow to see related threads and dive even deeper.&quot;
评论 #40524175 未加载
ivanhoe12 months ago
I really like this.<p>How about enabling user to select any piece of text and use it the way the links work now?<p>For instance, I&#x27;ve noticed it doesn&#x27;t linkify peoples&#x27; names, and one thing I love on Wikipedia is that you can easily lookup people mentioned in the text. So, rather than having thousand links in the page, it would be handy if I could just select the name in the text, click some button (or right-click menu item), and get the new prompt based on the selection (user having to click a few extra times for this custom use-case wouldn&#x27;t be a problem IMO)
评论 #40536562 未加载
roldie12 months ago
Great idea. I also see inspiration from Andy Matuschak&#x27;s notes [0], of which I&#x27;m a bit fan<p>[0] <a href="https:&#x2F;&#x2F;notes.andymatuschak.org&#x2F;zB74H9CuWrosEuqve7jZyCo?stackedNotes=z5E5QawiXCMbtNtupvxeoEX&amp;stackedNotes=zDcuS8A5uxGR8hQygsqP83A" rel="nofollow">https:&#x2F;&#x2F;notes.andymatuschak.org&#x2F;zB74H9CuWrosEuqve7jZyCo?stac...</a>
评论 #40524024 未加载
mbil12 months ago
This is great and something that I&#x27;ve wished existed. Thanks for making it! Right now, the tiles are linear. E.g. if tile A links to tiles B and C, clicking either B or C will open a new tile directly next to A (and only one of B or C is visible at a time). What do you think about making more of a tree layout where B and C both branch from A and can be viewed simultaneously?
评论 #40527341 未加载
评论 #40523776 未加载
评论 #40533220 未加载
globalise8312 months ago
I like it a lot. Feels like idle Wikipedia link surfing but with the key difference that each new step keeps track of the previous context. To me it is both novel and useful.
评论 #40527636 未加载
freedomben12 months ago
Wow, this is really neat! I usually don&#x27;t comment on Show HN&#x27;s because I&#x27;m rarely impressed by them and I don&#x27;t want my lack of enthusiasm to be a detterent for people showing their work, but occasionally one like this comes up that is very cool. I also really appreciate that absence of tracking other than Cloudflare Insights (which seems very reasonable to me).<p>There&#x27;s an old truism in the business, that the more &quot;suggestions&quot; people give about your idea, the more they like it, and it&#x27;s absolutely true. Solid work!<p>Do you have plans to monetize and&#x2F;or open source it?
ajdegol12 months ago
Branching conversations are great for a whole bunch of reasons. I posted a demo of a prototype: <a href="https:&#x2F;&#x2F;x.com&#x2F;ajdegol&#x2F;status&#x2F;1788689011302682657" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;ajdegol&#x2F;status&#x2F;1788689011302682657</a><p>And Jake Collins just announced he’s open sourcing an obsidian plugin which has a ton of features: <a href="https:&#x2F;&#x2F;x.com&#x2F;JacobColling&#x2F;status&#x2F;1795462258258002255" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;JacobColling&#x2F;status&#x2F;1795462258258002255</a>
评论 #40533241 未加载
评论 #40535137 未加载
albumen12 months ago
Perhaps have it not scroll down as it generates the text? Invariably I have to scroll back to the top to start reading. You could have a mini-hud (growing line, with a small rectangle at the top showing the first page of text) which would let you see at a glance how much text is being generated, without interrupting reading. Or not; ChatGPT just keeps on vibrating the phone (iOS app) during text gen, with hovering arrow in the middle-bottom as a shortcut to jump to the end.
评论 #40525187 未加载
评论 #40525513 未加载
skadamat12 months ago
Reminds me of Andy Matuschak&#x27;s UI as well: <a href="https:&#x2F;&#x2F;notes.andymatuschak.org&#x2F;z9C7piFz8mthkGUUd1W2CR?stackedNotes=zExJhXA1doKnxyQQ4ZjX4dV" rel="nofollow">https:&#x2F;&#x2F;notes.andymatuschak.org&#x2F;z9C7piFz8mthkGUUd1W2CR?stack...</a>
goosethe12 months ago
I&#x27;d love it if it had a &quot;zoomed out&quot; tree-like view that makes all the different paths of conversation viewable at once
评论 #40524939 未加载
评论 #40525010 未加载
GrinningFool12 months ago
This is pretty great, provides a nice set of breadcrumbs for a deep dive into any rabbit hole.<p>Once thing that threw me off was when I went to the original panel and clicked a second topic, it cleared out the panes that I had explored off the first topic. I had to discover they weren&#x27;t really lost by re-clicking. I think it would be better if there was some visual indicator they were still there - perhaps the topic (and sub-topics) get collapsed but are still visible with the heading of the selected topic?
