All AI updates

Merge builds product and business tools with OpenAI technology

Merge uses GPT-6 Astra, ChatGPT, and Codex to review product code and support marketing. Its teams also use AI to plan executive dinners and prepare sales updates.

0.167.0

Codex 0.167.0 adds an agent command center opened with `codex agents --no-daemon`, unread-result markers in the command center and session picker, cross-backend session switching from the terminal picker, local image attachment previews, and a Google Cloud Vertex AI setup wizard. Bug fixes keep saved chats visible across model provider changes, preserve transcript history when Ctrl+L clears the terminal, and resolve network approval prompts when their requests end. Documentation and chores cover managed network constraints under [permissions.network], DotSlash manifest updates, and reduced jemalloc arena counts for Linux musl builds.

Clearing up a robot’s fuzzy memory

MIT researchers developed UQ-DAAAM, a memory system that quantifies a robot's uncertainty about object descriptions by comparing multiple viewpoints and fusing them into more accurate captions, resolving ambiguities for about 92 percent of objects with only two additional captions each. The system improved object description accuracy by 5 percent over the state of the art and produced more accurate answers about objects, their locations, and when they were seen. The work, led by senior author Luca Carlone with co-lead authors Harry Zhang and Nicolas Gorlo, will be presented at the Conference on Neural Information Processing Systems.

Accelerators take care of the basics

Sierra describes Accelerators, prebuilt packages for high-volume workflows such as order tracking, appointment scheduling, claims intake and ID verification, which come with basic integrations and guardrails and are customized for a business using Ghostwriter. Accelerators are maintained by Sierra but track customer changes separately so updates can be reviewed before being applied, and they connect to a company's existing software systems. The first Accelerators are available today in Agent Studio, with user-built and shareable Accelerators planned.

SRT file guide: Create, edit, and use subtitles in video

This guide explains that an SRT file is a plain-text subtitle format storing numbered blocks, each pairing a start and end timestamp with on-screen caption text, and that it is widely supported across players and editing software. It covers the format's structure with an example, how to open SRT files on any device, converting transcripts to SRT via speech to text, editing to fix timing and readability errors, adding subtitles to video, and choosing between SRT and WebVTT. It also notes ElevenCreative and ElevenAPI options for generating and exporting subtitle files, including the Scribe v2 model's word-level timestamps across over 90 languages.

Getting started with Web Answer on Amazon Bedrock AgentCore

Amazon announced the general availability of Web Answer on Amazon Bedrock AgentCore, a managed Gateway tool that takes a natural-language question and returns a synthesized answer with source citations in seconds, replacing a self-built search, retrieve, synthesize, and citation pipeline. The post contrasts it with Web Search (which returns ranked URLs and lets you keep your own model), then walks through granting the Gateway service role bedrock-agentcore:InvokeWebAnswer, creating a Web Answer target, and calling it from a Strands agent. It is available in us-east-1 at $10 per 1,000 tool calls, with a 10 transactions per second per account rate limit.

0.168.0-alpha.7

This page is a GitHub Changelog entry titled "0.168.0-alpha.7"; the supplied material consists only of site navigation and product menu text, with no details about the release itself.

Build Live Apps in Cursor

Cursor now lets users build and deploy live apps directly from the editor: a cloud agent scaffolds the web app, sets up a Postgres database if needed, configures authentication behind Cursor SSO, provisions hosting, and returns a shareable link at a chosen domain. Publishing runs deterministic and agentic security checks, and nothing goes live until checks pass and the user explicitly publishes; team admins can view all apps and secrets, add review rules, approve blocked publishes, and unpublish apps. The capability is available on all paid plans with cloud agents enabled, deploying from new or existing repos on Cursor Origin and GitHub, with GitLab support coming soon.

Vera Delivers the Best of Both Worlds: Highest Throughput and Fastest Agents

NVIDIA says its Vera CPU outperforms both the fastest and the most energy-efficient x86 CPUs in a 100-megawatt data center comparison, delivering 1.6x more agents per watt versus the fastest x86 CPU and 2x faster agent speed with leading energy efficiency versus the most energy-efficient x86 CPU. Early adopters report production-style gains, including Perplexity running agentic coding tasks 1.5x faster, DeepInfra reaching up to 2.2x the speed of the best x86 processor on agentic orchestration, and ClickHouse scaling near-linearly across 176 cores on its standard binary. Daytona brought 1,000 agent sandboxes online in 10.8 seconds versus 88.4 seconds with x86, and Modal achieved up to 2x sandbox throughput per core versus comparable x86 hardware.

