First screen, captured . Open the video.
Captured . Click for the whole screenshot.
136 points by porridgeraisin | 155 comments
A LessWrong shortform comment by Arjun Panickssery arguing AI agents show poor research judgment in autonomous R&D side projects like a Scrabble engine and philosophy research.
25 points by kimi | 1 comments
GitHub repository for Mocktastic, an educational C++ Meshtastic-compatible LoRa client for ESP32 boards (LilyGO T3) built with Arduino IDE, with decoder and DHT22 sensor modes.
37 points by boldi | 25 comments
GitHub repository for tmux-restore, a tmux plugin that restores sessions, layouts, scrollback and AI coding agent conversations after reboots.
115 points by adrbn | 48 comments
GitHub repository for bo2-mac, a tool that statically recompiles Call of Duty: Black Ops II from 32-bit x86 Windows to native ARM64 macOS using Remill, LLVM and Metal.
22 points by triska | 6 comments
GitHub repository for zipc, a work-in-progress compiler from Prolog to Clocksin's ZIP abstract machine plus a C VM, intended as a backend for trilog.
60 points by GloVin | 76 comments
A Stork AI daily newsletter issue led by a report that Apple cut some engineers' daily AI spending cap to $200, plus other AI news items.
338 points by canjobear | 229 comments
arXiv paper modeling the spread of LLM use as a viral contagion, showing tipping points toward persistent cognitive dependence and conditions for cognitive immunization.
285 points by emschwartz | 41 comments
Blog post arguing that a model's weight edit fixing Qwen3.5-0.8B thinking loops can be reproduced by solving a linear program, with fewer changed weights.
334 points by Cynddl | 333 comments
arXiv paper finding that fine-tuning language models to be warm and empathetic raises error rates and makes them more sycophantic on safety-critical tasks.
497 points by allinonai | 231 comments
arXiv paper introducing HybridAgent, LLM agents that switch between latent-space reasoning and text-based tool interaction, trained with HybridSFT and HybridRL.
545 points by davidbarker | 369 comments
OpenAI post announcing that GPT-5.2 derived a new result in theoretical physics; the captured page shows only a loading logo.
3 points by merksittich | 2 comments
Epoch AI report showing the open-weight model GLM 5.3 exploited Firefox's patch gap to compromise six recent releases for about $1,000 each.