State Dept picks OpenAI. Plus: who verifies AI-written code? And new kernels surpass expert-written ones.
The US State Department yesterday announced it will phase out Anthropic in favor of OpenAI for internal use. This follows weeks of public friction between Anthropic and the Pentagon, signaling that government agencies are making concrete choices about their AI vendors. Meanwhile, a new analysis raises a critical question for builders: who verifies the software when AI is the author?
doubleAI's WarpSpeed uses AI for GPU performance engineering, generating and verifying hyper-optimized graph algorithm kernels. Its doubleGraph product, a drop-in replacement for NVIDIA's cuGraph, shows a mean 3.6x speedup (18% over 10x) across architectures, with 100% verified correctness.
'recall' is a new skill for Claude and Codex, enabling full-text search across past conversation sessions. It builds a SQLite FTS5 index over chat logs, using BM25 ranking with recency boosts, and is dependency-free for easy installation.
Ben Thompson argues geopolitical factors like Taiwan and chip manufacturing will influence AI development more than abstract safety debates. He questions if private AI labs can resist government control as AI becomes a critical state asset. This provides builders with context on how geopolitical forces, particularly around chip supply and state control, could shape future AI development and access to resources.
An investigation reveals data annotators in Kenya are exposed to sensitive personal videos and images from Meta's AI smart glasses, including intimate moments. This happens due to inadequate anonymization and automatic data processing mandatory for AI functionality. The report highlights a gap between Meta's marketing claims of user control and the reality of data sharing.
The U.S. State Department, Treasury, and HHS are phasing out Anthropic's AI products, including Claude. This decision comes from a disagreement with the Trump administration over limits on military use of its AI, specifically for autonomous weapons targeting and domestic surveillance. These agencies will now use OpenAI and Google Gemini models, while OpenAI simultaneously secured a deal to deploy its tech within the Defense Department's classified network.
Intent launched as a developer workspace for AI apps, featuring coordinated agents that break down tasks. It uses a 'living spec' that syncs with code changes and integrates coding, browsing, and Git. The platform supports multiple LLMs and custom models to keep code, specs, and agent outputs aligned.
A new paper introduces Evo 2, a large biological foundation model trained on 9 trillion DNA base pairs with a 1 million token context window. Evo 2 predicts genetic variation impacts and generates new genomic sequences across all domains of life, with model parameters and training dataset released open-source.
As AI speeds up software development, traditional code review and testing can't keep pace with AI-generated errors. Leo de Moura argues that formal verification, powered by AI and tools like Lean, is the answer, shifting verification from manual friction to mathematical proof.
Sean Goedecke argues that giving LLMs a 'personality' is a fundamental engineering requirement, not a marketing gimmick. He states base models are unmanageable, needing engineers to define a behavioral 'region' for usefulness and safety. This persona makes LLMs controllable tools.
This analysis argues the current AI boom is an 'information war' by companies like OpenAI and Anthropic. The analysis claims these firms use media leaks to inflate financial prospects while downplaying significant costs and unprofitability. It also criticizes their involvement with the US military, suggesting ethical stances are performative.
Unsloth now supports fine-tuning for the Qwen3.5 model family, covering both text and vision models. It trains 1.5x faster and uses 50% less VRAM than standard setups, with options to preserve reasoning ability and export to GGUF/vLLM.