#Deployment/Engineering

AIZ.ai reports that GLM-5.3 helped build and optimize the inference infrastructure for GLM-5.3-Flash. The system went from first successful run to production readiness in under two weeks, with end-to-end throughput tripling over the initial baseline. The team credited dense feedback from local correctness tests, execution traces, microbenchmarks, and end-to-end measurements for enabling targeted hypothesis testing.

AIAmp runners can now serve multiple directories, specified with repeated --dir flags or found automatically with --discover-dirs, which scans Git checkouts up to two levels deep by default. Runners also check for new releases about once an hour, install them, and restart into the new version once no thread is running, at most once every 12 hours, with auto-update disabled via amp.runner.autoUpdate.enabled: false.

AISpeakeasy is open-sourcing its full OpenAPI client suite under the AGPLv3 license, including generators for seven languages (Python, TypeScript, Go, Java, C#, PHP, Ruby), an agent-native CLI generator, and a documentation MCP server generator. Google said the move followed the May 2026 shutdown of the SDK generation provider it had been using, which it cited as evidence that closed-source generators pose platform risk. Google's new Google GenAI SDKs for the Interactions, Agents, and Webhooks APIs were built with this pipeline across six targets.

AIDatabricks rolled out Astra to every engineer, about 3,500 people, after a pilot with around 200 users. Engineers given Astra increased coding spend by roughly 60% compared to baseline. The company reports Astra outperforms Opus 5 and Sol 5.6 on highly complex system design tasks, but sees no clear gain on medium or low complexity coding. Astra gets a separate sub-budget in Unity Gateway to encourage selective use.

Why it matters: The post reports internal rollout data on cost and performance, showing how a company manages model access and budgets for engineers at scale.

AIMicrosoft signed a landmark agreement with the American Federation of Teachers and introduced a Privacy & Safety Standard for Schools covering Microsoft Education products. The standard limits how student and educator data is used, requires human oversight for consequential decisions and keeps school-created knowledge owned by schools. Microsoft also introduced Teach in Microsoft 365 Copilot, an education-first AI experience for educators.

AIMidjourney's September 16, 2026 alpha site update fixes the v8.2 editor, adds a Korean language option, and makes its mobile and tablet layouts fill the screen. Users can now drag images into folders with the sidebar collapsed, use Add to Folder without existing folders, and delete uploads from a menu. Default parameters and sref previews are listed as upcoming work.

AIDatabricks rolled out Astra to all of its engineers and reported internal measurements. Engineers given Astra increased overall coding spend by around 60% compared with baseline, and Astra outperformed prior top models on highly complex system design tasks, while the gain on medium or low complexity tasks was unclear.

AIGrok Voice is now live on fal, and the post says it answers in 0.70 seconds and can finish tool calls before a sentence ends. Its listed capabilities include transcription with word-level timestamps, text to speech in 30+ voices across 25+ languages, and voice cloning from two minutes of audio.

AIPerplexity says its Search SDK extracts passages relevant to a query from user-provided URLs. Agents can use those passages instead of full pages, keeping unrelated content out of the model context. The company also points to an Agent Skill for installing the Search SDK in coding agents.

AISundial fine-tuned Thinking Machines' Inkling-Small with RLVR on 3,978 verified TeX.StackExchange fixes, using rewards for compilation and PDF match and penalties for removed content. The trained model fixes 83.7% of LaTeX errors in under one second at $0.0013 per fix, according to the post. Sundial says it is rolling out the model in its editor, applying fixes as suggestions and rebuilding the PDF.

AIMosaic, Ponder Studio, and Revyl used early access to Google's Gemini Flash models to test agentic video understanding on long footage. Mosaic reports a 97% cut in median token usage and nearly double the ability to handle complex edits, while Ponder Studio reports a 0.967 F1 score and about 72% lower token costs for B-roll selection. Revyl says the approach improved mobile UI bug-catching accuracy by 65%. The capability is available now for video uploads and YouTube videos via the Gemini API.

AIBaseten has launched Hosted Tools, starting with Baseten Grounded Inference, a server-side web search capability for models hosted on Baseten. Developers enable it by adding a hosted search tool to a Messages, Chat Completions, or Responses request, and the platform runs the search loop with partners Exa, Keenable, Parallel, and You.com. In Baseten's benchmarks, agents using the hosted tools saw a 15% reduction in end-to-end latency compared with client-side tools, and the feature is in playground preview with 25 RPM rate limits and $2 of free credits.

