Updated
Updated
Showing low-relevance items too. Hide low-relevance items
LangChain@LangChainAI score34AILangChain says Managed Deep Agents 0.9 adds a Reactions API for Slack channels, letting developers build loading states and receipts. The developer can build these states using heuristics, decision models, or custom logic.
SemiAnalysis@SemiAnalysis_AI score22AISemiAnalysis says NVIDIA's Rubin delivers 3.2x better profit per gigawatt and up to 10x better performance per dollar than GB300 NVL72 on the vLLM production LLM engine. The post is the first of a three-part thread that will explain the results.
Suno@sunoAI score34AISuno says Studio 2.0 lets users click Add Effect on any audio or MIDI track to build an effect rack. Users can also open the chat bar to create their own plugin.
RadixArk@radixarkAI score28AIRadixArk releases Miles v0.1.2, adding score centering for stable async RL and an experimental native Kubernetes backend. RL jobs can run as ordinary cluster workloads, and orchestration can restart while training continues. The release also adds torchtitan as a third training backend alongside Megatron and FSDP, plus support for DeepSeek-V4.1-Flash and MiMo-V2.6-Flash-RL.
RadixArk@radixarkAI score25AINASCAR says it is using increased automation and AI to expand what it offers fans and speed up its development work. The source gives no further details on specific tools, figures, or timelines.
AIMeta's Muse AI agent has millions of users and can handle tasks from refunds to insurance shopping. Its broad access to personal data has already led to some unnerving mistakes, according to Fast Company.
AIDanu Robotics, an Edinburgh startup founded by Amy Ma, is bringing its H.E.R.O. recycling-sorting robot to market, using a pincer claw instead of suction. The company has letters of interest from two large customers, $500,000 in signed contracts, and more than 200 prospective customers in its sales pipeline. Ma estimates a working site would earn $485,000 in added revenue against a $160,000 initial investment and $24,000 in annual maintenance fees.
AILangChain's Managed Deep Agents v0.9 adds a reactions attribute for Slack channels that accepts either an emoji string or a callable returning one. The article shows a function that returns a bug emoji when a message contains "broken" and eyes otherwise. It also shows a TypeSafe Classifier that picks from a seven-emoji vocabulary and falls back to eyes below 25% confidence.
AIThe Trump administration announced a voluntary agreement with major AI companies calling for internal safety monitoring, external audits, and independent board reviews, and the Federal Trade Commission launched an investigation into OpenAI, Anthropic, and other AI companies over potential consumer risks. OpenAI released Dots, a proactive assistant that retains context, works across applications, and acts without waiting for prompts. Google says Gemini 4 Argon can generate up to a million output tokens in a single response.
🚨 AI News | TestingCatalog@testingcatalogAI score41AIPine AI launched Pine Computer, a cloud computer, harness, and runtime layer built for agentic tasks. On the publisher's SaaS-Bench v1.1, it posts a 78.3% checkpoint score against 74.3% for Opus 5 with Claude Code, but completes fewer whole tasks, 27.4% against 31.1%. Instead of simulating clicks and screenshots, it reads web pages as structured data, and access is through a private beta waitlist.
🚨 AI News | TestingCatalog@testingcatalogAI score62AIAnthropic has expanded dynamic workflows in Claude Managed Agents into a public beta, according to Testing Catalog. Users can configure their agents for multiagent orchestration, with Claude planning and operating a fleet of agents to achieve a goal. The post also links a video from Anthropic's ClaudeDevs account, which the author describes as a new SWE norm.
Santiago@svpinoAI score44AIPine has released a cloud computer service with a built-in AI agent that applications can control through its SDK. Developers give the agent a plain-English task, and it can use a browser, files, and a shell while the app receives notifications and final outputs. Pine's Stanley Wei says the computer is built for AI rather than humans.
