Search “Meta”

1Password Launches 1Password for Agents, Letting AI Agents Complete Logins Without Seeing Passwords

1Password has launched 1Password for Agents, connecting its password vault to AI agents: when an agent reaches a login form, it authenticates to 1Password using a digital key; after the user confirms which credentials are needed and what they are for, 1Password fills in the form directly—credentials are never handed to the agent and are removed once login succeeds, so the agent cannot read or reuse them. The integration currently supports only Meta's Muse and Grok Bot, with OpenAI Dots and Anthropic Claude Duet to follow.

1Password launches AI agent password authorization feature: agents can log into accounts but cannot read passwords

1Password has launched a credential authorization feature for AI agents. After an agent proves its identity with a unique digital key, it sends a request to the user, who approves or rejects it on their own device. Once approved, 1Password fills the login information directly into the website login interface; the password is not handed to the agent in text form. The feature will support agents and tools such as Meta Muse, ChatGPT Work, ChatGPT Dots, Claude Duet, Grok Bot and Town, and platforms including Browserbase, Browser Use and Perplexity are also advancing integration. The original text also notes that even if an agent cannot see the password, it may still perform actions the user does not want after approval, so users need to carefully review authorization requests.

Google paper proposes OO-Harness: evolving specialized agent harnesses by task type

A new Google paper proposes OO-Harness, which evolves agent harnesses by code class: each specialized harness inherits from a parent class and changes only the parts that its task cluster needs adjusted, after which a router selects a specialized harness or keeps the base result based on the base run's outcome. On AutomationBench business workflows, this method scores 77.41%, versus 61.62% for the base harness and 63.92% for the global method Meta-Harness; on unfamiliar applications it retains the base harness and maintains a 59.18% pass rate, while Meta-Harness's global harness drops to 44.90%. The authors recommend using the base harness as a fallback for mixed workloads and routing to a specialized harness only when a run exposes a known problem.

Stanford CS 329Z Lecture 9 Outlines the Selection of Agent System Optimization Methods

Lecture 9 of Stanford's CS 329Z course presents a selection diagram for Agent system optimization: the prerequisite for optimization is having data, a system, and evaluation metrics all in place; generally, context optimization (prompts and few-shots) is done first because it is cheap, fast, and sample-efficient; when context optimization is insufficient, or when you want to distill capabilities into a small model, or when prompts grow longer and longer, then you move on to changing the weights. The lecture notes also review reflective prompt optimization methods such as OPRO, MIPROv2, TextGrad, and GEPA; weight optimization paths such as SFT, RL (PPO vs. GRPO), and on-policy distillation; and approaches that let agents self-optimize, such as the Darwin Gödel Machine, Meta-Harness, and karpathy's autoresearch. They also point out that agents are already decent optimizers, making evaluation metrics the link that most needs to be controlled.

Recursive releases its first model Flint: local coding MoE, 125+ tokens/s on a MacBook

Recursive releases its first model Flint, a MoE with 40 billion total parameters and 2.5 billion activated parameters per token, aimed at local Agent coding, with weights open-sourced under Apache 2.0. With short context, it exceeds 125 tokens per second on an M4 Pro MacBook Pro and over 150 on a 64GB RTX Spark, and supports llama.cpp, Ollama, LM Studio, vLLM, SGLang, and Unsloth. Nearly every stage of its R&D involved the automated research system Eureka Machine, including architecture search, quantizing the output layer to 4 bit with offline distillation, and fusing Metal kernels for M4/M5.

MiniMax responds to H3 integration with TPU: looking forward to open-source work on efficient video inference

MiniMax officially responded to SGLang adding MiniMax H3 to the TPU support list of SGLang-TorchTPU, and said it looks forward to more open-source work on efficient video inference. SGLang said that SGLang-TorchTPU was built jointly by SGLang, the Google TPU team, and the Meta PyTorch team, has been used for production inference at Meta, and supports models including Muse Spark Open, GLM-5.3 Flash, Kimi K3, Ling 3.1 Flash, and MiniMax H3.

Artificial Analysis releases AA-PersonalityBench sycophancy evaluation

Artificial Analysis launches AA-PersonalityBench, which evaluates the sycophantic personality trait of models. The three oldest models tested—GPT-4o, Llama 4 Maverick, and GPT-3.5 Turbo—ranked highest, while Grok 4.7, Claude Fable 5.1, and Inkling ranked lowest. All five Meta models tested scored above 65 (Llama 4 Maverick 97, Muse Spark 1.3 87, Muse Glimmer 79, Muse Spark Open 76, Muse Wonder 65), all Claude models scored 55 or below, and Grok 4.7 was the lowest overall.

Anthropic releases personal agent Claude Duet, built on the Opus 5.5 model

Anthropic has released a personal agent app called Claude Duet, positioned as a "proactive agent" that can run in the background and proactively push tasks to users; it is currently in research preview with a waitlist. The tool is built on the flagship Opus 5.5 model released at the end of September, and it pushes a daily morning briefing from connected apps, notifies users when something needs attention, and requires user approval for some tasks. Mike Krieger, co-lead of the Anthropic Labs team, says it can judge when to act proactively; Karan Ahuja, who leads Duet development, previously headed product management at Meta Superintelligence Labs.

Trump says large data centers must generate their own power and pay cash dividends to communities

In a Truth Social post, Trump said large new data centers must generate their own power and pay cash "dividends" to their host communities. The closest existing commitment is the Ratepayer Protection Pledge signed in March this year by Amazon, Google, Meta, Microsoft, OpenAI, Oracle and xAI, under which signatories pledged to build or purchase generating capacity and bear infrastructure upgrade costs, but experts believe the pledge may not be enforceable at the federal level. Trump said some projects that were initially highly unpopular became popular after financial agreements were reached, and the dividend proposal goes further than those pledges.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

SGLang-TorchTPU Open-Sourced: Bringing the SGLang Serving Stack to TPU

The SGLang team, in collaboration with Google and Meta, has open-sourced SGLang-TorchTPU, a PyTorch-based native backend that brings the full SGLang serving stack to Google Cloud TPU and is already in large-scale production use at Meta. Initial support covers models including Muse Spark Open, GLM-5.3, Kimi K3, DeepSeek V4, Qwen3.8, and Gemma 4, with features such as speculative decoding, KV cache offloading, PD disaggregation, and FP8 KV cache. The team also used TorchTitan and Miles to complete GRPO training of Qwen3-4B on a single 4-chip TPU v7x, with the reward curve matching a 6× H200 baseline within random fluctuation.

English translation is being prepared.