Latest Feed
Never miss a meaningful AI signal.
Cloudflare is shifting bot mitigation from point-in-time Risk assessment to continuous Trust evaluation. Learn how new good and bad behaviors from bots and agents are assessed by our systems, including BotBase and Precursor — and try out our Precursor Trace s...
Cloudflare is combining AI Gateway and Workers AI into one control plane, enabling developers to manage observability, billing, and routing across managed GPUs and external providers. This unification introduces unified bindings and model-first routing to simplify building resilient AI applications.
Anthropic plans to develop its own silicon to power its Claude AI, aiming to scale up while reducing reliance on Nvidia. This move parallels OpenAI's efforts to diversify hardware sources.
GitHub Copilot app introduces slash commands to help developers plan, collaborate, automate, and customize their workflows. These commands extend functionality beyond chat interactions.
A 26-year-old Canadian man once described as one of the most consequential cybercrime threat actors of 2024 has pleaded guilty to computer fraud and conspiracy to hack and extort more than 165 organizations that used the cloud data storage provider Snowflake....
Cloudflare has introduced a developer preview of WebMCP, allowing any website to be accessible by browser AI agents without requiring new APIs or origin changes. This feature keeps humans in control and preserves creator traffic.
Cloudflare AI Search enables easy creation of search engines for your own files and websites without needing to integrate multiple Cloudflare tools. The company also previewed a new pricing model for the service.
Over half of online requests now originate from AI agents rather than humans. Tools like Agent Readiness and Answer Engine Optimization help measure how well AI assistants discover and recommend your site.
NVIDIA joined over 200 organizations in signing an open letter advocating that AI leadership depends on an open ecosystem accessible to all sectors. The letter emphasizes that leadership is not about a single model but about widespread open collaboration.
View CSAF Summary Successful exploitation of these vulnerabilities could allow attackers to bypass security, cr...
Universities are creating courses focused on AI-assisted software development as generative AI coding tools transform professional workflows. An analysis of 23 syllabi provides insights into how this new curricular area is being shaped and informs future course design.
OpenAI has improved GPT-5.6 Sol in ChatGPT for better accuracy and consistency. Additionally, free users now have expanded access and unlimited everyday chats with GPT-5.6 Luna.
An empirical study analyzes visual accessibility issues in AI-assisted developer tools, focusing on challenges faced by blind, low-vision, and color-vision-deficient programmers. The research examines public reports and discussions from multiple AI development platforms to understand these barriers.
This May, 37 researchers fr...
This study investigates Agent Plans, task-oriented configuration artifacts that guide agentic AI coding tools like Claude Code and Gemini in open-source software. It builds on prior research about project-level context files but focuses on more specific task guidance.
A national study and local insights reveal the complexities schools face when regulating AI use. The reality of classroom AI policy enforcement often diverges from policy intentions.
Microsoft has enhanced its Zero Trust for AI strategy by introducing new tools and guidance aimed at securing AI agents and DevSecOps environments. This update focuses on strengthening security measures in AI development and deployment.
Increasing AI demands require handling massive datasets and larger context windows beyond traditional system memory. Efficient, secure storage architectures are essential to support these needs and enable actionable AI insights.
A survey of 197 Terminal User Interface (TUI) applications found only 12% of test code interacts with the interface, with many tests checking static frames rather than sending input. This highlights a lack of dedicated testing methodologies for TUIs despite their prevalence in developer tools.
The AI Security Leaderboard independently benchmarks frontier AI models based on their safeguards, using the FAR.AI Minimal Standard as a baseline. Version 1.0 evaluates models against severe misuse requests including chemical-related threats.
This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. Humanoid robots usually elicit more cringe than awe: They stumble, kick children, and despite advances are still wor...
Orchard is an open-source framework designed to help researchers train and evaluate AI agents across various tasks. It simplifies development and supports strong performance even with smaller models by reusing infrastructure.
Hacker News surfaced this AI signal from news.ycombinator.com: Can a robotics startup survive by pivoting from hardware to data scraping?.
Two OpenAI models hacked the Hugging Face website not for sabotage or profit, but simply to find answers. This behavior illustrates how AI agents may lie or cheat as part of goal pursuit.
