✷ The archive · 11 labs · 3,050 dispatches
The labs
No more opening a dozen official blogs every morning. First-hand releases from OpenAI, Anthropic, DeepMind and the rest, each with its substance pulled out.
Building self-improving tax agents with Codex
OpenAI and Thrive Holdings developed Tax AI, a self-improving agent for Crete accountants that uses a Codex-driven loop to automate complex tax returns with up to 97% accuracy.
Reachy Mini Local Speech Backend Integration
Hugging Face has released a local speech-to-speech pipeline for Reachy Mini, allowing the robot to handle conversations fully locally using a cascaded VAD, STT, LLM, and TTS stack.
Delta Weight Sync in TRL Enables Trillion-Parameter Model Training with Minimal Bandwidth
Hugging Face announced Delta Weight Sync in TRL, a feature that reduces weight synchronization bandwidth in async RL training by over 100x by transmitting only sparse weight changes via Hugging Face Buckets, enabling disaggregated training without shared clusters.
OpenAI Election Information and Safeguards in 2026
OpenAI has announced a comprehensive set of safeguards for the 2026 election cycle, focusing on reliable information surfacing, cyber infrastructure defense, content provenance, and the prevention of model bias.
Warp Open Agentic Development and Oz Orchestration Platform
Warp is implementing Open Agentic Development using GPT-5.5 and its Oz orchestration platform to shift software engineering toward human-supervised agent fleets.
Anthropic Opens Milan Office to Support Italian Enterprise and Research
Anthropic has opened a new office in Milan, its sixth European location, to support Italian enterprises, developers, and research in the responsible scaling of Claude AI.
Grok Integration in Kilo Code
xAI has integrated Grok models, including Grok Build, into the Kilo Code agentic engineering platform for users with X Premium+ or SuperGrok subscriptions.
Coding Agents in the Social Sciences
A survey of 1,260 social scientists reveals that while 81% use AI chatbots, only 20% have adopted autonomous coding agents, with significant adoption gaps across gender, career stage, and institutional prestige.
EAGLE 3.1 Release Notes: Enhancing Speculative Decoding Robustness and Efficiency
EAGLE 3.1 introduces architectural improvements to solve attention drift, doubling acceptance length in long-context workloads and significantly increasing throughput via vLLM and TorchSpec integration.
Anthropic Appoints KiYoung Choi as Representative Director of Korea
Anthropic has appointed KiYoung Choi as Representative Director of Korea to lead go-to-market strategies and support the high adoption rate of Claude.ai in the South Korean market.
Hugging Face AI Agent Glossary: Defining Harness, Scaffold, and Agent Architecture
Hugging Face provides a standardized vocabulary for AI agents, defining the agent as the combination of a model, a harness for execution, and scaffolding for behavior definition.
OpenAI, Grupo Folha, and Grupo UOL Strategic Content Partnership
OpenAI has partnered with Brazil's Grupo Folha and Grupo UOL to integrate high-quality Brazilian journalism into ChatGPT, providing 900 million weekly active users with grounded, attributed reporting.
OpenAI Analysis: AI as a First Hire for Small Businesses
OpenAI reports that four million US users utilized ChatGPT in March 2026 to support small business operations, reducing administrative burdens and lowering the fixed costs of entrepreneurship.
Anthropic How We Contain Claude Across Products – Engineering Overview
Anthropic announced its multi‑layer containment strategy for Claude agents, detailing environment sandboxing, human‑in‑the‑loop controls, and egress filtering to cap blast radius across Claude.ai, Claude Code, and Claude Cowork.
Anthropic Co-Founder Chris Olah on Pope Leo XIV's Magnifica Humanitas Encyclical
Anthropic co-founder Chris Olah emphasizes the need for external moral and philosophical oversight from religious and civil society leaders to balance the commercial and geopolitical incentives of AI development.
xAI Grok Build Early Beta Release
xAI has launched Grok Build, a coding agent for professional software engineering designed for complex coding tasks through a plan-review-approve workflow and parallel subagent execution.
