The best Hacker News stories from Show from the past day
Latest posts:
Show HN: Wordllama – Things you can do with the token embeddings of an LLM
After working with LLMs for long enough, I found myself wanting a lightweight utility for doing various small tasks to prepare inputs, locate information and create evaluators. This library is two things: a very simple model and utilities that inference it (eg. fuzzy deduplication). The target platform is CPU, and it’s intended to be light, fast and pip installable — a library that lowers the barrier to working with strings <i>semantically</i>. You don’t need to install pytorch to use it, or any deep learning runtimes.<p>How can this be accomplished? The model is simply token embeddings that are average pooled. To create this model, I extracted token embedding (nn.Embedding) vectors from LLMs, concatenated them along the embedding dimension, added a learnable weight parameter, and projected them to a smaller dimension. Using the sentence transformers framework and datasets, I trained the pooled embedding with multiple negatives ranking loss and matryoshka representation learning so they can be truncated. After training, the weights and projections are no longer needed, because there is no contextual calculations. I inference the entire token vocabulary and save the new token embeddings to be loaded to numpy.<p>While the results are not impressive compared to transformer models, they perform well on MTEB benchmarks compared to word embedding models (which they are most similar to), while being much smaller in size (smallest model, 32k vocab, 64-dim is only 4MB).<p>On the utility side, I’ve been adding some tools that I think it’ll be useful for. In addition to general embedding, there’s algorithms for ranking, filtering, clustering, deduplicating and similarity. Some of them have a cython implementation, and I’m continuing to work on benchmarking them and improving them as I have time. In addition to “standard” models that use cosine similarity for some algorithms, there are binarized models that use hamming distance. This is a slightly faster, similarity algorithm, with significantly less memory per embedding (float32 -> 1 bit).<p>Hope you enjoy it, and find it useful. PS I haven’t figured out Windows builds yet, but Linux and Mac are supported.
Show HN: Bullshit Remover
Show HN: The Hitchhiker's Guide to the Galaxy Online
I thought to make a funny website with a fine-tuned model that acts like the guide from Douglas Adams's novel. This way, you can find the right answer when in a tight spot (like when you need a ride to Alpha Centauri or smth).
Show HN: The Hitchhiker's Guide to the Galaxy Online
I thought to make a funny website with a fine-tuned model that acts like the guide from Douglas Adams's novel. This way, you can find the right answer when in a tight spot (like when you need a ride to Alpha Centauri or smth).
Show HN: The Hitchhiker's Guide to the Galaxy Online
I thought to make a funny website with a fine-tuned model that acts like the guide from Douglas Adams's novel. This way, you can find the right answer when in a tight spot (like when you need a ride to Alpha Centauri or smth).
Show HN: I removed politics from Twitter with AI
I'm sure you all saw the debate a few days ago, alongside all the tech influencoors giving their mid-takes about politics.<p>Doesn't it get tiring?<p>That's why I built mindfirewall a chrome extension that uses AI to filter out tweets.<p>You'll never have to see politics, war, or negativity on your timeline again.<p>In the future you'll be shocked you even allowed this stuff into your brain in the first place.<p>Let me know what you think below, does it feel too much like censorship? would you use this?
Show HN: Meet.hn – Meet the Hacker News community in your city
Hey HN!<p>I just published <a href="https://meet.hn" rel="nofollow">https://meet.hn</a>, a map to find hackers in your city.<p>How it works?<p>Demo of the signup process: <a href="https://x.com/meet_hn/status/1834918518904746329" rel="nofollow">https://x.com/meet_hn/status/1834918518904746329</a><p>1. Fill the form: username, city+country<p>2. Copy the text generated in the box below the form, and paste it in your HN description.<p>3. Click "Add me on the map"<p>Optionnaly (it's recommended!) you can add links to your socials as well as some tags to showcase your interests.<p>Why does it exist?<p>I created this because, despite its harsh reputation on the internet, I love the HN community.
I have fewer than a handful of friends who are as curious and eager to think and reflect as the people on HN.