qwertox12 months ago
For some reason I feel like the Wikipedia should use this (or something like this) as a backend to serve a more &quot;dialogish&quot; UI (without replacing the current static UI). The name &quot;delve&quot; is spot-on, but it lacks Wikipedia&#x27;s intelligence and interconnectedness.<p>And maybe add some locally stored &quot;Microsoft Recall&quot;-like feature to revisit paths you&#x27;ve made. It would be text-only, so use up almost no space, and be quickly searchable.<p>Well done it could even work in a terminal.
evantravers12 months ago
I can&#x27;t believe no one has mentioned Andy Matuschak&#x27;s work on his notes, &quot;Andy-mode:&quot; <a href="https:&#x2F;&#x2F;twitter.com&#x2F;andy_matuschak&#x2F;status&#x2F;1568032773025431552?lang=en" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;andy_matuschak&#x2F;status&#x2F;156803277302543155...</a><p>It&#x27;s a really clever use of the UI, and I think he&#x27;d be happy with it.
kepano12 months ago
It would be so cool to be able to save a delve exploration to a set of linked Markdown files that you can open in a local wiki app.
评论 #40530595 未加载
iknownthing12 months ago
Are you paying for the API calls yourself here?
评论 #40526144 未加载
评论 #40524413 未加载
brandonhorst12 months ago
I really love this. A book was recommended to me that I&#x27;m not going to have time to read, but this UI is an amazing way to figure out the main ideas and dive deeper into the interesting ones.<p>No idea if the things it&#x27;s telling me are true or not, but that doesn&#x27;t matter quite as much in this case.
评论 #40524719 未加载
wordpad2512 months ago
How does it work?<p>Are hyperlinks generated as part of original prompt or you do post processing on a response with another LLM?
rvnx12 months ago
You may find interesting to look at Google&#x27;s AI Kitchen and the early versions of Bard (the LaMDA version) because they were specifically optimizing user scenarios where the user wants bullet points (though intuitively, your tool seems already much better than what Google did).
distalx12 months ago
This UI wrapper looks fantastic! Great job on making the GPT API more accessible. Are there any plans to roll it out as a UI component that can integrate with existing systems, or does it require some backend work too? I see huge potential for clients like GPT4All and Prompta to use pluggable UI components like this, allowing people to keep their data locally.<p>I believe we might see AI clients evolving into open-source editors like Atom or VS Code, with plugins and packages that can be shared and iterated upon rapidly.<p>Also, I received the message &#x27;You&#x27;ve gone too deep!&#x27; and I&#x27;m wondering what the context length limit is. Thanks!
instagraham12 months ago
This could replace the hours I spend on Wikipedia. Hope it&#x27;s not too expensive to run.
评论 #40526544 未加载
评论 #40528135 未加载
blixt12 months ago
Really fun! I realized each delve carries the context of the previous ones. So I got to StarCraft II from the initial example of &quot;Faster language models&quot;, but it mostly talked about how SC2 can be used for reinforcement learning. It&#x27;d be nice to have a key I can hold down to start a new delve on the topic (bonus points if you can stack multiple delves so you can keep going deeper on the old track as well!)<p>Another thing that would be interesting is if there was minimal markup for the LLM to indicate &quot;here should be an image of [search term]&quot; or maybe even interactive code blocks etc. But obviously this is scope creep deluxe.
yungtriggz12 months ago
This is really cool! I love the rabbit hole stuff you can do when you give GPT more capabilities. I was playing around with this stuff and found I was most often wanting to use it when wanting to learn about something so made Instaclass: <a href="https:&#x2F;&#x2F;myinstaclass.com&#x2F;">https:&#x2F;&#x2F;myinstaclass.com&#x2F;</a>. It finds videos, images, makes quizzes and gets more relevant web links for you to keep exploring, and structures it like a class (basically a list of bullet points like you mentioned). Try it out and lmk what you think!