François Chollet@fcholletAI score32AIBusabase releases an MIT-licensed open-source database and workspace that lets AI agents write results to a shared, structured space instead of isolated chat histories. The tool turns agent outputs into reusable data, documents, and skills that a whole team can access.
merve@mervenoyannAI score28AIHugging Face launches an arena where users bring their own agent, which gets Nebius GPUs to build RL environments that improve Qwen3.8-27B across eight domains. The arena runs on PostTrainArena from BenchFlow, with compute from Nebius. Setup requires only a few steps through the linked OpenEnv Arena space.
ClaudeDevs@ClaudeDevsPickAI score60AIAnthropic's ClaudeDevs account announces that dynamic workflows for Claude Managed Agents are now available in public beta. The feature is a new type of multiagent orchestration in which a lead agent writes a plan that runs across many agents in phases, then combines their results at the end.
Why it matters: The post describes how a lead agent plans work across many agents in phases and merges their results, a structure useful for understanding complex agent orchestration.
ClaudeDevs@ClaudeDevsAI score46AIAnthropic's ClaudeDevs says users can configure an agent with multiagent type multiagent_20261001 and ask Claude to run a workflow. Claude writes the plan and orchestrates up to 1,000 agents per run.
elvis@omarsar0AI score34AIElvis Saravia says AI models are already smart, but they need better harnesses and environments, with major cost implications. He recommends reading a report on how Pine Computer can help teams, and says he will test it himself and share more later. The quoted post from Stanley Wei argues that real-world AI tasks remain slow, expensive and unreliable because AI runs on computers built for humans, and announces Pine Computer.
LangChain@LangChainAI score16AILangChain says its Interrupt archive is available, with every session watchable on-demand at Context from the quoted post: LangChain simplified agent authentication, memory, and channels, and added web search as a pre-built tool for Managed Deep Agents.
AIFactory says it is acquiring Mentlio (YC S26), a startup building in-harness model routing, context compression, and AI usage measurement to optimize cost and quality. The founders, Ashank Shah and Ahmet Demirbas, will join Factory to bring this work into its autonomous software engineering platform for large enterprises.
Atai Barkai@ataiiamAI score52AIAtai Barkai announces Open Intelligent UI, a template for adding interactive generative UI to any agent-based product. It supports a team's own design system, components, and real data and actions, and is available on Web, Mobile, and Slack, powered by CopilotKit and AG-UI. The code is at
ComfyUI@ComfyUIAI score22AIComfyUI posts a demo of real-time character control using Reactor. The post gives no further details on the method, performance, or availability.
Ai2@allen_aiAI score22AIAi2 published the engineering details of its new GPU scheduler, which allocates compute across research teams. On its largest H100 cluster, median queue wait fell from 5 minutes to 24 seconds.
Ai2@allen_aiAI score22AIAi2 says its old scheduler made every scheduled workload eventually run at HIGH priority. Researchers kept idle jobs running to reserve GPUs for experiments, because the incentives rewarded holding capacity even with no active work.
Ai2@allen_aiAI score13AIAi2 says managers now assign GPU-time budgets to research programs and projects. Its fair-share scheduler compares recent usage with those budgets and moves work from underused allocations ahead in the queue.
Ai2@allen_aiAI score25AIAi2 reports that over a 30-day test of its new scheduler, teams received 98% of the GPU hours they were owed, based on actual demand. Cluster occupancy stayed at 98%, and spare capacity went to interruptible work without drawing down team budgets.
Ai2@allen_aiAI score14AIAi2 has published a blog post on the design decisions behind its scheduler, along with rollout lessons and next priorities. The post links to The text provided does not include further details of the scheduler's features or results.
AITessl Blog describes an agentic code review workflow built for teams whose coding agents produce pull requests faster than humans can review them. The workflow runs review against a written standard in the repository, applies four parallel perspectives covering correctness, security and privacy, scale and resilience, and maintainability, then records each finding, verdict, and response. Tessl Code Review, which the post says is free to start, runs these perspectives as skills, and the team's memory of past decisions is fed back into the standard.
AIAi2's AI Infrastructure team replaced its priority-based scheduler for GPU clusters with GPU time budgets, hierarchical fair-share allocation, and a time-slicing contract. The team says the change moved debates over how much GPU time each research project deserves from case-by-case operational decisions into a transparent budgeting process. The clusters range from 88 to 1024 GPUs across NVIDIA H100, B200, and B300 hardware, and serve about 150 internal researchers.