New NVIDIA Cosmos 3 Open Models for Interactive World Simulation

NVIDIA's Cosmos world model platform adds four new open-weight checkpoints to its Cosmos 3 family: Cosmos3-Nano-Transfer-Auto for multi-view driving data with varied weather, lighting and time of day; Cosmos3-Nano-Sim-Auto for streaming synchronized camera and LiDAR driving simulation; Cosmos3-Nano-Sim-Transfer for turning simulator controls and text prompts into photorealistic video; and Cosmos3-Nano-Sim-Bimanual for predicting video from robot or camera actions. All four are post-trained examples of Cosmos 3 Nano with 16B trainable parameters and use OpenMDW1.1, permitting commercial and non-commercial use under its terms. The post also details a four-stage training recipe (bidirectional adaptation, teacher forcing, consistency distillation, self-forcing) and post-training guidance for adapting the models to new sensor rigs and embodiments.

Frameboy: Atlas in Real-time

The team behind Atlas, an autoregressive world model that generates frames one at a time with pixel-perfect camera control, released Frameboy, a research preview that streams frames from cloud GPUs to a browser at up to 15 fps as the user controls the camera. Frameboy serves a highly optimized Atlas variant, and the real-time capability is also integrated into Viewfinder, a web app (currently in closed beta) for building scenes frame-by-frame and exploring them live via a streaming Frameboy overlay. The post argues that faster generation lowers the cost of a "bad idea" and shifts interaction from cautious manual prompting toward intuitive exploration.

How Amazon Marketplace evaluates content quality with Amazon SageMaker AI

Amazon describes a system that renders a full A+ Content page as a single snapshot and scores it with a fixed Vision-Language Model across four rubric dimensions, then fuses those scores with four OCR readability signals in a CatBoost classifier trained on expert annotations. On a 20 percent holdout set of 5,478 product detail pages from 10 categories, the fused model reached 0.94 accuracy and 0.95 F1, versus 0.80 and 0.84 for an averaged-score baseline, and SHAP values plus an LLM turn the results into prioritized recommendations. The pipeline is implemented on Amazon SageMaker AI using Processing Jobs, Inference endpoints, Feature Store, Training Jobs, Model Registry and monitoring, with a diffusion model generating improved image assets.

GPT‑6 Cyber: Advancing the frontier of aligned cybersecurity agents

OpenAI introduces GPT‑6 Cyber, a cybersecurity model designed to find exploitable vulnerabilities and develop fixes, trained to stay within authorized scope.

More visibility and control over Claude Enterprise for security teams

Anthropic announced new Claude Enterprise security and admin features, including a security center that compares organization settings with Anthropic's recommended configuration, generally available prompt and pre-tool-call inference hooks that send prompts and tool calls to existing data security tools, and centralized management with a parent-organization admin console at admin.claude.com. New access controls let admins sign members out via the Admin API, manage their own IP allowlist, and block personal and outside Claude accounts in Claude Desktop on managed computers, while an EDR integration runs an organization's sensor inside the Cowork virtual machine. The security center, centralized management, Admin API sign-out, self-serve IP allowlist, and Claude Desktop settings management are in beta, and EDR in Cowork is in beta on macOS and Windows for Claude Team and Enterprise.

Introducing Personal Finance in Claude

Anthropic is launching Personal Finance in Claude, which lets users in the US securely connect bank and card accounts through Plaid with read-only access and ask questions about balances, transactions and spending in everyday language. It starts rolling out in beta this week on the Pro, Max and Ultra plans in the US, and Claude cannot move, manage or invest money, shows no ads or sponsored recommendations, and does not offer its own financial products. Conversations in the Personal Finance hub are not used to train models, accounts can be unlinked anytime, and Anthropic says Claude can make mistakes, so important figures should be checked with your bank.

Nanbeige has released Nanbeige4.3-33B scoring 20 on the Intelligence Index

Nanbeige, the foundation model research team at BOSS Zhipin (operated by Kanzhun Limited), has released Nanbeige4.3-33B, an Apache 2.0 open-weights dense reasoning model with text-only input and output and a 524K-token context window. It scores 20 on the Artificial Analysis Intelligence Index v5, matching Qwen3.8 27B (xhigh) at 20, with relative strengths in terminal use (30% on Terminal-Bench 4.0 vs. 6%) and physics reasoning (16% on CritPt v2's Hard Subset vs. 4%), but trails on AA-Briefcase v1.1 (1162 vs. 1391 Elo), FrontierCode 2.0 (11% vs. 15%) and HLE-Diamond (7% vs. 15%). It shows high abstention and low hallucination on Omniscience (18% vs. 29% hallucination rate, declining 71% of questions) and used 158K output tokens per Intelligence Index task vs. 110K for Qwen3.8 27B.

