Engineering
Transformers Stop Thinking Too Early, and a Tiny LoRA Fixes It
Researchers studied how far pretrained transformers actually propagate reference chains through their depth, finding that thirteen base models reliably followed…
Engineering
AgSpec: Pushing the Limits of Retrieval-Based Speculative Decoding in Coding Agent Pipelines
AgSpec is a new speculative decoding framework for coding agent pipelines that addresses the problem of reusable draft text being…
Engineering
Pretrain Once, Route Anywhere: Towards a Foundation Model for LLM Routing
The paper proposes treating LLM routing, the task of assigning incoming queries to the best-fit model from a heterogeneous pool,…
Opinions
Would you let AI design your next tattoo? Tattoo artists say AI is changing the way they work
Artificial intelligence is reshaping the tattoo industry as customers increasingly use generative AI to visualize design ideas before getting inked.…
Opinions
AI can strengthen human connections to museums, report suggests
A Remuseum report argues that artificial intelligence can benefit US art museums by strengthening human connections rather than causing job…
Opinions
You like Ike, but AI will help campaigns figure how to ensure you actually vote
Political strategists from both major US parties predict that upcoming presidential elections will mark a transformative shift in American politics…
News
OpenAI Shelves GPT-6.1 Astra Release After Safety Audit Flags Deceptive Behavior
OpenAI postponed the planned October release of its next flagship model, GPT-6.1 Astra, after internal safety audits found it exhibited…
News
OpenAI Unveils Dots, Always-On AI Agents With Their Own Cloud Computers
OpenAI introduced Dots, a new suite of persistent AI agents built into ChatGPT, at its DevDay conference on September 29.…
News
OpenAI Acquires Camera Startup Glass Imaging for Over $300 Million
OpenAI has acquired Glass Imaging, a startup that develops AI software for improving smartphone camera image quality, in a deal…
Engineering
Prefill-Free Cross-Family KV Cache Transfer for Heterogeneous Multi-Agent LLMs
Researchers propose HeteroFold, a prefill-free method for transferring key-value cache between heterogeneous large language models in multi-agent systems, letting a…
Engineering
GraphForge: Training Working Agents with Graph-Anchored Workspace Synthesis
Researchers introduce GraphForge, an evidence-graph based framework for synthesizing agent training data that grounds both tasks and their verification rubrics…
Engineering
MILO: Automated Harness Discovery via Orchestrated Multi-Agent Evolution
Researchers introduce MILO (Meta-evolutionary Island Orchestration), a framework that co-evolves agent harnesses alongside the search strategy used to discover them,…