frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

The recently proposed o(n log n) integer multiplication algorithm [pdf]

https://web.maths.unsw.edu.au/~davidharvey/research/subnlogn.pdf
123•craigcitro•3h ago•73 comments

WorldTUI: I trained a 360M model to pretend to be a terminal

https://simedw.com/2026/10/21/world-model-terminal/
60•alphax314•3h ago•21 comments

There isn't an easy exploit for this PHP vulnerability

https://eliastaylor.com/noexploit/
391•zkldi•7h ago•183 comments

Security Disclosure

https://blog.railway.com/p/security-disclosure-wednesday-october-21st
401•passenger•8h ago•233 comments

GitLab critical patch release

https://docs.gitlab.com/releases/patches/patch-release-gitlab-19-5-1-released/
301•artooro•11h ago•73 comments

Rust Style Guide

https://google.github.io/styleguide/rustguide.html
268•kaycebasques•11h ago•149 comments

Compromised LiteLLM 1.107.0 on PyPI Installs a Credential Stealer and Backdoor

https://safedep.io/litellm-pypi-compromise-credential-stealer-backdoor/
477•abhisek•16h ago•162 comments

Using LLM for confidence scoring is a bad idea

https://rumple.github.io/blog/2026/10/20/asking-llm-for-confidence.html
48•jimauthors•1d ago•30 comments

I can't believe PowerShell is good

https://www.spacebar.news/powershell-is-actually-good/
46•corbindavenport•2h ago•39 comments

OpenAI CLI

https://github.com/openai/openai-cli
12•armcat•55m ago•3 comments

mitmproxy adds --emudhra option

https://github.com/mitmproxy/mitmproxy/commit/169234499088fc2c414644a67efaed3d64ae2e3b
382•ahlCVA•6h ago•95 comments

No more icons in every menu by default

https://blog.jim-nielsen.com/2026/icons-in-every-menu/
246•riffraff•2d ago•163 comments

Chain of Thought for Humans

https://www.boristhebrave.com/2026/10/18/chain-of-thought-for-humans/
76•ibobev•2d ago•33 comments

Beyond the &

https://lwn.net/Articles/1096028/
41•PaulHoule•1d ago•2 comments

Meta ordered to pay $25B penalty in New Mexico Cambridge Analytica case

https://www.cnbc.com/2026/10/21/meta-new-mexico-cambridge-analytica-penalty.html
484•cbg0•12h ago•301 comments

PingCap TiDB Leaks

https://fulcrumsec.vg/pingcap/
288•swyx•1d ago•43 comments

Tiny possum and glider thought extinct for 6000 years found in remote West Papua

https://www.bbc.com/news/articles/cp814l80pw2o
185•bryanrasmussen•1d ago•37 comments

Show HN: Omnara – Open-source alternative to Claude Managed Agents

https://github.com/omnara-ai/omnara
59•kmansm27•13h ago•16 comments

What Is the Karakoram Anomaly?

https://nsidc.org/learn/ask-scientist/karakoram-anomaly-it-real
12•thunderbong•2d ago•3 comments

OpenAI models are coming to Google Cloud

https://openai.com/index/openai-models-google-cloud/
276•armcat•17h ago•113 comments

Today's PHP security release, and how we're strengthening security

https://thephp.foundation/blog/2026/10/21/todays-php-security-release-and-how-were-strengthening-...
709•ejohnso49•14h ago•348 comments

DuckDB 2.0.0 LTS “Cinnamon Teal”

https://duckdb.org/2026/10/21/announcing-duckdb-200
395•nnx•16h ago•50 comments

Code is clay now

https://www.albertodenia.com/writing/code-is-clay-now
149•albertodenia•2d ago•82 comments

Patch the Planet: global response coordination as agents uncover vulnerabilities

https://openai.com/index/coordinated-disclosure-of-critical-vulnerabilities/
249•nr378•16h ago•167 comments

Oregon lawmakers approve money to buy an iconic waterfall listed on Redfin

https://apnews.com/article/oregon-waterfall-for-sale-lawmakers-996de63462d49ec3ad9bee1629fb9488
183•surprisetalk•14h ago•91 comments

