#Agent

AIGoodfire Research describes probe-based cyber monitors for Kimi K3 and GLM 5.3 deployed on a production inference stack. The probe filters suspicious exchanges before an LLM judge reviews them, reaching about 93% recall at a 5.5% benign-session interruption rate at roughly 50x lower judge cost. In FAR.AI's red-teaming, the monitor reduced universal jailbreaks to zero across 140 tested strategies.

AI3D is still impressive with GPT-6 Astra. I had it read my Memory and, based on my preferences, make a poster with several models built in Blender. The model details, materials, and lighting all come out very well.

AIAnthropic says the fastest way to use its new credits is through Claude Managed Agents. Developers can build an agent with the claude-api skill in Claude Code by running /claude-api managed-agents-onboard.

AIOracle is using ChatGPT Work and Codex to turn specialist knowledge into fast, repeatable workflows across recruiting, engineering, and operations. The source does not provide figures, timelines, or specific results beyond the headline's claim that days of work can take minutes.

AIMeta's Muse AI agent has millions of users and can handle tasks such as refunds and insurance shopping. Its broad access to personal data is already producing unnerving mistakes.

AIStepFun's flagship model, Step 5 Preview, is available free for one week through Nous Portal, the platform from NousResearch. The post invites users to try it with Hermes Agent and share what they build.

AIAnthropic released Claude Haiku 5.5, raising its OSWorld 2.1 score from 15.7% to 72.4% and supporting a 1 million token context window. The article notes Haiku 5.5 still scores 39.2% on Terminal-Bench 4.0 against Sonnet 5.5's 70.6%, and that prompts above 100,000 tokens are priced higher, so migration costs need to be measured on real workloads.

AICogent Attack Path Analysis uses Cogent's own cyber models to find the longer attack routes an AI agent swarm could take into a company. It then identifies the single fix that closes each route, since agent swarms persist where human attackers give up.

AIMicrosoft has released MxC, an open-source sandboxing library that runs on Windows, macOS, and Linux. It builds on processcontainer, bubblewrap, and seatbelt as its underlying mechanisms. The project is aimed at developers struggling with fragmented sandbox tools for AI agents.

AISatya Nadella outlines Microsoft's vision of Copilot as a new operating system for work spanning every model and task, backed by a governed "headless" business layer. Microsoft announced over 30 new Copilot skills across Dynamics 365 Sales, Service and Customer Insights, plus Microsoft Copilot Managed Runtime for hosting code inside a company's IT-governed environment. Nadella describes this as an "infinite SaaS factory" where users can describe needs and build customizations connected to existing systems of record.

AIAugment Code is selling select assets, including Cosmos, Auggie CLI, and the Code Context Engine, to Harness, and the product team is moving to Harness. The company says Harness's integrated platform delivers these capabilities to customers more effectively than building them independently. Harness describes itself as building the Autonomous SDLC Platform for shipping AI-written code across enterprises.

Why it matters: The announcement shows how a coding AI company is folding its products into a larger software delivery platform, a shift that shapes how enterprise teams will buy these tools.

AIA security platform uses agents to map a system's potential vulnerabilities and identify routes an attacker could take to reach sensitive data. It then recommends changes to close those paths. The quoted post cites a 700-agent swarm that breached Hugging Face with over 17,000 actions, and presents this tool, Cogent Attack Path Analysis, as the defensive counterpart.

AIThe post promotes a curated collection of recursive self-improvement (RSI) papers covering the field's history and latest techniques. The author describes RSI as a hot, still-early AI topic and links to the collection at academy.dair.ai.

AIThe author says GoodCase's case clustering, similar-case recommendations, mobile layout, and multi-country web acceleration, including how storage is split across Vercel, Cloudflare, and Supabase, were all rewritten with Opus 5.5. The post also notes Claude's usage quota has held up well for this work, with a longer write-up planned.

AIAmazon unveiled 8-inch, 11-inch and 12-inch Alexa tablets priced from $230 to $550 and said it is ditching its budget Fire lineup. The devices run Android rather than Fire OS, and preorders open Thursday with shipping starting Oct. 14. Amazon said it will support the Fire lineup for four years after the final shipment but is no longer manufacturing new units.

AIStepFun is offering its flagship Step 5 Preview model free in OpenCode for one week, targeting agentic and professional work. The preview supports a 1M context window, multimodal input, and zero data retention.

AIA Kyndryl survey of 2,000 senior IT decision-makers found only 10% are applying agentic AI as a modernization tool, and nearly half report being behind schedule with cost overruns. Researchers say agentic AI shows early promise for mapping hidden dependencies, generating code, and creating documentation, while Andy Thurai, a former IBM chief strategist, warns that AI-driven infrastructure sprawl could make compute costs unpredictable.

AIThe author argues that agents working across a company fail because they lack the decisions and context recorded in threads, meetings, and DMs, not because the model is weak. The approach stores distilled claims with source evidence and time, never overwrites facts, labels missing information explicitly, and resolves permissions before the model runs. The report cites results on LongMemEval, including 99.8% top-ten evidence recall and $8.24 ingestion cost, and says an open-weight model can match frontier extraction quality.

AIGoogle is launching a "universal" Gemini agent that works across apps and devices in the background, available in private preview to enterprise customers. Users can chat with it and assign tasks from the Gemini Enterprise app, and it works inside Gmail, Drive, Docs, Sheets, and Calendar as well as third-party apps like Slack and Microsoft 365. The source notes it runs in the cloud, keeps the same context across devices, and can use job-specific sub-agents.

AINous Research says Hermes Agent is now available on the new ASUS ProArt RTX Spark PCs, with MuseTree and ComfyUI integrations and local model support on up to 128 GB of unified memory. The post presents it as a creative stack that runs entirely on the user's own machine.