0.168.0-alpha.6

This page is a GitHub Changelog entry titled "0.168.0-alpha.6," but the supplied material consists only of site navigation and product menu text (Platform, Copilot, Actions, Codespaces, Advanced Security, pricing, etc.) with no release notes or details about the version.

Independent Assessment of AI R&D Acceleration at Anthropic in October 2026

METR conducted a high-access evaluation at Anthropic from August 12 to October 29, 2026, concluding that Anthropic is likely improving model capabilities on well-specified tasks at around twice its March 2024–October 2025 rate, with AECI benchmark scores growing at an estimated 1.9x the pre-October 2025 rate (95% CI: 1.5–2.8x). METR estimates that a majority (perhaps ~80%) of the current acceleration is due to AI, found no clear bottlenecks to continued progress over the next 12 months, and cautioned that Anthropic redacted most of the evidence it found most informative.

Review of the "Risks from automated R&D" section in Anthropic's Risk Report (August 2026)

METR reviewed the "Risks from automated R&D" section of Anthropic's August 2026 Risk Report and agrees that internally deployed models as of July 15th posed low risk of directly automating R&D or causing acute catastrophic risks. However, METR argues the report's claim that Anthropic had not crossed its 2x Acceleration Threshold relies on contestable judgment calls not clearly communicated publicly, and that the public report understates the acceleration occurring; METR believes it is plausible the threshold had been crossed. METR finds the argument that the Full Substitution Threshold had not been crossed broadly compelling, while noting lower confidence due to a lack of high-confidence unsaturated benchmarks.

A builder’s guide to scaling Databricks Apps

This guide outlines scaling Databricks Apps across four stages—design, build, test and deploy, and operate—starting with agreed latency and peak-load targets before any code is written. It recommends keeping app compute for serving the UI, sending heavy work to purpose-built services, using async code, splitting soft and hard state so sessions can be rebuilt on any instance, and letting load tests and telemetry determine instance size, worker count and instance count. It also covers connection budgeting for Lakebase, graceful shutdown within the 15-second SIGTERM window, and monitoring and alerting in production.

GitLab Integration

Amp now integrates with GitLab on gitlab.com and self-managed instances, letting users start projects from GitLab repositories and giving Orbs a signed-in glab so the agent can open and comment on merge requests and read issues and pipelines. Amp can subscribe to comments and CI for merge requests it opens, and users can ask Amp to start a new orb for every new merge request or issue in a project, managed under Settings → Integrations & MCP → GitLab. Each person connects GitLab in Settings → Integrations & MCP, then a group Owner installs Push Updates once for the workspace.

Faster Retrieval for Better, More Efficient RAG

Apple researchers present amortized maximum inner product search (MIPS), training small neural networks offline on a fixed key database to learn the support function of the key set, which is convex and whose gradient returns the best-matching key. SupportNet learns per-cluster support functions to route queries more accurately than centroid-based routing, while KeyNet rewrites queries before a standard IVF index; on Natural Questions, SupportNet reached 95% routing recall about 1.9× faster than centroid routing, and KeyNet cut the time to 99% recall by up to ~8×. The work was presented at ICML 2026.

Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips

Google DeepMind released Gemini Omni 1.2 Flash, now generally available across the Gemini API and Gemini for business, adding native 15-second video generation, multi-turn conversational editing, and stronger reference adherence. The model is priced 30% below Gemini Omni 1.1 at $0.07 per second for 720p video, with duration settable in one-second increments from 3 to 15 seconds. It also adds more natural speech and lip-sync, more controllable cinematic camera moves, and outputs carrying an imperceptible SynthID watermark.

Formally verifying agent-written full-stack web apps with PDD Kit

GitHub Next describes PDD Kit, a toolkit for Proof-Driven Development that turns a developer's intent into a formal specification in Verus, proves the agent's code meets that spec for every input, and falls back to tests where proofs cannot reach. The team applied it to Pet Store, a roughly 5,000-line Rust full-stack app, formally verifying its login logic while the developer reviewed the intent rather than reading generated code line by line. PDD Kit labels each result as either formal verification or empirical evidence and asks developers to sign off on each trusted component.

