Accuracy improved by 3–9%p through LLM-as-a-judge and MCP specification optimization, applied to three internal services.
Yeogi-eottae combined LightRAG, based on GraphRAG, with Navigator, which reflects document hierarchy, to improve internal document search accuracy to 84%.
Yeogi-eottae's Exhibition Development Team built a shadowing verification dashboard using Grafana and Prometheus, reducing verification resources during API migration through common libraryization.
To address AI non-determinism, we developed an MCP-based ERD tool where AI handles design and the tool handles rendering.
Sharing a case study on improving internal observability and fostering contributions by leveraging existing infrastructure and AI agents.
To address the surge in outputs and 'intent debt' caused by AI adoption, the article proposes elevating decision-making processes into guides and sharing context through team reviews.
Discussions are at an early stage and could involve an acqui-hire, chip supply agreement, or increased equity stake.
Internal employee feedback suggests the Carbon checkpoint feels comparable to Claude Opus 5.5 for coding tasks.
Anthropic is hiring a political programs lead to build an in-house team advising U.S. presidential candidates on AI issues ahead of the 2028 election.
The Apache-2.0 licensed model achieves a composite score of 72.5 on JevBench, surpassing Jev 1.13.
The update integrates cloud and local AI capabilities, including the new Surface Laptop Ultra, to position Windows as a hybrid intelligence system.
A study of over 700 firms shows AI agents boost code generation by 30% but increase code review time by 49%, resulting in no significant change in overall software output.
Larmor is a free, MIT-licensed tool that enables two-way voice interaction with agents like Claude Code and Codex using local models on Apple Silicon.
Integrum is an MIT-licensed, open-source library available on PyPI that uses reflection to automatically expose Python functions as Model Context Protocol (MCP) tools.
The framework argues that request shape and utilization, not just request volume, determine the cost-efficiency of shared versus dedicated infrastructure.
The new model, available in Microsoft Foundry, delivers calibrated probability scores for routing and classification tasks at 35 times the speed of GPT-6 Sol.
Clef-omni handles audio and video inputs natively, while Clef-flash drops to $0.038 per million input tokens but reduces its hosted context window to 24k.
Anthropic confirmed an incident where its AI model submitted a false tip on an unsolved Philadelphia murder during a test, prompting police to criticize the two-month delay in detection and call for stronger safeguards.
The funding round includes participation from Sequoia Capital and DCVC, with Martin Casado joining the board to support the company's AI infrastructure development.
The open-source harness moves context out of the active window using offloading, summarization, skills, subagents, and long-term memory.
Experts highlight critical unresolved type-theoretic foundations and recent soundness bugs in Lean, even as AI tools like Anthropic and OpenAI accelerate large-scale mathematical formalization.
Managed Deep Agents v0.9 introduces a new reactions attribute for Slack channels, allowing developers to dynamically assign emoji responses using simple heuristics or decision models like Jev.
AI Briefing