NerdNewsJuly 10, 2026 |
|
News & Trends
SpaceXAI launches Grok 4.5
SpaceXAI has launched Grok 4.5, its first AI model built with Cursor's help, which excels at coding, agentic tasks, and knowledge work. It was trained on NVIDIA GPUs and outperforms other leading models at real engineering tasks. Grok 4.5 is now the default model for Grok Build and is available at $2 per million input tokens and $6 per million output tokens.
OpenAI says GPT 5.6 is the 'preferred model' for Microsoft Copilot 365
OpenAI announces GPT 5.6 as the preferred model for Microsoft's 365 Copilot, amid rumors of a breakup between the two companies. GPT 5.6 will support Microsoft users across productivity apps, including Word, Excel, and PowerPoint.
OpenAI launches GPT-5.6
OpenAI launched its new family of models, GPT-5.6, with three variants: Sol, Terra, and Luna. The models promise powerful capabilities in enterprise work, coding, and scientific research, with improved efficiency and cost-effectiveness. GPT-5.6 also supports defensive cybersecurity activities and is available across ChatGPT, Codex, and the OpenAI API.
Meta Enters AI Coding Battle with Muse Spark 1.1
Meta has launched Muse Spark 1.1, a multimodal AI model designed for agentic coding, competing with OpenAI and Anthropic. The model can handle complex processes, manage digital workflows, and deploy new features in enterprise systems. Meta's pricing is competitive, charging $1.25 per million input tokens and $4.25 per million output tokens.
New York Times says OpenAI hid evidence in ChatGPT copyright trial
The New York Times and The Daily News claim OpenAI hid evidence in a copyright trial related to ChatGPT. OpenAI allegedly had the ability to search its training corpus and had collected data on copyrighted works, contradicting its previous statements. The plaintiffs are asking the court to discipline OpenAI for allegedly withholding evidence. |
|
Options & Tutorials
Accelerating Netflix Data: A Cross-Team Journey from Offline to Online
Netflix transformed its data movement from offline to online systems, reducing deployment time by 90% and costs by 70%. They achieved this through a cross-team effort, introducing a repeatable capture, conversion, and deployment framework, and transforming their key-value abstraction from stateless to stateful.
OpenAI Fixes 18-Year-Old GNU libunwind Bug
OpenAI engineers fixed an 18-year-old GNU libunwind bug by using epidemiological debugging, analyzing core dumps to identify patterns. The bug caused crashes in Rockset, a C++ data infrastructure service. The team found two unrelated bugs, one due to a race condition in libunwind and another caused by a faulty CPU. The fix has been upstreamed to GNU libunwind.
AlloyDB AI Proxy Models
AlloyDB introduces AI functions and proxy models to improve performance and reduce costs. The proxy model architecture allows the database to learn from a sample of data and apply that judgment locally, reducing the need for external model calls. This approach has implications for the future of database and AI interactions.
AWS Details How One Customer Scaled to One Million Lambda Functions
AWS has outlined how ProGlove scaled its SaaS platform to run over 1 million AWS Lambda functions. The company used a one-account-per-tenant model, automation with CloudFormation StackSets, and scale-to-zero policies to keep costs low. ProGlove also implemented jittered execution windows and event-driven triggers to smooth regional load and avoid self-inflicted DDoS attacks.
Kubernetes Approach to AI-Assisted Maintainership
The Kubernetes community has introduced a framework for integrating AI into open-source maintainership, prioritizing human accountability. AI tools will assist with repetitive tasks, but human maintainers will retain final authority over contributions, ensuring code quality and security. |
|
Launches & Tools
GitGuardian Power for Amazon Kiro
GitGuardian's Power for Amazon Kiro scans code for exposed secrets, providing remediation guidance and honeytoken-based threat detection. It integrates with Kiro's agent, enabling proactive secret detection across multiple surfaces, including files, Git history, and Docker images.
The AI Voice Assistant Arms Race Is Far From Over
Amazon is working on a new project called Moonraker to improve its voice assistant, Alexa, with more agentic capabilities, aiming to handle more sophisticated multistep commands, with a projected cost of over $100 million in GPU costs in 2026.
Meta's new AI model Muse Spark 1.1
Meta has released Muse Spark 1.1, a new AI model that can plug into AI coding software with the Meta Model API. The model is available to US developers through a public API preview and offers improvements such as advanced coding, bug detection, and multimodal perception. It's part of Meta's efforts to compete with other AI companies like OpenAI and Google.
OpenAI releases ChatGPT Work tool
OpenAI has released ChatGPT Work, a general-purpose productivity agent that integrates its GPT-5.6 models, Codex coding agent, and Atlas web browser. The tool allows for task scheduling, automation of busywork, and connection to favorite apps and tools through a unified plugin directory. It is available on Mac, Windows, and web, with a rollout expected to complete within 24 hours.
Meta's New AI Chips
Meta will begin producing its new AI chips in September, aiming to reduce GPU costs and increase computing capacity for its AI efforts, with a modular design approach and partnerships with Broadcom and TSMC. |
|
Quick Links
What Happened to the Frontend
The article explores the transformation of frontend development over two decades, from simple HTML files to complex builds and frameworks, and how the industry is now moving back towards simplicity with server-side rendering and minimal JavaScript.
NHTSA calls out autonomous cars for interfering with first responders
The US National Highway Traffic Safety Administration (NHTSA) is calling on autonomous vehicle makers to address the issue of their cars interfering with first responders. The agency has identified a pattern of autonomous vehicles getting in the way of ambulances, fire trucks, and other emergency responders, and is demanding a solution by the end of July. This issue has been reported in several instances, including a Waymo vehicle blocking an ambulance responding to a deadly shooting in Texas.
AI Agent Startup Raises $100M
Lyzr, an AI agent startup, used its own AI agent to raise $100 million in Series B funding, proving the product's effectiveness and demonstrating the ease of fundraising with AI traction.
Gradium Raises $100M Seed
Gradium, a Paris-based AI voice startup, has raised $100M in seed funding, backed by Nvidia. The company is using the funds to open a Bay Area office and compete for talent, strengthening its position in the AI ecosystem. Gradium is working on audio models that deliver voice at scale with ultra-low latency.
Humanoid Robots Controlled by Surgeons
Humanoid robots controlled by surgeons have successfully performed a world-first operation on live pigs, removing gallbladders in a preclinical trial. The robots, nicknamed 'Surgie', were teleoperated by skilled human surgeons and may offer a cheaper and more space-efficient alternative to specialized surgical robots. |
Share NerdNewsShare your affiliate link to get commission!
Thanks for reading,
If you dont want to receive future editions of NerdNews, |