frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Show HN: I saw this cool navigation reveal, so I made a simple HTML+CSS version

https://github.com/Momciloo/fun-with-clip-path
10•momciloo•2h ago•0 comments

Show HN: Look Ma, No Linux: Shell, App Installer, Vi, Cc on ESP32-S3 / BreezyBox

https://github.com/valdanylchuk/breezydemo
275•isitcontent•22h ago•38 comments

Show HN: Kappal – CLI to Run Docker Compose YML on Kubernetes for Local Dev

https://github.com/sandys/kappal
22•sandGorgon•2d ago•12 comments

Show HN: If you lose your memory, how to regain access to your computer?

https://eljojo.github.io/rememory/
351•eljojo•1d ago•216 comments

Show HN: I built a <400ms latency voice agent that runs on a 4gb vram GTX 1650"

https://github.com/pheonix-delta/axiom-voice-agent
2•shubham-coder•1h ago•0 comments

Show HN: I spent 4 years building a UI design tool with only the features I use

https://vecti.com
367•vecti•1d ago•169 comments

Show HN: Stacky – certain block game clone

https://www.susmel.com/stacky/
3•Keyframe•2h ago•0 comments

Show HN: A toy compiler I built in high school (runs in browser)

https://vire-lang.web.app
3•xeouz•2h ago•1 comments

Show HN: Smooth CLI – Token-efficient browser for AI agents

https://docs.smooth.sh/cli/overview
94•antves•2d ago•70 comments

Show HN: R3forth, a ColorForth-inspired language with a tiny VM

https://github.com/phreda4/r3
82•phreda4•21h ago•16 comments

Show HN: Nginx-defender – realtime abuse blocking for Nginx

https://github.com/Anipaleja/nginx-defender
3•anipaleja•4h ago•0 comments

Show HN: BioTradingArena – Benchmark for LLMs to predict biotech stock movements

https://www.biotradingarena.com/hn
28•dchu17•1d ago•12 comments

Show HN: Artifact Keeper – Open-Source Artifactory/Nexus Alternative in Rust

https://github.com/artifact-keeper
154•bsgeraci•1d ago•64 comments

Show HN: Slack CLI for Agents

https://github.com/stablyai/agent-slack
52•nwparker•1d ago•12 comments

Show HN: ARM64 Android Dev Kit

https://github.com/denuoweb/ARM64-ADK
18•denuoweb•2d ago•2 comments

Show HN: MCP App to play backgammon with your LLM

https://github.com/sam-mfb/backgammon-mcp
3•sam256•6h ago•1 comments

Show HN: Gigacode – Use OpenCode's UI with Claude Code/Codex/Amp

https://github.com/rivet-dev/sandbox-agent/tree/main/gigacode
21•NathanFlurry•1d ago•10 comments

Show HN: I'm 75, building an OSS Virtual Protest Protocol for digital activism

https://github.com/voice-of-japan/Virtual-Protest-Protocol/blob/main/README.md
7•sakanakana00•7h ago•1 comments

Show HN: I built Divvy to split restaurant bills from a photo

https://divvyai.app/
3•pieterdy•7h ago•1 comments

Show HN: Which chef knife steels are good? Data from 540 Reddit tread

https://new.knife.day/blog/reddit-steel-sentiment-analysis
2•p-s-v•2h ago•0 comments

Show HN: Micropolis/SimCity Clone in Emacs Lisp

https://github.com/vkazanov/elcity
173•vkazanov•2d ago•49 comments

Show HN: I Hacked My Family's Meal Planning with an App

https://mealjar.app
2•melvinzammit•9h ago•0 comments

Show HN: Slop News – HN front page now, but it's all slop

https://dosaygo-studio.github.io/hn-front-page-2035/slop-news
21•keepamovin•12h ago•6 comments

Show HN: I built a free UCP checker – see if AI agents can find your store

https://ucphub.ai/ucp-store-check/
2•vladeta•10h ago•2 comments

Show HN: Daily-updated database of malicious browser extensions

https://github.com/toborrm9/malicious_extension_sentry
14•toborrm9•1d ago•8 comments

Show HN: XAPIs.dev – Twitter API Alternative at 90% Lower Cost

https://xapis.dev
3•nmfccodes•4h ago•1 comments

Show HN: Falcon's Eye (isometric NetHack) running in the browser via WebAssembly

https://rahuljaguste.github.io/Nethack_Falcons_Eye/
7•rahuljaguste•21h ago•1 comments

Show HN: Horizons – OSS agent execution engine

https://github.com/synth-laboratories/Horizons
24•JoshPurtell•1d ago•5 comments

Show HN: Compile-Time Vibe Coding

https://github.com/Michael-JB/vibecode
10•michaelchicory•11h ago•3 comments

Show HN: Local task classifier and dispatcher on RTX 3080

https://github.com/resilientworkflowsentinel/resilient-workflow-sentinel
25•Shubham_Amb•1d ago•2 comments
Open in hackernews

Show HN: Tiny Diffusion – A character-level text diffusion model from scratch

https://github.com/nathan-barry/tiny-diffusion
172•nathan-barry•2mo ago
This is a character-level language diffusion model for text generation.