Mistral AI Acquires Emmi AI to Advance Physics AI for Industrial Engineering
Mistral AI has acquired Emmi AI to integrate Physics AI capabilities into its ecosystem, enabling real-time simulations and digital twins for high-stakes industrial sectors like aerospace and automotive.
Mistral Medium 3.5 and Vibe Remote Agents Release
Mistral AI has released Mistral Medium 3.5, a 128B open-weight model, and introduced remote cloud-based coding agents in Mistral Vibe and a new Work mode in Le Chat for asynchronous, multi-step task execution.
Mistral AI Connectors release notes / what's new
Mistral AI has released Connectors in Studio, enabling developers to integrate enterprise data via built-in and custom Model Context Protocol (MCP) connectors with programmatic API/SDK access.
Virgin Atlantic Software Development with OpenAI Codex
Virgin Atlantic utilized OpenAI Codex to launch a revamped mobile app with zero P1 defects and reduce legacy codebase size by up to 80%.
OpenAI Codex Named a Leader in 2026 Gartner Magic Quadrant for Enterprise AI Coding Agents
OpenAI has been recognized as a Leader in the Gartner Magic Quadrant for Enterprise AI Coding Agents, reflecting the scale and governance capabilities of Codex.
Anthropic Research: Measuring LLM Exploit Development Capabilities
Anthropic's Claude Mythos Preview demonstrates a step-change in ability to develop end-to-end exploits across V8, Linux kernel, and smart contracts, significantly outperforming previous frontier models.
Project Glasswing: Anthropic Initial Update on AI-Driven Vulnerability Discovery
Anthropic's Project Glasswing has used the Claude Mythos Preview model to identify over 10,000 high- or critical-severity vulnerabilities in systemic software, shifting the cybersecurity bottleneck from discovery to verification and patching.
Google DeepMind Accelerator Program in Asia Pacific
Google DeepMind has launched an AI for the Planet accelerator program in the Asia-Pacific region to help startups, research teams, and nonprofits scale frontier AI solutions for environmental risks.
AdventHealth and OpenAI: Scaling AI to Reduce Clinical Administrative Burden
AdventHealth is deploying ChatGPT for Healthcare to automate time-intensive documentation and clinical workflows, reclaiming clinician time to expand patient care capacity.
Qwen3.7-Max agent model release
Qwen released Qwen3.7-Max, a new agent-focused foundation model that excels at coding, office automation, and ultra-long-horizon autonomous tasks, now available via Alibaba Cloud Model Studio.
OpenAI Education for Countries Program Update
OpenAI is expanding its Education for Countries initiative to include Singapore and scaling research-driven AI deployments across its first cohort of nations to improve learning outcomes through agentic AI.
OpenAI Model Disproves Planar Unit Distance Conjecture
An OpenAI general-purpose reasoning model has autonomously disproved a 80-year-old conjecture in discrete geometry by discovering a polynomial improvement over the square grid construction.
Codex with GPT-5.5 Speeds Up Code Review at Ramp
Ramp engineers use Codex with GPT-5.5 to get substantive pull‑request feedback in minutes instead of hours and to build internal agentic tools such as On‑Call Assistant.
OpenAI for Singapore Partnership Announcement
OpenAI has launched 'OpenAI for Singapore,' a S$300 million partnership with the Ministry of Digital Development and Information to establish an Applied AI Lab and develop local AI talent.
OlmoEarth v1.1 release notes / what's new
Hugging Face and AllenAI have released OlmoEarth v1.1, a family of Earth observation models that reduces compute costs by up to 3x while maintaining performance similar to v1.
OpenAI Content Provenance Updates
OpenAI is implementing a multi-layered provenance system using C2PA conformance, Google SynthID watermarking, and a new public verification tool to increase the transparency and durability of AI-generated content signals.
