AI News

Daily AI News for Builders

The stories worth reading, for builders and indie hackers. Updated all day.

Hybrid AI Workloads Gain Energy Boost From Quantum Hardware

IonQ and QuantumBasel released a study showing hybrid AI workloads could gain significant energy advantages by leveraging quantum hardware as they scale. This is big for teams looking to optimize their AI systems' efficiency, especially in data centers where power costs are high. The study highlights that as these systems grow, the benefits of using quantum tech become more pronounced. Interlune's cryogenic technology aims to support this growth with a new Helium-3 supply expansion.

Jul 21, 2026 · 3 min read

tech

GitLab 19.2 Adds Agentic Automation for Security and Review

GitLab has released version 19.2 of its DevSecOps platform, introducing agentic automation aimed at enhancing security and review processes. The release includes four key features: Dependency Scanning Auto-Remediation, Security Review Flow, GitLab Duo CLI, and Custom Flows. These tools address the challenge of AI-assisted coding creating a faster stream of changes than traditional security and review processes can handle. For instance, Dependency Scanning Auto-Remediation automatically opens merge requests to fix vulnerable dependencies, iterating until the build passes. Security Review Flow identifies logic flaws that pattern-matching scanners often miss, such as broken authorization checks and race conditions.

Jul 21, 2026 · 3 min read

tech

NixOS Module for $ORIGIN Support in Shebangs Awaits Kernel Merge

Christian Brauner proposed a Linux kernel patch series enabling PT_INTERP and shebang support for $ORIGIN in relocatable binaries during TacoSprint 2026. This change allows binaries to be dynamically loaded with an interpreter selected via eBPF, enhancing flexibility for NixOS users. The patches will soon land in the -next branch, paving the way for a new NixOS module that registers $ORIGIN support at boot, ensuring seamless integration and backward compatibility.

Jul 21, 2026 · 3 min read

tech

Google Cloud Outage Hits Europe Due to Cooling Failure

Google Cloud experienced a significant outage due to a cooling failure in its europe-west4-a zone, affecting VMware Engine, NetApp Volumes, and Bare Metal Solutions for over 15 hours. The incident underscores the importance of understanding hidden dependencies in cloud services, as some managed services may rely on hardware or infrastructure tied to a single datacenter within a zone. This means that even if you're using multiple zones and regions for resilience, critical services could still be at risk during localized failures. Google's Spanner tool, designed for cross-zone replication, didn't function properly once one building became unavailable, highlighting the need for more robust failover mechanisms.

Jul 21, 2026 · 3 min read

EU Orders Google to Open Android to Rival AI Tools

The European Commission issued binding DMA measures forcing Google to open 11 Android feature groups to rival AI assistants and share search data with competitors. Android changes land by July 2027; search-data sharing starts January 2027.

Jul 21, 2026 · 3 min read

Trump's AI Safety Agency Head Resigns After 3 Months

The head of CAISI, the US AI safety body, has resigned barely three months into the job. The exit leaves federal AI evaluation leadership vacant just as frontier-model access policy is being rewritten in Washington.

Jul 21, 2026 · 3 min read

The daily that does your AI homework.

One email, 7am, free. Reads in 5 minutes.

White House Now Steers Access to Frontier AI Models

The White House is dictating who gets access to frontier AI models, shifting leverage away from OpenAI and Anthropic toward government, per sources. It signals a more hands-on federal role in the AI supply chain.

Jul 21, 2026 · 3 min read

Anthropic Locks In 3.5GW of Google TPU Compute

Anthropic expanded its deal with Google and Broadcom for about 3.5 gigawatts of next-gen TPU capacity coming online from 2027. Most of it sits in the US, extending Anthropic's $50B American compute commitment.

Jul 21, 2026 · 3 min read

Huawei Debuts Atlas 950 AI Cluster to Rival Nvidia

At Shanghai's World AI Conference, Huawei gave the first physical showing of its Atlas 950 SuperPoD, an 8,192-NPU Ascend cluster built for large-model training. It anchors China's pitch for a homegrown alternative to Nvidia systems.

Jul 21, 2026 · 3 min read

Alibaba's Qwen3.8 Max Hits 2.4T Params, Trails Only Fable 5

Alibaba previewed Qwen3.8 Max, a 2.4-trillion-parameter flagship it ranks second only to Anthropic's Claude Fable 5. Shares rose about 5% Monday. The preview is live via Alibaba's Token Plan, Qoder, and QoderWork.

Jul 21, 2026 · 3 min read

Simon Willison Puts Kimi K3 Through the Pelican Benchmark

Simon Willison test-drives Moonshot's Kimi K3, the 2.8-trillion-parameter open-weight model, with his SVG-pelican benchmark. His read: the strongest open model yet, still a notch below Claude Fable 5 and GPT-5.6 Sol.

Jul 21, 2026 · 3 min read

Natural Raises $30M to Build a Payments Rail for AI Agents

Natural raised a $30M Series A led by Forerunner to let AI agents move and store money, taking direct aim at Stripe. The round brings total funding to $40M as agent-to-agent commerce heats up.

Jul 21, 2026 · 3 min read

Page 1 of 397