georgehotelling12 months ago
I&#x27;m not sure how to use this, I just see a blank screen with the word &quot;delve&quot; at the top and typing doesn&#x27;t do anything. I&#x27;m on Firefox on macOS.
nullbio12 months ago
Well this is incredible. Love the interface and the speed. My only wish is that it gave more links, right now it only seems to generate a few of them for each response.<p>Also it seems to be way faster than ChatGPT, yet just as intelligent, how is that possible? Can you elaborate a little bit on the architecture you&#x27;re using and how it all works? What model are you using? Is it just a straight up API to ChatGPT, or are you also using additional embeddings and fine-tuning?
theflyestpilot12 months ago
I&#x27;ve been craving a gpt ui where I could fork conversations that stem from a genesis thought.<p>Put this on a canvas too so I can zoom out and look at the footprint I left to retrace my steps
评论 #40529402 未加载
评论 #40535212 未加载
sergiotapia12 months ago
crrriispy ui dude. This is like that custom prompt people were sharing for gpt where it would preempt three relevant follow up questions.<p>what heuristic are you using for making words clickable?<p>i recommend making the links just hardcode old-school blue and purple. make it obvious you can click these things. &quot;dive on in&quot; <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=DElxVXS7PD0" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=DElxVXS7PD0</a>
amsterdorn12 months ago
Interesting concept, but in practice the experience is just a simplified Wikipedia hole. I enjoy all the hyperlinks in Wikipedia articles: they let me decide how deep I want to go rather than get a short summary with only a few pre-generated links.<p>This adds a lot more barriers to knowledge exploration (and the GPT is likely trained on Wikipedia anyways) and doesn&#x27;t provide sources.
yaj5412 months ago
awesome work! I&#x27;ve wanted to explore the same idea - glad to see it getting worked on. The chat interface into language models clearly works but it frequently feels like an inefficient way to explore the latent knowledge space of the model. Hypertext (the www) has also been shown to be a great way to explore a massive knowledge space. What this is doing is applying something like a hypertext layer as a way to navigate the model&#x27;s latent space. Very cool. It could become something of a dynamically generated personalized wikipedia. I&#x27;m curious what the prompts look like that you are using to generate subsequent &quot;pages&quot;. It could be as simple as &quot;write a wikipedia style summary of &lt;x&gt;&quot; but I think there is a lot of potential in including the context of my current &quot;rabbit hole&quot;: &quot;explain &lt;x&gt; in the context of &lt;y&gt; with a learning goal of &lt;z&gt;&quot;, etc. Another idea: grounding this kind of hypertext exploration with rag on a specific dataset, e.g., wikipedia or hn.
SubiculumCode12 months ago
This is pretty neat. What I really like is the tiling layout.<p>I subscribe to phind, which provides a nice search&#x2F;answer service, which also suggests followup questions, which works fairly well: <a href="https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;WfHSzdk" rel="nofollow">https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;WfHSzdk</a><p>But if it was in a tiling format, that would be pretty awesome for the flow, especially on mobile.
dinkleberg12 months ago
Nice work, this is really solid!<p>I&#x27;ve had something like this on my mind for a while. I really think there are some great use cases for AI around supporting&#x2F;enhancing human cognition rather than trying to outsource our thinking. In this case of this, being able to rapidly &quot;expand&quot; your working memory with whatever is present in these cards is promising.<p>I look forward to seeing what you do with this.
telesilla12 months ago
This is fantastic, like an encyclopedia that knows what context you are learning about as you skip pages. Nice work!<p>My minor recommendation is to highlight somehow that the input fields accepts any topic and the suggestions are just random try-it-out topics, it&#x27;s wasn&#x27;t immediately clear. Instead of &#x27;write a message&#x27; it could say maybe, &#x27;enter a topic to learn about&#x27;.
ProllyInfamous12 months ago
This reminds me of what <i>Perplexity . AI</i> &#x27;s interface was initially like, 2023Q1 (when the hyperlinks actually listed at the bottom of each reponse, instead of just as unlisted superscripts).<p>Nice clean interface, OC, <i>please keep it this way</i>. Thanks for sharing.
