#Agent

AIJetBrains released Mellum2.1, a 12B mixture-of-experts model with 2.5B active parameters under the Apache 2.0 license, built for coding agents. Post-training shifted to reinforcement learning across thousands of environments and millions of sandboxed runs, and the model is available on Hugging Face. The source reports gains over Mellum2 on LiveCodeBench, AIME, GPQA Diamond, BFCL v4, IFEval, and SWE-bench Verified, and says it serves almost twice the tokens of Qwen3.5-9B under heavy load.

Why it matters: The post shows how reinforcement learning in real sandboxed environments changed a compact open model's repository work, with benchmark gains against Mellum2 and two peers.

AIGoogle Cloud announced the Gemini agent, a single agent that answers questions, handles knowledge work, creates media, and writes and runs code from one prompt box. It runs in the cloud with persistent memory, uses multi-agent orchestration, and adds Workspace integration, domain skills for data and industries, identity-based governance through Agent Gateway, and spend caps. The source also cites customer deployments and says nearly 80% of Google Cloud customers use its AI products.

Why it matters: The announcement shows how a single work agent spans chat, Workspace, data analysis, governance, and cost controls, useful for judging enterprise agent deployment scope.

AIThe author suggests installing a coding agent such as Pi or DeepSeek Harness in a cloud computer and allocating unused Code plan quota to it. Grok Bot could then use those tokens to write code and render videos, so leftover quota is not wasted.

AIBy the way, could Grok Bot's model already be Opus 5.5? It works remarkably well on complex tasks.

AITencent Cloud has open-sourced Octop, a self-hosted multi-agent AI assistant platform aimed at families and small teams, with multi-user accounts and data kept on the user's own machine. The full text describes it as a single Python process that bundles the backend, web dashboard, CLI, IM gateway, cron jobs, and multi-agent runtime, with state rebuilt from SQLite on restart.

AIGoogle's Antigravity agent has been renamed "Chief of stuff" in its latest update, which the poster reads as a promotion. The poster wonders whether Antigravity could become a home for Google's own agents, and background notes that Google is prototyping a voice agent internally called "Concierge," which appears to be a very early version.

AILangChain has added three engineering upgrades to Skills in its Deep Agents framework: tool-binding Skills, pinned Skills, and mid-thread reloading. Tool-binding lets a SKILL.md declare tools via metadata.include_tools, so tools are injected only when the Skill is read, and pinned Skills inject full instructions before the next model call, skipping a round trip. Setting skills_metadata to None rescans the Skills library mid-thread without restarting, at the cost of invalidating the cache.

AINVIDIA AI highlights how H-Company optimizes vision-language model (VLM) serving for computer use agents using NVIDIA Dynamo. The post is a broadcast announcement with a link, and no performance figures or technical details are given in the text.

AIMuse, Dots, or other agents with cloud computers can all be used.

AIThe author set up a Grok bot to produce a daily morning AI news video on a schedule, running content collection, code writing, and video rendering entirely on Grok's cloud virtual machine without local computers. The author says the results are quite good and shares the full prompt so others can run the same workflow with their own Grok bot.

AIAnthropic released Claude Haiku 5.5, its fastest and most affordable small model, at prices up to 90 percent lower for most prompts under 100,000 tokens. Artificial Analysis ranks it first among small-class models on its Intelligence Index with a score of 43, but it consumes about three times the output tokens per task that GPT-6 Luna needs at maximum effort.

AILenovo's YOGA Pro 15 RTX Spark laptop, powered by the NVIDIA RTX Spark N1X superchip, opened blind pre-orders on October 8 at 9:00. The machine offers up to 1 Petaflop FP4 local AI compute, up to 128GB unified memory and 2TB storage, and comes preloaded with a 35B local model.

AIWall Street is increasingly letting AI agents interpret market information, so a bad source or missing context can spread into real-world decisions before anyone detects the error. The source article does not identify specific models, products, or institutions.

AIManus parent company Butterfly Effect has completed a new financing round of over $500 million, led by Boyu Capital and IDG Capital, and is rebuilding a team in Beijing to develop AI Agent products for the Chinese market. Its recruitment page lists 17 open positions, up from 11 before the National Day holiday, including AI Agent product manager, Agent Harness engineer, Agent evaluation engineer, and LLM algorithm engineer roles.

