What an unprocessed photo looks like

Kidnapped by Deutsche Bahn

You can make up HTML tags

How small can a language model be while still doing something useful? I wanted to find out, and had some spare time over the holidays.Z80-μLM is a character-level language model with 2-bit quantized weights ({-2,-1,0,+1}) that runs on a Z80 with 64KB RAM. The entire thing: inference, weights, chat UI, it all fits in a 40KB .COM file that you can run in a CP&#x2F;M emulator and hopefully even real hardware!It won&#x27;t write your emails, but it can be trained to play a stripped down version of 20 Questions, and is sometimes able to maintain the illusion of having simple but terse conversations with a distinct personality.--The extreme constraints nerd-sniped me and forced interesting trade-offs: trigram hashing (typo-tolerant, loses word order), 16-bit integer math, and some careful massaging of the training data meant I could keep the examples &#x27;interesting&#x27;.The key was quantization-aware training that accurately models the inference code limitations. The training loop runs both float and integer-quantized forward passes in parallel, scoring the model on how well its knowledge survives quantization. The weights are progressively pushed toward the 2-bit grid using straight-through estimators, with overflow penalties matching the Z80&#x27;s 16-bit accumulator limits. By the end of training, the model has already adapted to its constraints, so no post-hoc quantization collapse.Eventually I ended up spending a few dollars on Claude API to generate 20 questions data (see examples&#x2F;guess&#x2F;GUESS.COM), I hope Anthropic won&#x27;t send me a C&amp;D for distilling their model against the ToS ;PBut anyway, happy code-golf season everybody :)

Show HN: Z80-μLM, a 'Conversational AI' That Fits in 40KB

As AI gobbles up chips, prices for devices may rise

Software engineers should be a little bit cynical

Unity's Mono problem: Why your C# code runs slower than it should

MongoBleed Explained Simply

PySDR: A Guide to SDR and DSP Using Python

Staying ahead of censors in 2025

Show HN: Vibe coding a bookshelf with Claude Code

Researchers discover molecular difference in autistic brains

Huge Binaries

I&#x27;ve been working on a little open source [1] search engine, nilch. I noticed that nearly all well known search engines, including the alternative ones, tend to be run by companies of various sizes with the goal to make money, so they either fill your results with ads or charge you money, and I dislike this because search is the backbone of the internet and should not be commercial, so it runs in a not-for-profit style and aims to survive on donations. Additionally I&#x27;m personally really sick of AI in my search results so I got rid of that, and I wanted DuckDuckGo bangs so it supports all of them. Like many alternative search engines, it is fully private.Sadly, it currently does not have its own index but rather uses the Brave search API. Once I&#x27;m in a financial position that it&#x27;s possible, I would absolutely love to build a completely new index from the ground up which is open source, as well as an open source ranking and search algorithm, to back it.I posted on Reddit and got an amazing amount of feedback which I implemented a number of feature requests, so I would really like your ideas, critiques, and bug reports as well. Thank you and sorry for the long post![1] <a href="https:&#x2F;&#x2F;github.com&#x2F;UnmappedStack&#x2F;nilch" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;UnmappedStack&#x2F;nilch</a>

HN

What an unprocessed photo looks like

Kidnapped by Deutsche Bahn

You can make up HTML tags

Show HN: Z80-μLM, a 'Conversational AI' That Fits in 40KB

As AI gobbles up chips, prices for devices may rise

Software engineers should be a little bit cynical

Unity's Mono problem: Why your C# code runs slower than it should

MongoBleed Explained Simply

PySDR: A Guide to SDR and DSP Using Python

Staying ahead of censors in 2025

Show HN: Vibe coding a bookshelf with Claude Code

Researchers discover molecular difference in autistic brains

Huge Binaries

Show HN: My not-for-profit search engine with no ads, no AI, & all DDG bangs

My First Meshtastic Network

UK accounting body to halt remote exams amid AI cheating

Spherical Cow

Developing a Beautiful and Performant Block Editor in Qt C++ and QML

Feynman's Hughes Lectures: 950 pages of notes

GOG is getting acquired by its original co-founder: What it means for you