Also, the city I currently live in is more focused on industry than on technology and entrepreneurship, which are core to HN.<p>This led me to want to meet the HN community IRL.
After trying `site:news.ycombinator.com/user toulouse` on Google and getting only one result, I decided to create meet.hn.<p>My first goal with this is to meet at least one HN member in my city: Toulouse, France.
If you are ever in the area, hit me up! I'm sirobg at <a href="https://meet.hn/city/fr-Toulouse" rel="nofollow">https://meet.hn/city/fr-Toulouse</a><p>Additional details:<p>- meet.hn has a twitter page: <a href="https://x.com/meet_hn" rel="nofollow">https://x.com/meet_hn</a>. If you meet IRL thanks to meet.hn, don't hesitate to tag it with a picture, it would mean the world to me.<p>- the code is open source: <a href="https://github.com/borisghidaglia/meet-hn">https://github.com/borisghidaglia/meet-hn</a><p>- meet.hn integrates with <a href="https://at.hn/" rel="nofollow">https://at.hn/</a> from @padolsey (<a href="https://padolsey.at.hn/" rel="nofollow">https://padolsey.at.hn/</a>), registered on meet.hn at <a href="https://meet.hn/city/cn-Beijing" rel="nofollow">https://meet.hn/city/cn-Beijing</a><p>Finally, many thanks to these people for their help and/or feedbacks!
Ordered alphabetically:<p>- <a href="https://x.com/ericbureltech" rel="nofollow">https://x.com/ericbureltech</a><p>- <a href="https://x.com/fredkisss" rel="nofollow">https://x.com/fredkisss</a><p>- <a href="https://x.com/JulienDuquesne1" rel="nofollow">https://x.com/JulienDuquesne1</a><p>- <a href="https://www.linkedin.com/in/lbasseto/" rel="nofollow">https://www.linkedin.com/in/lbasseto/</a><p>- <a href="https://x.com/lcswillems" rel="nofollow">https://x.com/lcswillems</a><p>- <a href="https://x.com/leeerob" rel="nofollow">https://x.com/leeerob</a><p>- <a href="https://x.com/padolsey" rel="nofollow">https://x.com/padolsey</a><p>- <a href="https://x.com/tomlienard" rel="nofollow">https://x.com/tomlienard</a><p>I hope you will enjoy this!
Please share any feedback in the comments.
Show HN: Meet.hn – Meet the Hacker News community in your city
Hey HN!<p>I just published <a href="https://meet.hn" rel="nofollow">https://meet.hn</a>, a map to find hackers in your city.<p>How it works?<p>Demo of the signup process: <a href="https://x.com/meet_hn/status/1834918518904746329" rel="nofollow">https://x.com/meet_hn/status/1834918518904746329</a><p>1. Fill the form: username, city+country<p>2. Copy the text generated in the box below the form, and paste it in your HN description.<p>3. Click "Add me on the map"<p>Optionnaly (it's recommended!) you can add links to your socials as well as some tags to showcase your interests.<p>Why does it exist?<p>I created this because, despite its harsh reputation on the internet, I love the HN community.
I have fewer than a handful of friends who are as curious and eager to think and reflect as the people on HN.
Also, the city I currently live in is more focused on industry than on technology and entrepreneurship, which are core to HN.<p>This led me to want to meet the HN community IRL.
After trying `site:news.ycombinator.com/user toulouse` on Google and getting only one result, I decided to create meet.hn.<p>My first goal with this is to meet at least one HN member in my city: Toulouse, France.
If you are ever in the area, hit me up! I'm sirobg at <a href="https://meet.hn/city/fr-Toulouse" rel="nofollow">https://meet.hn/city/fr-Toulouse</a><p>Additional details:<p>- meet.hn has a twitter page: <a href="https://x.com/meet_hn" rel="nofollow">https://x.com/meet_hn</a>. If you meet IRL thanks to meet.hn, don't hesitate to tag it with a picture, it would mean the world to me.<p>- the code is open source: <a href="https://github.com/borisghidaglia/meet-hn">https://github.com/borisghidaglia/meet-hn</a><p>- meet.hn integrates with <a href="https://at.hn/" rel="nofollow">https://at.hn/</a> from @padolsey (<a href="https://padolsey.at.hn/" rel="nofollow">https://padolsey.at.hn/</a>), registered on meet.hn at <a href="https://meet.hn/city/cn-Beijing" rel="nofollow">https://meet.hn/city/cn-Beijing</a><p>Finally, many thanks to these people for their help and/or feedbacks!