AIAlibaba's official handbook on AI Native R&D identifies four open challenges: infrastructure engineering complexity, enterprise knowledge assets not yet agent-friendly, organizational design, and the pace of AI iteration. The post's author argues that Agent Infra must suit non-deterministic agent operation and that enterprise knowledge needs top-down structuring and governance. The author also notes that organizational resistance in large companies makes AI adoption harder than in startups.

AIResearchers at robotics labs say that AI advances behind chatbots like ChatGPT and Claude will not quickly produce robots that are useful in everyday life. Many skeptics argue that using language- and image-based intelligence to master the physical world is far harder than it sounds, despite bold predictions from Elon Musk about Tesla's Optimus. Progress is real but incremental, as shown by Google DeepMind's Gemini Robotics controlling ALOHA 2 arms to pack a lunchbox.

AINVIDIA, Princeton University, and the University of Maryland introduced PivotOPD, an on-policy distillation method that teaches multi-turn LLM agents to recover from their most damaging early mistake. Tested on Qwen3-1.7B and Qwen3-8B students, it posts the best average against 13 baselines on ALFWorld, WebShop, and Search-based QA. It recovers from 72.7% of replayed pivotal mistakes, versus 20.3% for standard OPD, with no added inference cost.

AITeknium announced that Hermes Agent will be available on the ASUS RTX Spark PC when it releases soon. The post links the upcoming support to ASUS's ProArt P16 and P14 creator laptops, which ASUS says are powered by NVIDIA and Hermes AI.

AIAVEVA chief technologist Arti Garg argues industrial AI should augment rather than replace human supervisors in critical decisions, with guardrails defining where automated systems can act. She says organizations must rethink business processes and safeguards as foundation models, physical AI, and agentic AI enable more complex automation.

AILast week, several of South Korea's largest banks were hit by a cyberattack. A CrowdStrike report reportedly indicates the entire attack may have been carried out by one person. The attacker reportedly combined the open-source AI penetration tool ARTEX, DeepSeek v4.1-Flash, GLM-5.3, Grok 4.6, and Claude Code.

AIAnt Ling's Ling-3.1-flash is now available as a day-0 partner release on Vercel AI Gateway, free through October 13. The model is positioned as token-efficient, built for agentic and coding work, and suited to multi-step analysis.

AIAnt Ling announced a day-zero collaboration with AI/ML API, making Ling-3.1-flash available there for agentic and cowork scenarios. AI/ML API describes it as a 560B-parameter MoE model with about 25B active per token and up to 1M context, built for agents, coding, and long documents. The model is free to try on AI/ML API until October 13.

AIKargoBot has started a scaled AI freight network in Qipanjing, Ordos, combining human-driven trucks, autonomous trucks with cabs, and cabless transport robots on one system. The company says cabless robots could raise economic gain per vehicle from 20 percent to more than 30 percent, a target it has not audited. Platooning reportedly improves gross margin by about 10 to 18 percent versus manned haulage, with one lead driver able to head up to five follower trucks.

AITencent's WorkBuddy agent can take a plain-language request through to a WeChat mini-program preview and a publish request, according to a hands-on product test reported on October 8. In the test, the agent built a voice notebook with cloud login, storage and files, using Tencent's wand-asr-v1 for speech-to-text and GLM-5.3-Flash for sorting notes. WeChat's own review and filing steps remain outside the agent, so the test does not show that every mini program goes live automatically.

AIAnthropic released Claude Haiku 5.5, priced the same as OpenAI's GPT-6 Luna, with a 1M-token context window. Artificial Analysis scored it 43 on its Intelligence Index, slightly ahead of GPT-6 Luna at 38, but it uses about 3x more output tokens at max effort.

AIOpenAI quietly re-shipped Codex Cloud, which its Tibo says is now "pretty good." Codex Cloud can now securely connect to resources on a user's tailnet through Tailscale.

AIThe Grok bot's rate limit on X is 1,000 calls per day, which the author finds more than sufficient. Previously, using X's official API cost $20 per top-up and was spent quickly, while now it can be used for free.

