Full-text search across 70867 articles. Combine with topic and date filters; results sorted by relevance.
Researchers propose DISTAL, a dual-prior framework for structure-agnostic materials property prediction that combines self-supervised compositional pretraining with knowledge distillation from a pretrained ALIGNN teacher…
Researchers have introduced Faster Flash Decoding (FFD), a hardware-algorithm co-design framework that achieves up to 11.6x kernel-level speedup and 2.37x end-to-end throughput improvement for long-context LLM decoding, …
Researchers propose AnySearch, a reinforcement learning framework that enables a single LLM-based search agent policy to adapt to any budget constraint at deployment, outperforming baselines across seven QA benchmarks an…
Researchers from ServiceNow introduced profile behavioral grounding, a framework that extracts open-ended user profiles from anonymized social media posts to personalize large language models (LLMs). In evaluations acros…
A developer rants on Hacker News that the AWS Management Console is broken for cross-region S3 bucket replication, citing missing monitoring and a batch operation that cannot select source and destination buckets in diff…
Databricks has released the Big Book of AgentOps, a guide to the people, processes, and technologies needed to deploy governed and reliable AI agents at enterprise scale, covering architecture, evaluation, observability,…
An evaluation by the SLayer team found that an AI agent using the open-source version of Cube, an incumbent open-core semantic layer, solved zero of 10 tasks from the LiveSQLBench-Lite benchmark, while a SLayer-using age…
Founder Namanyay Goel reports that using Claude to automate outreach to people who engaged with competitor or own posts on LinkedIn yielded a nearly 2.5x better response rate, calling it a 'cheat code' for early-stage st…
In a recent interview with Lex Fridman, DHH argued that users typically utilize only 5% of a software's features, and coding agents now enable building exactly that 5%, allowing replacement of existing tools. He is concu…
Inception released Mercury 2.5, a diffusion large language model (dLLM) that generates tokens in parallel, achieving 1,107 tokens/sec on standard GPUs and a 10+ point intelligence jump over Mercury 2, with pricing at $0.…
Anthropic released Claude Fable 5.1 and Mythos 5.1 on Tuesday, with Fable 5.1 priced an estimated 25% less than Fable 5 for typical workloads and about 45% less for agentic workloads due to reduced cache read costs. The …
GitHub and Codex Skin launched multiple open-source ecosystem tools for OpenAI's Codex, including oh-my-subagents, a runtime for persistent parent-subagent delegation, and Codex Skin, a theming platform for customizing t…
BMC's 2026 Mainframe Survey of over 1,300 professionals shows a shift from experimental AI to operational use, with 94% expressing long-term confidence in mainframes and 45% prioritizing AI implementation. However, trust…
Claude Fable 5.1 is now available on Agent Platform, bringing frontier intelligence to production environments for long-running, high-stakes tasks. The model is accessible via the Model Garden, enabling integration acros…
TaxisDB, a database project, published a series of blog posts in 2026 covering topics such as Substrate Architecture, AI-generated code readability, data types and constraints, and RDF/OWL ontology visualization, with th…
Anthropic released Claude Fable 5.1, an upgrade to its most capable AI model, claiming higher performance on coding and problem-solving tasks and up to 45% cost savings for agentic workloads. The company also launched My…
Simon Willison released a new GeoJSON Map Viewer tool that displays GeoJSON files on a map and exports them as PNG images, built with assistance from GPT-5.6-Sol, Claude Code for web, and Fable 5.1. The tool was created …
Joshua Penman released Semantic Overlays, a method using small trained adapters on a frozen model to mitigate LLM prompt injection, claiming it brings the injectable Qwen-3.5-9B to state-of-the-art scores on all prompt i…
Atri10 released Executor, an open-source initiative-scoped workflow system for coding agents that structures work from intake through review with a strict per-initiative ID namespace and separated thinking and execution …
Google DeepMind chief Koray Kavukcuoglu said frontier AI leadership is the only thing that matters for the company, admitting its current models are 'a little bit below the frontier' but expressing 100% certainty it will…