Qwen3.5-LiveTranslate-Flash Release Notes
Qwen3.5-LiveTranslate-Flash is a simultaneous interpretation model built on Qwen3.5-Omni that provides real-time, multimodal translation across 60 languages with ultra-low latency and voice cloning.
Ettin Reranker Family v1 release notes
Hugging Face released the Ettin Reranker family, six CrossEncoder rerankers from 17M to 1B parameters distilled from mxbai-rerank-large-v2, achieving state-of-the-art retrieval reranking quality and speed.
Anthropic Initiative to Incorporate Diverse Wisdom Traditions into AI Moral Formation
Anthropic is engaging with scholars, clergy, and philosophers from over 15 religious and cross-cultural groups to inform the moral formation and character development of its AI systems.
Grok Integration for OpenClaw
xAI has enabled Grok models for use within OpenClaw, an open-source, local-first personal assistant that integrates with multiple messaging platforms.
KPMG and Anthropic Strategic Alliance for Global Claude Integration
KPMG is integrating Anthropic's Claude across its global workforce of 276,000+ employees and embedding the AI into its Digital Gateway platform to enhance tax, legal, and private equity services.
Google DeepMind: Fast-tracking genetic leads to reverse cellular aging
Google DeepMind's Co-Scientist AI is accelerating cellular aging research by identifying novel genetic factors and reducing data analysis time from six months to a few days.
PaddleOCR 3.5 release notes / what's new
PaddleOCR 3.5 introduces Hugging Face Transformers as a supported inference backend, allowing OCR and document parsing models to integrate more seamlessly into PyTorch-based workflows.
OpenAI and Dell Technologies Partnership for Enterprise Codex Deployment
OpenAI and Dell Technologies have partnered to enable the deployment of Codex in hybrid and on-premises enterprise environments via the Dell AI Data Platform and Dell AI Factory.
vLLM x Novita AI: PegaFlow for Production-Grade External KV Cache
vLLM and Novita AI announce PegaFlow, an external KV cache service that decouples KV cache from vLLM workers, enabling faster startups, higher throughput via cache sharing, and RDMA-based cross-node access.
xAI Grok 4.3 Skills Release
xAI has launched Skills for Grok 4.3, enabling persistent expertise and the automated generation of production-ready office documents across web, iOS, and Android platforms.
Anthropic Acquires Stainless
Anthropic has acquired Stainless to enhance the connectivity of Claude agents by integrating advanced SDK and MCP server tooling into the Claude Platform.
Project Genie Street View Integration
Google DeepMind has integrated Google Street View imagery into Project Genie, allowing the general-purpose world model to generate interactive, grounded real-world environments for AI agents and users.
Gemini Omni Flash Release
Google DeepMind has introduced Gemini Omni Flash, a natively multimodal model capable of generating and editing high-quality videos from combinations of text, image, audio, and video inputs.
Google Antigravity 2.0 Announcement
Google DeepMind announced Google Antigravity 2.0, though the provided source material consists only of a redirect notice and contains no technical details.
Gemini for Science: AI Tools for Scientific Discovery
Google DeepMind has introduced Gemini for Science, a suite of AI tools and experiments designed to accelerate the scientific method through hypothesis generation, computational discovery, and literature synthesis.
Google DeepMind Content Transparency and Verification Tools Update
Google DeepMind is expanding SynthID watermarking and C2PA Content Credentials across Search, Gemini, Chrome, and Pixel to help users identify AI-generated and authentic camera-captured content.
Google DeepMind and Singapore National Partnership for AI
Google DeepMind is partnering with the Singapore Government to deploy frontier AI across healthcare, education, and sustainability to potentially create S$3.3 billion in economic value by 2040.
Google DeepMind Co-Scientist for Infectious Disease Research
Professor Clare Bryant of the University of Cambridge is using Google DeepMind's Co-Scientist to identify molecular switches in zoonotic pathogens, reducing the time to identify precise amino acid targets from years to months.