AIJerry Liu called the decision to route tasks to the best available back-end model a good business move that keeps him using the bot. The background post from Elon Musk says the Grok bot will use the best back-end model for each task, including Claude Opus 5.5, Midjourney, Suno and other leading APIs.

AIThe author argues that in agent workflows, the real bottleneck is whether users can precisely express requirements, not the model or agent capability. When people work outside their expertise, they lack the precision needed for prompts and plans, forcing many imprecise iterations that waste time and tokens. The suggested fix is to have the model first teach the unfamiliar domain knowledge before acting.

AIMy Grok bot has found its first powerful use case. Using a personal agent to collect information from Twitter for you is far better than scrolling your phone yourself, manually gathering information, or getting addicted.

AIThe post announces DevDay 2026 and frames its theme as a shift from single-player to multiplayer agents. It gives no further details about specific products, features, or dates.

AIAnthropic is halving the price of cache reads on Claude Sonnet 5.5, to $0.10 per million tokens. The company says this makes Sonnet 5.5 about 20% cheaper to run on most long-running work, where agents rely heavily on cached reads.

AIYuchen Jin argues that Apple, which controls the entire iOS ecosystem, is best positioned to build a personal AI agent that can do nearly everything on a phone. He says the main limitation of Instint and Muse is that they cannot control most apps on his phone, and he criticizes Siri's current performance.

AIDeveloper Dex Horthy highlights Fold, a Effect-based agent harness by @0xblacklight that stores state as a portable append-only log. It runs on Cloudflare Durable Objects out of the box and includes a retro OpenTUI interface with 10 themes.

AIEvery moved its Slack-based Every Agent from its own OpenClaw fleet to Anthropic's Claude Managed Agents, letting the team focus on building a shared coworker. The newsletter, written by engineer Paridhi Agarwal, covers the gains and trade-offs of stopping self-run agent infrastructure.

AIMastra launched the Mastra Agency Program, a network of certified agencies and consultancies that build Mastra agents for clients. The launch includes five partners: Deerfield Group, Blue Drop Labs, Frontleap, Handpicked, and Young Security. Every partner has been vetted by Mastra's FDE team and receives direct access to Mastra's leadership and regular roadmap updates.

AIAnthropic has launched the Anthropic Cyber Mission, which starts with the Critical Infrastructure Defense Program for operational technology and OSS Scanner for open-source projects. The defense program brings frontier Claude models, on-site engineers and threat research to trusted providers such as Accenture, CrowdStrike and Palo Alto Networks. OSS Scanner gives enrolled open-source projects periodic free scans from its strongest models, with reports sent without human review and an expected true-positive rate above 90%.

Why it matters: The announcement shows how a frontier AI lab is packaging cyber defense around critical infrastructure and open-source maintainers, including the program's partners and access routes.

AIAnthropic is launching OSS Scanner, an opt-in service that runs periodic security scans of enrolled open-source projects using its strongest models at no cost. Its outputs are fully model-generated without human review, so some reports may be incorrect or invalid, though a pilot found 85 of 97 checked critical and high-severity findings met Anthropic's disclosure bar. Core maintainers of eligible projects can enroll through a GitHub pull request.

AIJohns Hopkins astrophysicist Brice Ménard, working as an Anthropic researcher, used Claude Science to produce the first complete map of the sky in ultraviolet light. Claude orchestrated agents to merge GALEX, Swift, and FIMS/SPEAR data, then predicted roughly a third of the sky that no UV telescope had observed, using relationships to visible, infrared, and radio data. Hidden test regions were reconstructed to within about 10% of real measurements, and each pixel is labeled measured or predicted with uncertainty estimates.

Why it matters: The post shows how an astrophysicist used Claude Science agents to merge UV surveys and predict missing sky regions, with a validation step that makes the method reusable.

AIBlock's AI capabilities lead describes using Claude Fable to plan large code migrations and direct smaller models like Opus and Sonnet on individual tasks. He says Block routes frontier and smaller models by task and keeps merges and production deploys behind human dual approval.

Why it matters: Block's engineering lead describes how frontier models orchestrate large migrations and how access, effort levels, and safeguards are managed across an organization.