Hansenq12 months ago
I used to go down rabbit holes on Wikipedia all the time--could spend hours doing this.<p>This to me seems like an Infinite Wikipedia! Really cool use-case!
patrickhogan112 months ago
I love the left to right view. Super cool. I got stuck thought like 5 levels deep.<p>I really like this for source level information. Like drill down on research studies and then drill down to authors and concepts. Would be cool if it was also building a mind map or semantic tree that you could see. Like how to separate the topics by their level of generalization.
aaronharnly12 months ago
This is fun! It feels like infinite hyperlinks. It&#x27;s the kind of wonder I had in exploring Wikipedia for the first time.
viveknathani_12 months ago
This is so cool! I can see myself using this all the time for research work. Would love to manage multiple sessions. Would also like to see if the individual tiles could be resized. I would like to decrease the size of slightly less important tiles perhaps. Also, how are you managing the API costs right now?<p>Great work!
rnavi12 months ago
Follow up questions generated with the answer is easy picking. Much like Google (traditional search) does. To spice it up, there can be a slider using which - questions would change. The axis of slider itself can be an interesting one eg: simple to advanced. It will depend on the topic being discussed.
gsuuon12 months ago
This is cool, it&#x27;s almost like a wiki you can talk to. I also wanted to make a thread-based UI for LLM chat since I realized that&#x27;s how I typically interact with them (almost like git branches) but hadn&#x27;t gotten around to it yet. Neat to see others are interested in branching conversations as well!
hahattpro12 months ago
<a href="https:&#x2F;&#x2F;delve.a9.io&#x2F;" rel="nofollow">https:&#x2F;&#x2F;delve.a9.io&#x2F;</a> no longer give answer. Please give me code&#x2F;self-host frontend, or let me input my api key
dxhdr12 months ago
Very cool!<p>It seems to dead-end unexpectedly on some topics. For example, I delved &quot;path tracing&quot; -&gt; &quot;importance sampling&quot; and it output a section on probability distribution functions (PDFs) but didn&#x27;t offer any links to explore those further.<p>Highlight-to-delve would probably fix that.
maxutility12 months ago
It would be great to implement a browser extension that lets you highlight a term or phrase on any webpage and open a GPT rabbit hole for that term or phrase.<p>@maxkreiger - if I were to build one as a proof of concept, would you object to me having it hyperlink to your UI?
archon81012 months ago
Since the page did not provide any information on the language model used, I asked it and it said it was Anthropic. An About page would go a long way for this project.
andreygrehov12 months ago
I see there are many comments about Wikipedia. I can&#x27;t find the link, but many years ago (like 7+), there was a concept Wikipedia redesign that proposed the same UX. This was a &quot;marketing&quot; project of a non-Wikipedia related designer. Does anybody remember?
评论 #40529948 未加载
tomatosuoop12 months ago
This is great, I love it. I&#x27;d love to have it be keyboard only. Maybe using a command syntax like &quot;&#x2F;link title&quot;, so it knows we don&#x27;t want to continue the chat in the same panel, but want to follow the keywords. You&#x27;ve got something great here!
iamwil12 months ago
Reminds me of the Smalltalk IDE and browser.<p>Also, WorldSim could use something like this, and perhaps web browsers.<p>Anyway, I suspect this resonates for anyone that has to do research on the web or in GPT. I often end up with multiple threads on GPT anyway trying to learn about something.
jpcookie12 months ago
I love how it feels like obsidian
abhayhegde12 months ago
This reminds me of Andy Matuschak&#x27;s notes website [0]. I dig this view for much better understanding.<p>[0]: <a href="https:&#x2F;&#x2F;notes.andymatuschak.org&#x2F;" rel="nofollow">https:&#x2F;&#x2F;notes.andymatuschak.org&#x2F;</a>
cjf10112 months ago
I dig this. Reminds me of column mode in MacOS&#x27;s finder, which is similarly helpful in &quot;rewinding&quot; an exploration of a file system.<p>Would be interesting to rabbit other rabbit hole resources like Wikipedia or IMDB in this way too.
评论 #40526533 未加载
jabza12 months ago
Neat! I made a free browser extension with a somewhat similar behaviour (<a href="https:&#x2F;&#x2F;hoverflow.io" rel="nofollow">https:&#x2F;&#x2F;hoverflow.io</a>) Inspired by nested tooltips in Crusader Kings 3.
spacebacon12 months ago
For coding it would be nice to add task to links. So when a link is clicked you could simply choose to follow it or create a new agent with the link. Each agent will tune the output as one goes down the rabbit hole.
geepytee12 months ago
I like it! Would like to have some sort of map&#x2F;visualization of what topics i&#x27;ve explored for when diving into parallel sessions. Right now the UI only supports diving into one branch at the time.
lukew312 months ago
There should be something a UI like this but it just sources data from Wikipedia. Could just be the Abstract and allow you to open a wikipedia iframe or something. You should add this as a mode to this app
评论 #40529816 未加载
edngibson12 months ago
This is cool. I&#x27;d love an option to make the output elaborate more, if I&#x27;m interested in a topic it&#x27;s a bit disappointing that I can only read a few sentences. Other than that, nice work!