Introducing Latency Metrics for Mastra Observability

Mastra now records TTFC (time to first chunk) and TTFT (time to first token) on every span, from model calls up to agent runs, so traces, spans, and threads can be filtered and sorted by these metrics on Mastra platform and Studio. The metrics work on self-hosted projects using DuckDB, PostgreSQL, or ClickHouse, and are queryable via the Mastra CLI ([email protected]+), Client SDK (@mastra/[email protected]+), and observability endpoints. Studio support requires [email protected] and @mastra/[email protected] or later.

Configure, Certify, and Operate Kubernetes GPU Clusters with NVIDIA DSX OS

NVIDIA DSX OS addresses the problem that a single underperforming GPU can slow an entire distributed training job, even when the node passes basic health checks and shows normal utilization, so operators may only discover the issue once the workload is running. The article notes these issues can surface at any point in a cluster's lifecycle, since a GPU cluster that passes readiness checks at handover changes as nodes are replaced and drivers are patched.

AI for scientific discovery: How researchers use Claude

Anthropic researchers used privacy-preserving classifiers to label 207,049 Claude and Claude Code research sessions (August 2–September 30, 2026) by field and task, finding heavier use in formal sciences like computer science, mathematics, and statistics than in health and lab-based physical sciences, with data analysis and writing dominating. Claude use rises with a constructed "theoretical AI capability" score, which correlates with reinforcement-learning suitability and cognitive work and negatively with physical, interpersonal, or regulated work. Scenario modeling suggests current capabilities could double research output in some fields, and full automation of non-physical work could raise output up to 15-fold in fields like computer science and statistics, versus three- or fourfold in bench sciences.

Agents aren’t humans (but they kinda are)

This post argues that remotely hosted personal AI Agents act as intermediaries between Users and Sites, stripping Sites of the identity, reputation and accountability signals they rely on for security, fraud prevention and monetization. It proposes the Intermediation Principle and surveys emerging standards — Web Bot Auth, MOLE, OAuth, AIPREFs and the new Trusted Agent Delegation coalition — while noting Cloudflare's new tools for seeing Agents, setting per-Agent access policies, and tying Agents back to users via Account Abuse Protection.

rust-v0.168.0-alpha.5

This GitHub Changelog entry lists the release rust-v0.168.0-alpha.5. The supplied material contains only the page's navigation menu and site sections, with no details about the release's contents.

Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips

Google DeepMind released Gemini Omni 1.2 Flash, now generally available across the Gemini API and Gemini for business, adding native 15-second video generation, multi-turn conversational editing, and stronger reference adherence. The model is priced 30% below Gemini Omni 1.1 at $0.07 per second for 720p video, with duration settable in one-second increments from 3 to 15 seconds. It also adds more natural speech and lip-sync, more controllable cinematic camera moves, and outputs carrying an imperceptible SynthID watermark.

Featured coverage of this story: 《Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips》

From Delta Sharing to OpenSharing: what changed, what's next

In June 2026 Databricks and the Linux Foundation announced that Delta Sharing has evolved into OpenSharing, an open, vendor-neutral protocol under the Linux Foundation AI & Data Foundation that extends sharing beyond tables to volumes, agent skills, and AI models, and adds Apache Iceberg REST Catalog support and on-premises or private cloud reach. Existing Delta Sharing clients and integrations remain compatible, and enabling OpenSharing on a Unity Catalog metastore is a configuration step rather than a migration. Databricks also introduced capabilities on top of the protocol, including SecureConnect (generally available), Global Distribution (private preview), Genie Agent Sharing (beta), and Lakebase table and change data feed sharing (public preview).

Which Data Collaboration Tool: OpenSharing, Clean Rooms, Marketplace

This Databricks blog compares three data-sharing methods: OpenSharing, which gives recipients direct read-only access to live assets such as tables, views, notebooks and AI models; Clean Rooms, where up to ten parties combine data and run only mutually approved code without exposing full underlying datasets; and Marketplace, a public storefront for discovering and publishing data and AI products. It advises choosing based on the relationship with the recipient and notes the tools are not mutually exclusive.