Ordered alphabetically:<p>- <a href="https://x.com/ericbureltech" rel="nofollow">https://x.com/ericbureltech</a><p>- <a href="https://x.com/fredkisss" rel="nofollow">https://x.com/fredkisss</a><p>- <a href="https://x.com/JulienDuquesne1" rel="nofollow">https://x.com/JulienDuquesne1</a><p>- <a href="https://www.linkedin.com/in/lbasseto/" rel="nofollow">https://www.linkedin.com/in/lbasseto/</a><p>- <a href="https://x.com/lcswillems" rel="nofollow">https://x.com/lcswillems</a><p>- <a href="https://x.com/leeerob" rel="nofollow">https://x.com/leeerob</a><p>- <a href="https://x.com/padolsey" rel="nofollow">https://x.com/padolsey</a><p>- <a href="https://x.com/tomlienard" rel="nofollow">https://x.com/tomlienard</a><p>I hope you will enjoy this!
Please share any feedback in the comments.
Show HN: Meet.hn – Meet the Hacker News community in your city
Hey HN!<p>I just published <a href="https://meet.hn" rel="nofollow">https://meet.hn</a>, a map to find hackers in your city.<p>How it works?<p>Demo of the signup process: <a href="https://x.com/meet_hn/status/1834918518904746329" rel="nofollow">https://x.com/meet_hn/status/1834918518904746329</a><p>1. Fill the form: username, city+country<p>2. Copy the text generated in the box below the form, and paste it in your HN description.<p>3. Click "Add me on the map"<p>Optionnaly (it's recommended!) you can add links to your socials as well as some tags to showcase your interests.<p>Why does it exist?<p>I created this because, despite its harsh reputation on the internet, I love the HN community.
I have fewer than a handful of friends who are as curious and eager to think and reflect as the people on HN.
Also, the city I currently live in is more focused on industry than on technology and entrepreneurship, which are core to HN.<p>This led me to want to meet the HN community IRL.
After trying `site:news.ycombinator.com/user toulouse` on Google and getting only one result, I decided to create meet.hn.<p>My first goal with this is to meet at least one HN member in my city: Toulouse, France.
If you are ever in the area, hit me up! I'm sirobg at <a href="https://meet.hn/city/fr-Toulouse" rel="nofollow">https://meet.hn/city/fr-Toulouse</a><p>Additional details:<p>- meet.hn has a twitter page: <a href="https://x.com/meet_hn" rel="nofollow">https://x.com/meet_hn</a>. If you meet IRL thanks to meet.hn, don't hesitate to tag it with a picture, it would mean the world to me.<p>- the code is open source: <a href="https://github.com/borisghidaglia/meet-hn">https://github.com/borisghidaglia/meet-hn</a><p>- meet.hn integrates with <a href="https://at.hn/" rel="nofollow">https://at.hn/</a> from @padolsey (<a href="https://padolsey.at.hn/" rel="nofollow">https://padolsey.at.hn/</a>), registered on meet.hn at <a href="https://meet.hn/city/cn-Beijing" rel="nofollow">https://meet.hn/city/cn-Beijing</a><p>Finally, many thanks to these people for their help and/or feedbacks!
Ordered alphabetically:<p>- <a href="https://x.com/ericbureltech" rel="nofollow">https://x.com/ericbureltech</a><p>- <a href="https://x.com/fredkisss" rel="nofollow">https://x.com/fredkisss</a><p>- <a href="https://x.com/JulienDuquesne1" rel="nofollow">https://x.com/JulienDuquesne1</a><p>- <a href="https://www.linkedin.com/in/lbasseto/" rel="nofollow">https://www.linkedin.com/in/lbasseto/</a><p>- <a href="https://x.com/lcswillems" rel="nofollow">https://x.com/lcswillems</a><p>- <a href="https://x.com/leeerob" rel="nofollow">https://x.com/leeerob</a><p>- <a href="https://x.com/padolsey" rel="nofollow">https://x.com/padolsey</a><p>- <a href="https://x.com/tomlienard" rel="nofollow">https://x.com/tomlienard</a><p>I hope you will enjoy this!