评论 #40530481 未加载
andrewmutz12 months ago
UI is very clean. Left right scroll is awkward without a trackpad, however
kaladin_112 months ago
First impression, a fast and neat interface. I went into data indexing rabbit hole as that has been my obsession these days.<p>Cudos! Particularly impressed with the lack of clutter and the speed.
davedx12 months ago
I REALLY like how snappy it is. I&#x27;ve always been impressed with how fast Wikipedia managed to stay over the years, but this is even better. Really nice work.
cglendenning12 months ago
This is what hyperlinking should have always been...context-sensitive. Very useful. You have earned a coveted slot on my bookmark list. Thank you!
horacemorace12 months ago
Super cool!<p>What’s the logic used to determine what words are links? I found that it was possible to feed it new vocabulary and that it would turn those into links
owenpalmer12 months ago
I really, really love this. Even after using it for just a few minutes, I&#x27;m sold. Excellent work, will it be open source at some point?
评论 #40527273 未加载
评论 #40529270 未加载
sunbum12 months ago
I just get a lightblue screen (Firefox Windows 11)
bardan12 months ago
Might be nice to display a couple of related articles from eg. Wikipedia or Wiktionary if the user wants to cross-reference.
ionwake12 months ago
This is absolutely brilliant. Like Tetris for the Gameboy. I have no suggestions how to monetise this but its brilliant.
peperunas12 months ago
This is an outstanding idea, I love it!<p>It would be great if you could support Ollama (or an OpenAI compatible endpoint) for private LLMs.
coder-312 months ago
I love this! My suggestion is to not close panels when going back up levels but instead put them in a parallel branch.
NayamAmarshe12 months ago
It seems that commonly used terms are cached, pretty neat strategy since this is wikipedia-like information website.
rexreed12 months ago
Which LLM is this using because it is providing hallucinated facts that differ from ChatGPT 4o on the latest model?
igorguerrero12 months ago
This UI is fantastic! Was the name &quot;delve&quot; AI generated? I just cannot believe a human uses that word...
评论 #40527812 未加载
评论 #40527805 未加载
toisanji12 months ago
some feedback:<p>* can you make it so we can share links of sessions?<p>* can you describe on the homepage or in a link from the homepage what it does.
评论 #40523427 未加载
nkotov12 months ago
I love the experience. It&#x27;s almost like a mind map that keeps getting better and better.
dereknee12 months ago
great work! We are also trying to design a new different way to use LLM: <a href="https:&#x2F;&#x2F;flowith.io" rel="nofollow">https:&#x2F;&#x2F;flowith.io</a> you can generate stuff on an infinite canvas.
jmkni12 months ago
Very nice idea, hope it&#x27;s not costing you a fortune to run!<p>Maybe let us put in our own api keys?
robertlagrant12 months ago
Warning: if you plug this into tvtropes then global productivity will drop sharply.
评论 #40524167 未加载
jackphilson12 months ago
How to combine the learning efficiency of this with the structure of a textbook?
obiefernandez12 months ago
Really curious about the data structure you&#x27;re using behind the scenes.
SoothingSorbet12 months ago
Can you please explain how you extract (or get it to generate) the links?
aster0id12 months ago
Love this. It would really benefit from a back&#x2F;forward button though!
atentaten12 months ago
This made me chuckle with delight once I understood what it was doing.
hahattpro12 months ago
What model&#x2F;api are you using for this ?
dilliwal12 months ago
Pretty cool, love it, why not opening a new tab instead?
spazmoth12 months ago
It&#x27;s important to remember that Wikipedia is
DrNosferatu12 months ago
Dark mode please!
ApolIllo12 months ago
i wish perplexity.ai did this
kkukshtel12 months ago
rip your api credits
byyoung312 months ago
this UI feels right.
mxpxrocks1012 months ago
nice job!
jpcookie12 months ago
Source code?