The model is a modified version of Nanochat's GPT implementation and is trained on Tiny Shakespeare!

It is only 10.7 million parameters, so you can try it out locally.

Comments

yugretcx•2mo ago
Why do these text diffusion demos always look like the number of allowed tokens is fixed for a specific unfilled region?

Is this the case?

Ie. if the region only has four tokens(here characters) but calculates the best word is “forget” does it just abandon the best fit or truncate it to fit?

Are there text diffusion models with lax infill directives?

rand0mwalk•2mo ago
Tokens start as a special [MASK] token. Then as the diffusion process runs they are "unmasked" i.e. sampled.

So yes, you define a sequence of [MASK] tokens with some length ahead of time.

In practice, if a model wants to write a shorter sequence, it'll just fill the remaining tokens with empty content. If it wants to write a longer sequence, you'll have to identify this and extend the sequence with more [MASK] tokens. This is typically obvious since there's no "end of sequence" token present if the model wants to generate more.

nathan-barry•2mo ago
Yes, this is the case. During training, the model will get a sequence of text (ex, 512 tokens long) with a percentage of them masked out (with a special <MASK> token). It learns how to unmask those tokens to construct the original text.

In the case that you mentioned, if we had 4 <MASK> tokens in a row, all we are doing for decoding is predicting what those 4 tokens should be.

Generally, this does not seem to be a significant problem, as there are usually multiple ways to express an idea in varying lengths. Also, with confidence-aware parallel decoding, it can usually avoid the scenario you mentioned, as focusing on decoding the highest confident tokens will generally avoid such scenarios with a well trained model.

simonw•2mo ago
This is really neat.

I noticed the diffusion-process.py demo was using matplotlib in a window, but I figured it would be cute if it used a terminal UI instead - so I had Claude Code convert it to use curses. Code and demo GIF here: https://gist.github.com/simonw/9033ebd8dd17b4c0ad101ddda7a54...

Majromax•2mo ago
The basic MLP block in this model uses a ReLU^2 activation function (x <- ReLU(x)^2). That seems to be copied from the nanochat project, and it's not present in nanoGPT. Is there some documentation on the choice of this activation function?
throwaway2027•2mo ago
Isn't it because ReLU is cheap and ^2 is squared loss?
kouteiheika•2mo ago
When it comes to compute cost the choice of activation function makes little difference nowadays (and it can often be fused with whatever operation comes before it, which makes it effectively free).

The real reason is simple: it was inherited.

The relu^2 was used in the nanogpt speedrun[1] because it produced the best empirical results, then Andrej based his nanochat on the nanogpt speedrun without changing the activation function, and then this project was based on nanochat.

[1] -- https://github.com/KellerJordan/modded-nanogpt

macleginn•2mo ago
There has been some experimentation with the use of ReLU^2 in language models in recent years, e.g., here: https://proceedings.neurips.cc/paper_files/paper/2021/file/2...
mlmonkey•2mo ago
I'm curious: has there been any work done on generating embedding vectors instead of discrete tokens via diffusion? What would that look like? Please point me to some references. Thanks!
volodia•2mo ago
There is also this one that was released in October: https://github.com/kuleshov/char-mdlm
embedding-shape•2mo ago
Fun project, easy to understand and nice looking results, everything one could ask for! I played around with it locally, did some optimizations of low hanging fruits without making it much more complicated, and was gonna send over a PR. But then I noticed there is no license attached to the project. What are your plans regarding the licensing for this?
nathan-barry•2mo ago
Hey, I’ll add the MIT licenses later today!
tell_me_whai•2mo ago
Looks fun, thanks for sharing. I see you're implementing game of life sampling, what's the reasoning for using this logic?
gdiamos•2mo ago
One year later and there is still no inference engine for diffusion LLMs

Students looking for a project to break into AI - please!

nathan-barry•2mo ago
Actually NVIDIA made one earlier this year, check out their Fast-dLLM paper
gdiamos•2mo ago
Thanks I’ll check it out!
gdiamos•2mo ago
Did I miss something? https://github.com/NVlabs/Fast-dLLM/blob/main/llada/chat.py

That’s inference code, but where is the high perf web server?

tough•2mo ago
training inspired on nanochat for diffusion models: https://github.com/ZHZisZZ/dllm

now someone needs to make it work with vllm or something

doppelgunner•2mo ago
This is impressive. Can it run on mobile?
aa_y_ush•2mo ago
this is so so awesome.