Please share any feedback in the comments.
Show HN: HypergraphZ – A Hypergraph Implementation in Zig
Show HN: I built TikTok but for studying with quizzes from your own notes
Show HN: I built TikTok but for studying with quizzes from your own notes
Show HN: Galaxy Visualization
100k real ( +100k random ) galaxies from a sector. Visualized with Raylib.
Show HN: Galaxy Visualization
100k real ( +100k random ) galaxies from a sector. Visualized with Raylib.
Show HN: A whiteboard that writes math equations
Show HN: A whiteboard that writes math equations
Show HN: FlowTracker – Track data flowing through Java programs
FlowTracker, a Java agent that tracks data flowing through Java programs.
It helps you understand where any program got its output from, what it means, and why it wrote it.<p>Watch the video or explore the live demo yourself, and read how it works at <a href="https://github.com/coekie/flowtracker">https://github.com/coekie/flowtracker</a>
Show HN: FlowTracker – Track data flowing through Java programs
FlowTracker, a Java agent that tracks data flowing through Java programs.
It helps you understand where any program got its output from, what it means, and why it wrote it.<p>Watch the video or explore the live demo yourself, and read how it works at <a href="https://github.com/coekie/flowtracker">https://github.com/coekie/flowtracker</a>
Show HN: FlowTracker – Track data flowing through Java programs
FlowTracker, a Java agent that tracks data flowing through Java programs.
It helps you understand where any program got its output from, what it means, and why it wrote it.<p>Watch the video or explore the live demo yourself, and read how it works at <a href="https://github.com/coekie/flowtracker">https://github.com/coekie/flowtracker</a>
Show HN: Repogather – copy relevant files to clipboard for LLM coding workflows
Hey HN, I wanted to share a simple command line tool I made that has sped up and simplified my LLM assisted coding workflow. Whenever possible, I’ve been trying to use Claude as a first pass when implementing new features / changes. But I found that depending on the type of change I was making, I was spending a lot of thought finding and deciding which source files should be included in the prompt. The need to copy/paste each file individually also becomes a mild annoyance.<p>First, I implemented `repogather --all` , which unintelligently copies <i>all</i> sources files in your repository to the clipboard (delimited by their relative filepaths). To my surprise, for less complex repositories, this alone is often completely workable for Claude — much better than pasting in the just the few files you are looking to update. But I never would have done it if I had to copy/paste everything individually. 200k is quite a lot of tokens!<p>But as soon as the repository grows to a certain complexity level (even if it is under the input token limit), I’ve found that Claude can get confused by different unrelated parts / concepts across the code. It performs much better if you make an attempt to exclude logic that is irrelevant to your current change. So I implemented `repogather "<query here>"` , e.g. `repogather "only files related to authentication"` . This uses gpt-4o-mini with structured outputs to provide a relevance score for each source file (with automatic exclusions for .gitignore patterns, tests, configuration, and other manual exclusions with `--exclude <pattern>` ).<p>gpt-4o-mini is so cheap and fast, that for my ~8 dev startup’s repo, it takes under 5 seconds and costs 3-4 cents (with appropriate exclusions). Plus, you get to watch the output stream while you wait which always feels fun.<p>The retrieval isn’t always perfect the first time — but it is fast, which allows you to see what files it returned, and iterate quickly on your command. I’ve found this to be much more satisfying than embedding-search based solutions I’ve used, which seem to fail in pretty opaque ways.<p><a href="https://github.com/gr-b/repogather">https://github.com/gr-b/repogather</a><p>Let me know if it is useful to you! Always love to talk about how to better integrate LLMs into coding workflows.