Loading activity…
@a1zhang
student @mit_csail @nlp_mit
AI
Arora shared defense slides titled “Causal Interpretability for Analysing and Improving Language Model Architectures.”
AI
PrimeIntellect says the rewrite reaches usable input about 13 times faster and uses 83% less startup memory than before.
AI
The poster said they planned to attend the Simons workshop during the week of October 6, 2026.
AI
Underdog says its Mac assistant stores personal data on-device, with iOS, Windows and Android versions planned.
AI
Posts on X congratulated Nathan Lambert and Trillium Labs after a launch, with several users describing the effort as open, science-oriented or focused on open post-training. Lambert later said some criticism came from Substack and more general-audience AI followers, not the technical AI community.
AI
A post about the project claims a 65% score gain with the same compute on a 24-hour multi-repository agent-swarm task.
AI
An organizer says the club will host talks on various topics. Alok, scheduled for October 1, is known for a blog post and video on group theory and positional encodings.
AI
A blog author explores designing models to fit their harnesses—the software that runs them—rather than designing harnesses around the models.
AI
Arena describes a comparison of Claude Code, Codex and Pi. The researchers say harness choice had little effect on task success rate but can significantly affect cost.
AI
Ehsan Amid from Google DeepMind shares the Core Automation post on X.
AI
Posts discuss OpenAI disclosure at Black Hat on agent behavior during model training.
AI
Company presents the harness as token-efficient for autonomous coding work.