AIPostman describes the architecture behind Agent Mode, its AI agent for API testing, documentation, discovery, and implementation. The post covers limiting tools per task, using schema-based queries, building purpose-shaped context handlers, and running on Amazon Bedrock with cross-Region inference and prompt caching. Postman reports that tool-selection errors rose once the visible toolset exceeded about 40 tools.
Why it matters: The post shows concrete patterns for tool scoping, context handling, and Bedrock routing and caching, which apply to any team moving an agent past a prototype.
AIAmazon Bedrock Managed Agents, powered by OpenAI, entered public preview, and OpenAI's GPT-6 Astra, GPT-6.1 Sol, and GPT-6.1 Luna became generally available on Amazon Bedrock. AWS also released Strands Decider 2B, a 2B-parameter open source decision model that answers in about 115ms locally, and said the Strands harness uses 28 percent fewer tokens than popular harnesses while matching their accuracy.
AIAi2's AI Infrastructure team replaced its priority-based GPU cluster scheduler with a system using GPU time budgets, hierarchical fair-share allocation, and a time-slicing contract. The team says the change turns decisions about how much GPU time each research project receives into a transparent administrative budgeting process. Its clusters, which range from 88 to 1024 GPUs including H100, B200, and B300 units, serve about 150 researchers facing demand two to three times available capacity.
ElevenLabs@ElevenLabsAI score32AIElevenLabs says it is partnering with Banner Health to answer patient calls with AI voice agents, starting with primary care scheduling. The ElevenAgents system books, reschedules, or cancels appointments directly in Banner's electronic medical record at any hour, and transfers calls to a Banner team member with context when a patient asks for a person.
Rohan Paul@rohanpaul_aiAI score46AISabi has raised a $50M seed round led by Vinod Khosla and Accel to build a baseball cap that turns brain signals into AI prompts. The cap can hold up to 100,000 sensors of 1 to 5 millimeters each, feeding a brain foundation model trained on 100,000 hours of labeled neural recordings. Sabi says the cap can predict a user's next 3-4 keystrokes from brain signals before they type them.
Santiago@svpinoAI score43AISabi has raised $50 million from Khosla Ventures, Accel, Initialized, Kevin Weil, and DST Global to build a wearable brain-computer interface. The company says its TSMC-fabricated chip reads brain signals without touching the scalp, with custom sensors collecting data and its own AI model decoding it into text. The device is a cap rather than an implant.
OpenAI Developers@OpenAIDevsAI score46AIOpenAI says Codex on Windows now has a new sandbox mode built on Microsoft's Execution Containers (MXC), offering faster setup, stronger network enforcement, and granular file access controls. The mode requires a compatible Windows 11 device. Background from Microsoft's announcement says MXC is now generally available on Windows 11, keeping agents within boundaries the operating system enforces.
OpenAI Developers@OpenAIDevsAI score22AIBaseten explains how to decide which layers of a model can run in 4-bit NVFP4 without losing needed information. The post compares architecture-based heuristics, isolated-layer sensitivity scoring, and SaturationQuant, which accounts for other quantized layers. It also covers calibration with representative data and block-level scales of 16 values.
Why it matters: The post explains how to choose which layers run at NVFP4 precision using heuristics, sensitivity scoring, and saturation-aware scoring, with clear calibration steps.
elvis@omarsar0AI score38AISabi has raised $50M from Khosla Ventures, Accel, Initialized, Kevin Weil, and DST Global to build a wearable brain-computer interface cap. The cap carries 100,000 sensors and aims to let people talk to AI agents by thinking. The post's author says improved hardware could unlock new research and personal agentic applications.
Databricks@databricksAI score25AIDatabricks has published a reference implementation pairing Temporal with Lakebase Postgres so cloud agents can survive worker, container, or deployment replacement. The design keeps recorded work and evidence and review state queryable, and lets human decisions arrive days later. Unity Catalog remains the governed policy source through synced tables.
