Masked Diffusion Language Agents for Tool-Integrated Chemical Reasoning
Shows a masked diffusion LM as the reasoning core of a tool-using chemistry agent.
AI Storyline
3 items · 2 sources · 3 days
Operational story trace
Diffusion Language
Follow in this browser to see new updates on your Live feed.
Current stateResearch · developing
Latest change
A new paper distills a decoding schedule so diffusion LLMs commit tokens in fewer rounds, targeting the parallel-generation speedup that has been hard to realize.
Earlier contextThe story so far
Diffusion language models first appeared here as the backbone of a tool-using chemistry agent. Days later they turned up on the offensive side, as a way to craft fluent black-box attacks on LLMs.
editor-curated · source-linked
Arc
Oct 9Oct 26 · now
AGENTS · Oct 9Masked diffusion LMs drive a tool-integrated chemistry agent1 source · show source ▾
RISK · Oct 22Diffusion LMs used for perplexity-aware black-box attacks1 source · show source ▾
Diffusion Language Models for Perplexity-Aware Black-Box Attacks
Uses a diffusion LM to generate low-perplexity attack prompts against black-box LLMs.
NOW · Oct 26Schedule distillation aims to cut dLLM decoding rounds1 source · show source ▾
Learning When to Commit: Schedule Distillation for Fast Diffusion Language Models
Proposes learning when to commit tokens, reducing the decoding rounds dLLMs need.
Day 1 Friday, Oct 9, 2026
Masked Diffusion Language Agents for Tool-Integrated Chemical Reasoning
Shows a masked diffusion LM as the reasoning core of a tool-using chemistry agent.
Day 2 Thursday, Oct 22, 2026
Diffusion Language Models for Perplexity-Aware Black-Box Attacks
Uses a diffusion LM to generate low-perplexity attack prompts against black-box LLMs.
Day 3 Monday, Oct 26, 2026
Learning When to Commit: Schedule Distillation for Fast Diffusion Language Models
Proposes learning when to commit tokens, reducing the decoding rounds dLLMs need.
What to watch — open questions
- Do the schedule-distillation speedups hold on independent benchmarks at equal output quality?
- Does any inference framework or vendor ship diffusion LLM serving with parallel decoding?
How this thread was built
editor wrote the arc · 3 beatswatcher 1 status change
Storylines are threaded mechanically from the feed: stories that share a distinctive anchor across multiple days and sources. Each item links to its original source. The evidence trace, current state, and open questions are written by the editor routine and refreshed whenever a new beat lands.