AIGoogle has announced a single, universal Gemini agent for Gemini Enterprise as part of its Gemini at Work updates. The agent answers questions, handles knowledge work, creates images and media, and writes and runs code. It works inline in Gmail, Drive, Docs, Slides, Sheets, Chat, and Calendar, with new data and analytics skills for plain-language insights and industry-specific tools for financial services and legal teams.

AIReplit has announced an integration with Databricks that lets teams build business apps and deploy them on live, governed enterprise data. A four-step setup guide covers Databricks admins, Replit org admins, and end users, aiming to take an app from zero to governed production in 15 minutes.

AITeamily AI now lets users add people and AI agents to the same group chat, so everyone works from shared context. The example shows a branding change handled by research, writing, and website-building agents, with a designer's feedback incorporated and the finished page shared in one continuous conversation. The platform's 2.0 release, which the post describes as opening to everyone, adds real-time human–agent collaboration and multi-model routing.

AIAI-native tools are removing the data bottleneck that long constrained CFOs, shifting the role toward designing the operating systems that turn data into decisions. Finance teams are adopting AI-native software for ERP, forecasting, procurement, and audit, and "finance engineers" are building custom automations and agents. OpenAI's CFO Sarah Friar describes finance moving toward a zero-day close and continuously updated forecasts.

AIGrok bot works great with foldable phones. Chat on the left, publish directly on the right.

AIMeta's Muse personal AI agent, launched on September 8, surpassed 2.5 million downloads in its first two weeks and topped free-app rankings on Apple and Google's US app stores. Its popularity has drawn attention to the emerging market for agent harnesses, where China's biggest internet companies are already competing for position. The excerpt does not provide full details on their specific products.

AIMeta's Muse and OpenAI's Dots are always-on AI agents with animated mascots, pitched to consumers and businesses for tasks like restaurant reservations and inbox triage. Muse is free, while Dots is not, and OpenAI also offers "specialist" Dots for marketing, legal work, and accounting. The discussion centers on privacy and security concerns about giving agents access to credit card details and email.

AIAllie K. Miller describes a workflow where she had Instinct write a large prompt for Codex to execute her 19-item task list, then emailed that prompt to herself. She also had Instinct write the prompt telling Codex to search her email and find the note Instinct left for it.

AIOpenRouter announces that StepFun's Step 5 Preview is live on its platform, priced at $1.00 per million input tokens and $2.70 per million output tokens. Cache hits are 50% off at launch, bringing them to $0.05 per million. A week of free access is rolling out across partners including opencode, Cline, Nous Research, and Kilo Code.

AIStepFun's Step 5 Preview outperforms Kimi K3 and GLM-5.3 on six of eight benchmarks in its own evaluation. Reported scores include 67.7 on DeepSWE, 80.5 on ProgramBench, 49.0 on StepCodeBench, and 66.4 on FrontierFinance.

AIStepFun's Step 5 Preview is now live on OpenRouter as the company's new flagship for agentic work. It uses a sparse MoE design with 27B active and 600B total parameters, a 1M context window, and accepts text, image, and video input. The post highlights strength in coding and professional knowledge work, especially finance.

AIReJev, an independent community project, applied LoRA post-training to OpenBMB's MiniCPM5-2B for bounded agent decisions: state, question, and candidate options yield one choice. On its sealed 1,892-sample holdout, accuracy rose from 51.11% to 80.50% (+29.39 percentage points) with 0% invalid outputs, at about $5.31 in cumulative Modal billing including earlier experimental overhead. The authors describe this as an early, task-specific result, not parity with Jev.

AISoftware developer Gergely Orosz questions why he is working more despite AI tools, quoting Sam Newman's view that AI was meant to free developers from drudgery. Newman says most developers are doing more work, with more context switching and a loss of the big picture. The quoted post adds that AI assistants are not human partners and that pairing with them fragments the shared mental model of a program.

AIHugging Face CEO Clément Delangue says the field urgently needs more public traces of AI agents attacking and defending systems, since defenders cannot learn from activity they cannot see. He invites anyone holding such traces who faces pressure to keep them private to contact him directly.

AITencent has added a new feature to its WorkBuddy: you can select files in different formats to open and preview them directly, plus AI chat for editing on the spot. For example, if I make a Markdown file, I can use AI to translate it directly, overwriting the original content.

AIStepFun's Step 5 Preview is now live on OpenRouter, with a week of free access rolling out across opencode, Cline, NousResearch, and Kilo Code. The company positions it as flagship-tier intelligence for agentic and professional work at substantially lower task cost, letting users keep their workflow while switching models.

AIZenity Labs researchers say a single publicly accessible agent on Amazon Bedrock AgentCore was enough to take over every AgentCore agent in the same AWS account and region. Using one chat prompt, the researchers got the agent to query the internal metadata service and send its AWS credentials to an external server, exposing private conversations, source code, and stored credentials. Zenity says AWS made IMDSv2 the default for new deployments and changed the default execution role around August.

AIReplit and Databricks integration, now generally available with native Lakebase support, lets enterprise teams build apps from plain-language prompts using Replit Agent and deploy them as Databricks Apps. Deployed apps inherit automatic user authentication and Unity Catalog access controls, and Replit Agent auto-provisions a managed Lakebase Postgres database for operational data. Lakebase keeps app-written data inside the Databricks perimeter instead of a separate external database.

AIStepFun has released Step 5 Preview, its flagship model for agentic work, built on a sparse Mixture-of-Experts architecture with 27B active and 600B total parameters. The source says it performs strongly in software engineering and professional tasks, but the feed supplied only an excerpt, so benchmark details are not available here.