Detecting signs of hypertension using wearable sensors

Google researchers describe AI-PPG-ACC-HTN, a multimodal AI system that uses smartwatch PPG and accelerometry data, without cuff calibration, to detect a persistent hypertensive state by aggregating 15-second waveform embeddings over multiple days alongside height. Validated against 24-hour ambulatory blood pressure monitoring in a 196-participant, five-site US study, it reached 65.8% sensitivity and 90.0% specificity for initial screening and 78.4% sensitivity, 90.9% specificity and 96.7% PPV for confirmatory testing among 48 participants with elevated office BP. The authors frame it as early feasibility research for passive opportunistic screening, not a replacement for clinical cuffs or a standalone diagnosis.

Building the Agent Cloud: Introducing Baseten Sandboxes

Baseten introduced Sandboxes, stateful agent runtimes built on microVMs that snapshot filesystem, memory, and running processes on disconnect and resume in about 25 ms, with boot from standby in 25 ms and scaling to thousands of sandboxes within seconds. They are the first piece of a planned agent cloud, to be followed by Agent Drive (a shared distributed filesystem) and Agent Proxy (external network control with credential injection, permissions, and audit logging). Sandboxes are free to run until November 16th.

Introducing the Mastra Custom Harness Program

Mastra launched a Custom Harness Program in which its forward-deployed engineers embed with customer teams to co-build agent harnesses on Mastra. The post cites Salesforce, MongoDB, Workday and Marsh McLennan as teams that built their own harnesses, including Workday's move from Amazon Bedrock Agents to Mastra in 31 days, which cut 3,000 lines of code and lowered vector-storage costs by 98%. The program includes dedicated FDEs, observability and evals, custom MCP and human-in-the-loop workflows, auth and governance controls, and early access to the Mastra Enterprise Platform.

Read more: How the Waymo Driver Manages Severe Winter Weather

Waymo describes how its Driver handles winter conditions, using cameras, radar and lidar with heaters and micro-wipers to keep sensors clear, and measuring real-time traction to adjust steering, braking and acceleration. The company says it may slow, limit or pause service when weather exceeds the boundaries defined by its Safety Framework, drawing on fleet-shared road data, forecasts and local officials. Waymo says it has accumulated tens of thousands of real-world miles in snowy regions including Upstate New York, Michigan's Upper Peninsula and the Sierra Nevada, and will conduct closed-course evaluations and heavy snowfall data collection in Alaska this season.

More Performance From Every Byte: How Full-Stack Codesign Solves AI’s Memory Challenge

NVIDIA describes how its full-stack codesign approach addresses growing GPU memory pressure from larger models, longer contexts and agentic applications by designing GPUs, CPUs, networking, storage and software together across a memory hierarchy. Techniques include NVFP4 for model weights and KV cache, distributing weights across up to 72 GPUs in NVL72, offloading inactive context to CMX context memory storage via Dynamo, and keeping RAG indexes outside HBM with Nemotron Retriever and cuVS. NVIDIA reports that from Blackwell Ultra to Vera Rubin, HBM capacity per GPU rose roughly 3% while agentic throughput per dollar rose up to 30x on SemiAnalysis AgentX.

The Wizard of Open Weights: Chris Alexiuk Makes AI Easy With Open Models

NVIDIA senior product research engineer Chris Alexiuk teaches developers to build, fine-tune, evaluate and ship production LLM applications and agents through YouTube, arguing that fully open source AI—with access to weights, data and training recipes—is necessary for democratizing intelligence. He calls himself the "No. 1 Nemotron Fan" and promotes NVIDIA's open source Nemotron model family, which spans models like Nemotron 3.5 Lightning for high-frequency execution work and Nemotron Ultra for long-running autonomous agents and enterprise tasks. Alexiuk traces his interest in machine learning to a computational physics class at Trent University and an early project building a classifier to detect his friend's pet bird on a webcam.

Sierra acquires Anon

Anon, which built infrastructure to help enterprise AI agents securely use websites the way people do—navigating pages and taking actions like signing in or filling out forms even without an API—is joining Sierra. Anon raised more than $13M and shipped a product used by hundreds of customers, and its team's expertise and technology will be brought to Sierra's agents, which already handle billions of customer interactions each year. Anon says it will help Sierra's agents complete steps that require signing in to a website, navigating a portal, or entering information through a browser.

Five Ways NVIDIA Co-Design Maximizes the Value of Every Byte in the AI Factory

NVIDIA outlines five co-design approaches for maximizing the value of every byte in an AI factory, where high bandwidth memory (HBM) is described as the fastest, most expensive and most contested memory tier. The piece argues that bytes spent on redundant weight copies, idle sessions' KV cache or recomputed prefixes are bytes not producing tokens, shifting the question from how much memory a GPU has to how much useful work each byte delivers.