Clockwork Mechanics, Maintenance, and Metaphors in the Late Middle Ages

https://www.jhiblog.org/2026/10/19/clockwork-metaphors/
7•benbreen•2d ago•1 comments

Postgres recovery_prefetch can cause outages

https://ardentperf.com/2026/10/20/postgres-recovery_prefetch-considered-harmful-and-causing-outages/
15•plaur782•7h ago•0 comments

Why Windows 11 still ships with Windows 95's memory manager hidden inside it

https://www.windowslatest.com/2026/10/19/why-windows-11-still-ships-with-windows-95s-memory-manag...
167•khalidkhair•2d ago•105 comments

Launch HN: Alter (YC S25) – Access control for multi-user agents

https://alterauth.com/
58•simfinii•23h ago•25 comments

PHP 8.2.35

https://github.com/php/php-src/releases/tag/php-8.2.35
994•tuyuri•16h ago•536 comments
Open in hackernews

WorldTUI: I trained a 360M model to pretend to be a terminal

https://simedw.com/2026/10/21/world-model-terminal/
57•alphax314•3h ago

Comments

obilgic•57m ago
There was a whole fake gui OS with similar concept right
prodigycorp•35m ago
athropic had an experiment that did exactly this. we need more experimentation in this space.
andai•9m ago
VibeOS - Fully Hallucinated Operating System

https://www.youtube.com/watch?v=z3pV6FHvcgM

htoz1•54m ago
Where is the point of this project if i have terminal ?
dotancohen•50m ago
Because it was there
mehmetoguzderin•38m ago
Terminal world models have many applications, but empirical observation of successful pretraining or fine-tuning can be significant for encapsulating from rather expensive Bash traces, such as ones obtained from systems with scarce accelerators, etc.
YmFzZTY0ZGFkZHk•34m ago
Sometimes it's just fun to do something absurd because you can. This is one of the reasons software is exciting.
good-idea•31m ago
because curiosity is the most fun of virtues!
oilkillsbirds•51m ago
Very cool. Fun project. I feel like there’s a wide field to be explored once you understand fine tuning good existing models
psandor•44m ago
Reminds me of this project from Microsoft. A Quake II “simulation” model where the whole application was just a model.

https://www.microsoft.com/en-us/research/articles/whamm-real...

soulofmischief•19m ago
These are called world models, which is why the project is called WorldTUI. For another example, take a look at Google's recent Genie models. https://deepmind.google/blog/genie-2-a-large-scale-foundatio...
__mp•40m ago
It would be interesting to expose this on the internet. I wonder what the scrapers etc… would do with it.
prodigycorp•36m ago
Wow! This this reminds me of when we tried doing this with gpt 3 and.. it worked!

The tmux thing had me confused for a moment until i realized it was emulating tmux. Super fun. Is the model window width aware?

yjftsjthsd-h•23m ago
Wait, how did it `sleep 3`? I thought LLMs didn't have a concept of time
peder•12m ago
It’s still rendering frames constantly, no? So it just figures the next X frames are all doing nothing?
sprobertson•12m ago
from the post:

Time is an input too. If nothing is pressed, the model gets a WAIT token with a duration instead of a key, which is how snake keeps moving on its own.

yjftsjthsd-h•5m ago
That seems as likely as anything, but that just moves the problem to counting, doesn't it? The model gets the duration, but I was under the impression LLMs struggled with the arithmetic that would make that useful.
thorum•16m ago
I regretfully suspect this actually is the future of UI. You just need smart local models running fast enough to keep the framerate high, but that’s inevitable in the long term, isn’t it?
gr_norm•4m ago
Why would you do that instead of having the model do the task you're using the UI for, and having it generate UI widgets on-the-fly to take rich input as needed?
andai•3m ago
>I used Qwen2.5-Coder-0.5B as the starting point. I needed a small model to get close to realtime performance. Being a coding model, it had also already seen plenty of CLI usage.

Interesting. I wanted to ask, how that model compares with the fine-tuned version, but I just saw the author provided benchmarks. Interesting that fine-tuning improved the memory so significantly, wouldn't have expected that.

Though, on second thought, I had an experience where I was able to, as a human, fine-tune my memory beyond what I previously thought was possible!