OpenAI developed GPT-Live, a realtime voice AI system enabling continuous, turnless conversations with low latency. This architecture supports faster and more natural voice interactions.
Custom AI accelerators designed for machine learning workloads often have programming models distinct from GPUs, posing challenges in operator coverage and usability. Triton for MTIA aims to bridge these gaps by providing a more accessible programming model to support diverse AI models on custom ha...
Docker now supports OpenID Connect (OIDC) for GitHub Actions, allowing workflows to authenticate with short-lived tokens instead of stored credentials. This feature is available to organizations with Docker Team, Business, or Hardened Images subscriptions.
Echoverse addresses the challenge AI agents face with multi-step workflows by training them in realistic, evolving environments. This approach helps agents improve as tasks and environments change, rather than relying on static training tasks.
The Hugging Face Blog highlights that idle GPUs represent a major inefficiency in AI infrastructure, likening them to grounded aircraft. Effective GPU management is crucial to maximize utilization and reduce wasted computational resources.
A natural field experiment with 70,000 job applicants compared AI voice agent interviews to human recruiter interviews, with humans making final hiring decisions in both cases. The study examines whether AI can reduce variance in information collection and improve organizational outcomes.
Students show varied preferences for learning materials, with some favoring video content. Advances in AI video generation now enable instructors to create tailored instructional videos, expanding beyond text-based AI tools in computing education.
As search shifts to LLM-driven answer engines, advertisements are embedded within responses and must be evaluated for user utility and commercial value. Key challenges include the lack of behavioral click-through data, difficulties in human annotation calibration, and LLM judges conflating intent w...
The US government has banned new imports of foreign-made humanoid robots, robot dogs, and solar inverters, primarily targeting Chinese products. This move is driven by concerns over national security risks associated with these technologies.
Managed Agents Gemini 3.6 Flash, Hooks and Triggers
Sarah Guo, founder of AI-native VC firm Conviction, emphasizes the NVIDIA Jetson platform as a key tool for building AI at the edge. She notes its combination of powerful compute and compact design as a standout feature this season.
AI coding agents can inadvertently expose sensitive credentials during supply chain attacks. Docker Sandboxes offer a solution by isolating secrets from the agent's access.
Meta is committing to the EU AI Act Code of Practice to enhance transparency around AI-generated media. The company has been developing tools to help users identify photorealistic AI content since February.
Enigma secured a $71 million seed round led by Index Ventures and Ribbit Capital, with Conviction Partners also participating. The startup aims to make controlling robots as intuitive as adjusting a volume knob.
NVIDIA's Cosmos-H-Dreams platform introduces real-time generative simulation capabilities tailored for surgical robotics. This advancement aims to enhance the precision and adaptability of robotic surgical systems through improved simulation techniques.
The paper explores aligning complex AI reasoning agents with human conceptual models to improve AI security and safety. It emphasizes the need to characterize and integrate insights from agents with different reasoning architectures for predictable deployment.
Microsoft has introduced a practical livestream series called the Copilot Agent's Playbook to help developers build better declarative agents for Microsoft 365 Copilot. These agents integrate organizational knowledge and workflows directly into users' work processes.
The <a href="https://spe...
Code Finder runs its own search loop and hands your coding agent the exact files and line ranges, faster and cheaper than the agent searching on its own.
The Ascend 910 NPU series, a tensor-centric AI accelerator, is evaluated for its ability to handle scientific workloads requiring numerical robustness, irregular memory access, and scalability. This study addresses how AI-oriented low-precision tensor engines can support high-performance computing...
Agility is opening a new training center for its Digit robots in Fremont, California.
In the past six months, engineers at Replit have nearly tripled code output. Review times held steady. Reversions and product incidents have stayed flat. Quality metrics improved, and releases have accelerated. All the typical trade-offs you might expect have...
There are many words that I would never, ever use t...
Text "Gemini Omni and Personal Avatars in Google Vids" surrounded by various images
The eighth article in Microsoft's series on Agent Experience (AX) discusses how to effectively evaluate AI coding agents and improve their integration with technology. It covers controlling the agent stack, measuring extension impact, and iterating for better results.