AI News
Latest AI engineering news, updated daily.
Ai Agents
AWS Tackles Agent Drift With Bedrock AgentCore Optimization
AWS has introduced AgentCore Optimization in preview to automate prompt updates and A/B testing, alongside a new desktop AI assistant called Amazon Quick.
Aws Bedrock · Agent Drift · Prompt Optimization
Ai Engineering
Runpod Flash Removes Container Overhead for AI Inference
The open-source Flash Python SDK allows developers to convert local functions into auto-scaling serverless AI inference endpoints without Dockerfiles.
Serverless Inference · Cloud Infrastructure · Open Source Sdk
Ai Engineering
DeepSeek V4 Pro Trails GPT-5.5 by 8 Months in NIST Benchmarks
The Center for AI Standards and Innovation evaluated DeepSeek-V4-Pro, placing its capabilities eight months behind U.S. frontier models while matching GPT-5.
Deepseek V4 · Nist Benchmarks · Llm Evaluation
Ai Engineering
TPU v5p Inference Speeds Triple With DFlash Block-Diffusion
Google and UCSD researchers released DFlash, a block-diffusion speculative decoding method that achieves a 3.13x average inference speedup on TPU v5p hardware.
Llm Inference · Google Tpu · Speculative Decoding
Ai Engineering
PyTorch Lightning 2.6.2 Drops Self-Spreading Credential Stealer
Threat actors hijacked the PyPI credentials for PyTorch Lightning to publish two malicious versions containing a self-propagating credential stealer.
Pypi Security · Pytorch Lightning · Supply Chain Attack
Ai Engineering
ChatGPT Images 2.0 Adds Multilingual Text and Thinking Mode
OpenAI released ChatGPT Images 2.0 with the gpt-image-2 model, adding agentic web search, 2K resolution, and non-Latin script rendering capabilities.
Generative Ai · Multilingual Support · Agentic Workflows
Ai Engineering
Hybrid ML-KEM Arrives in Cloudflare IPsec for WAN Tunnels
Cloudflare has introduced general availability for post-quantum IPsec tunnels using a hybrid ML-KEM handshake compatible with Cisco and Fortinet hardware.
Post Quantum Cryptography · Hybrid Ml Kem · Ipsec Tunnels
Ai Engineering
CVE-2026-31431 Grants Local Root via Linux Page Cache Write
A logic bug in the Linux kernel's userspace crypto API allows unprivileged local users to gain root access across major distributions dating back to 2017.
Linux Kernel · Privilege Escalation · Security Vulnerability
Ai Engineering
xAI Ships 2-Minute Voice Clones and Grok 4.3 APIs
xAI has introduced a fast custom voice cloning suite and a new Voice Library alongside the launch of its 1M-context Grok 4.3 model.
Voice Cloning · Grok Api · Multimodal Ai
Ai Engineering
Meta Acquires ARI for Open Humanoid Intelligence Platform
Meta has acquired robotics startup Assured Robot Intelligence to build foundational control and behavioral models for third-party humanoid hardware.
Robotics Ai · Foundational Models · Meta Acquisition
Ai Engineering
Amazon Bedrock Gains GPT-5.5 and Codex in $50B OpenAI Deal
Following the end of Microsoft's exclusive distribution rights, Amazon Web Services has introduced OpenAI's GPT-5.5 and Codex models to the Bedrock platform.
Amazon Bedrock · Openai Partnership · Gpt 5 5
Ai Engineering
Pentagon Approves Eight AI Vendors For IL7 Classified Networks
The Department of War has authorized models from OpenAI, Google, and six other vendors for classified networks following its dispute with Anthropic.
Defense Technology · Classified Networks · Government Procurement
Ai Engineering
Pre-Trial AI Toxicity Filters Isolate IRS4 Cancer Target
Researchers at St. Jude used AI safety filtering to identify IRS4 as a high-potential target for solid tumors by predicting toxicity before clinical trials.
Ai Safety Filters · Drug Discovery · Biotech Innovation
Ai Agents
On-Call Agent TasksMind Drops Incident Resolution to 60 Seconds
TasksMind has introduced an autonomous incident response agent that writes patches and resolves production alerts in under 60 seconds.
Autonomous Agents · Incident Response · Devops Automation
Ai Coding
RLHF Leak Forces OpenAI to Ban Goblin Metaphors in Codex
OpenAI hardcoded a ban on goblin metaphors in the GPT-5.5 Codex CLI after an unintended reinforcement learning generalization corrupted bug descriptions.
Rlhf · Openai Codex · Reinforcement Learning
Ai Engineering
Grok Training Partly Relied on OpenAI Model Distillation
Elon Musk testified in federal court that xAI partly relied on model distillation from OpenAI to validate and train the Grok chatbot.
Model Distillation · Large Language Models · Xai Grok
Ai Agents
GPT-5.5-Cyber Launch Restricted to Trusted Defense Partners
OpenAI has launched GPT-5.5-Cyber for autonomous vulnerability detection, restricting access to government and critical infrastructure through its TAC program.
Cybersecurity · Autonomous Agents · Government Ai
Ai Agents
128B Mistral Medium 3.5 Moves Vibe Coding Agents to the Cloud
Mistral AI's new 128-billion parameter dense model introduces configurable reasoning alongside asynchronous cloud-based execution for coding agents.
Mistral Medium · Large Language Models · Vibe Coding
Ai Coding
Anthropic's Claude Security Beta Patches Code With Opus 4.7
Anthropic released the public beta of Claude Security, an Opus 4.7-powered defensive tool that scans codebases for vulnerabilities and generates patches.
Anthropic · Vulnerability Detection · Automated Patching
Ai Agents
DeepMind AI Co-Clinician Logs Zero Critical Errors in 97 Cases
Google DeepMind introduced the AI co-clinician to support physicians in real-world care settings, logging zero critical errors across 97 primary care cases.
Healthcare Ai · Google Deepmind · Clinical Decision Support