Agents
Google Cloud's Open Knowledge Format turns scattered docs into Markdown files for AI agents
Google Cloud just launched the Open Knowledge Format, which converts scattered documents into Markdown files for AI agents. This makes it easier for AI to access and process information from various sources efficiently.
AI coding agents find the right file but miss the exact lines that matter, study shows
AI coding agents are successfully locating the right files but often overlook the specific lines of code that are crucial. This gap means developers might still need to manually check the outputs for accuracy.
Introducing Omnigent: A Meta-Harness to Combine, Control and Share Your Agents
Databricks just launched Omnigent, a new meta-harness for managing multiple AI agents. Users can now combine, control, and share their agents more efficiently, streamlining workflows across different tasks.
How we made GitHub Copilot CLI more selective about delegation
GitHub just improved Copilot CLI's delegation process to be more selective. This change helps users get more relevant suggestions and reduces unnecessary outputs during coding tasks.
NVIDIA Blackwell Leads on First Agentic AI Infrastructure Benchmark
NVIDIA just launched Blackwell, the first benchmark for agentic AI infrastructure. This sets a new standard for evaluating how well AI systems can operate autonomously and complete complex tasks.
Building Supercharger: How Rocket Close optimized title operations with agentic AI
Rocket Close optimized title operations using agentic AI to streamline processes. This enhancement allows for faster and more efficient title management, improving overall workflow efficiency.
Jeff Bezos’s Prometheus raises $12B to build an ‘artificial general engineer’ for the physical world
Jeff Bezos’s Prometheus just raised $12 billion to develop an 'artificial general engineer' focused on the physical world. This funding aims to advance AI capabilities in real-world engineering tasks, potentially transforming industries reliant on physical labor.
Claude Fable is relentlessly proactive
Claude Fable is taking a proactive approach to task management by autonomously generating and executing plans. This means users can expect more efficient workflows without constant oversight.
How an astrophysicist uses Codex to help simulate black holes
An astrophysicist is using OpenAI's Codex to simulate black holes. This integration helps streamline complex calculations and enhances research efficiency in astrophysics.
datasette-agent 0.2a0
Simon Willison just released datasette-agent 0.2a0, an update that enhances the capabilities of the Datasette framework. This version allows users to build more sophisticated data-driven applications by integrating AI agents for querying and processing data autonomously.
For Robotaxis, Safety Must Be Built In, Not Bolted On
NVIDIA is emphasizing that safety features for robotaxis need to be integrated from the start rather than added later. This approach aims to enhance the reliability and trustworthiness of autonomous vehicles for users.
If Claude Fable stops helping you, you'll never know
Claude Fable is changing its approach to user assistance, potentially leaving users unaware when it stops providing help. This shift could impact how users interact with the system and rely on its support.
Setting a custom price for a model in AgentsView
AgentsView now lets users set custom prices for their AI models. This gives developers more control over monetization strategies and can enhance their revenue potential.
Scale Robot Reinforcement Learning with NVIDIA Isaac Lab on Amazon SageMaker AI
NVIDIA just launched Isaac Lab for scaling robot reinforcement learning on Amazon SageMaker. This integration allows developers to train and deploy robotic AI models more efficiently in the cloud.
Can Voice Agents Handle Bilingual Customers? Benchmarking Frontier ASR on Code-Switched Speech
Hugging Face is benchmarking ASR systems for handling bilingual customers with code-switched speech. This means voice agents can improve their understanding and responsiveness in multilingual environments.
Claude Fable 5 is now available on Databricks, fully governed through Unity AI Gateway
Claude Fable 5 just launched on Databricks, fully governed through Unity AI Gateway. This integration enhances data governance and security for users leveraging AI in their workflows.
Transforming solar and wind maintenance reports with Genie and AI agents
Databricks is using Genie and AI agents to automate the analysis of solar and wind maintenance reports. This streamlines the reporting process, making it faster and more efficient for energy companies.
datasette-agent-edit 0.1a0
Simon Willison just released datasette-agent-edit 0.1a0. This tool allows users to edit Datasette data using an AI agent, streamlining data management tasks.
OpenAI says "chat is dead" and plans to rebuild ChatGPT as a full-blown agent app
OpenAI is rebuilding ChatGPT into a full-blown agent app, declaring that traditional chat is outdated. This shift means users will soon interact with a more capable AI that can handle complex tasks autonomously.
Perplexity's "Search as Code" lets AI models write their own search pipelines instead of calling fixed APIs
Perplexity just launched 'Search as Code,' allowing AI models to create their own search pipelines instead of relying on fixed APIs. This gives developers more flexibility and control over how search functionalities are implemented in their applications.
Five labs, five minds: building a multi-model finance drama on small models
Hugging Face is collaborating with five labs to create a multi-model finance drama using small AI models. This project aims to showcase how diverse models can work together in a cohesive narrative, enhancing user engagement in financial storytelling.
Sakana AI bets AI that improves itself can break the compute arms race of frontier labs
Sakana AI is developing self-improving AI systems to tackle the compute arms race in AI labs. This approach could lead to more efficient models that require less computational power while enhancing performance.
Meta's Hatch AI agent could cost up to $200 a month and marks its first paid AI product
Meta just launched its Hatch AI agent, which could cost users up to $200 a month. This marks Meta's first paid AI product, indicating a shift towards monetizing AI services.
Qwen3.7-Plus is Alibaba's bid to turn multimodal AI into a full-blown autonomous agent
Alibaba just launched Qwen3.7-Plus, aiming to transform multimodal AI into a fully autonomous agent. This move positions Alibaba to compete more aggressively in the AI landscape by enabling more complex, independent task execution.
Thousand Token Wood: shipping a multi-agent economy on a 3B model
Hugging Face just launched Thousand Token Wood, a multi-agent economy built on a 3B model. This setup allows multiple AI agents to interact and collaborate, enhancing their ability to perform complex tasks together.
Satya Nadella publicly torches a VP's plan to make Microsoft's AI agent deliberately addictive
Satya Nadella just criticized a VP's plan to make Microsoft's AI agent intentionally addictive. This pushback signals a focus on ethical AI design over engagement at all costs.
Quoting Andreas Kling
Andreas Kling is launching a new AI tool that automates coding tasks for developers. This tool aims to streamline workflows and reduce repetitive coding efforts.
Apple approves Poke as the first AI agent on its Messages for Business platform
Apple just approved Poke as the first AI agent on its Messages for Business platform. This allows businesses to automate customer interactions through AI, streamlining communication and improving response times.
Clay MCP is now available in Codex - The GTM with Clay Blog
Clay just launched MCP integration in Codex. This allows developers to create more complex workflows by leveraging multi-agent capabilities directly within their applications.
Industrial Software Leaders Build Secure, Autonomous AI Engineers With NVIDIA NemoClaw
NVIDIA just launched NemoClaw, a platform for creating secure, autonomous AI engineers. This tool enables industrial software leaders to develop AI that can operate independently in engineering tasks, enhancing productivity and security.