AICohere is offering a hosted, encrypted solution for running LLMs with full data protection. The Model Vault uses confidential computing so data stays protected while in use, with no access by anyone, including Cohere.

AIElad Gil added a brief endorsement of Harvey's appointment of Adam Sadovsky as Chief Research Officer. The quoted post says Sadovsky previously co-led post-training at Microsoft AI and Google DeepMind and helped train Gemini 1.0 through 2.5.

AIAnthropic is merging Claude Cowork and Chat into a single Claude starting today, which Mike Krieger says removes the friction of choosing which product to start with. Per the @claudeai announcement, Claude will carry tasks forward even after the laptop is closed, asking for clarification when needed while users keep final say. The rollout to Pro and Max plans will take place over the coming weeks.

AIAnthropic's Felix Rieseberg says Cowork users need to take no action, since chats, projects, artifacts, connectors, and skills all remain available. Users can simply open the app and continue where they left off.

AIPeking University's Yu Zhaofei team and the Beijing Academy of Artificial Intelligence (BAAI) released Brainμ-Spike, a small convolutional network for spike camera image reconstruction that is paired with the Brainμ model. The package includes weights, inference scripts, and evaluation tools, but the base large model and LoRA weights are not yet released, so the full generation pipeline cannot run from this repository alone.

AITencent's WeVisDoc-4B, fine-tuned from Qwen3-VL-4B-Instruct, converts page images into structured Markdown with LaTeX formulas and HTML tables. It scores 95.38 Overall on OmniDocBench v1.6 and a mean Overall of 75.54 across three PureDocBench tracks, ranking first among compared end-to-end parsers in all four reported settings. The model is available on Hugging Face and runs through vLLM, which requires version 0.11.1 or later.

AIZed has launched the public beta of Delta, a multiplayer environment for coding with agents and reviewing their work, which replaces pull requests with shared threads. Delta is built on DeltaDB, which records edits and messages between Git commits, and it is free during the beta, with paid plans for individuals and teams to follow.

Why it matters: The post explains how Delta replaces pull requests with shared agent threads and DeltaDB, showing a concrete alternative to the GitHub review workflow.

AIClaude now makes Cowork capabilities available from any conversation without choosing a mode first, with chats, tasks, projects, connectors, and skills carrying over. Users can also create designs, decks, and docs in any conversation, including Claude Code and the Artifacts tab, and edit them with Claude.

Why it matters: The release merges Cowork tasks into ordinary chats and adds design, slide, and doc creation, changing how Claude users start larger work.

AITogether AI says teams can migrate from closed to open-source models in weeks to months, not years, using a managed service. The guide recommends defining the use case, shortlisting candidates with relevant benchmarks, and replaying existing traffic against them to test accuracy, cost, and latency against current targets.

AIGoogle has put Agent Anomaly Detection into Private Preview on the Gemini Enterprise Agent Platform, a reasoning-based audit layer that reviews agent reasoning traces, tool calls, and execution flow to flag behavioral anomalies and policy violations. It runs asynchronously without adding runtime latency and publishes findings to Security Command Center. The preview requires ADK 1.2 or later.

AIGrok Build now has memory, which records conventions, decisions, and project facts after each completed turn and reads them in later sessions. Notes are stored per project plus a global set, and the /dream command organizes them into topic files while /memory opens a read-only browser. The feature is available now and applies to new sessions.

AIDelphi, which turns experts' knowledge into digital minds, runs its Python backend on Vercel with a 10-person team and no dedicated infrastructure role. The stack uses Vercel Workflows for long-running agents and Vercel Queues for background jobs. The team reports more than 100 production deploys a day.

AIMicrosoft has launched Foundry Dev Pack, an all-in-one installer that sets up tools for Microsoft Foundry development across the terminal, IDE, and coding agents. Depending on the environment, it installs Azure CLI (az), Azure Developer CLI (azd) with the Microsoft Foundry Extension for azd, the Microsoft Foundry Skill, the Microsoft Foundry Toolkit for Visual Studio Code, and Foundry Canvas (preview), with the last two conditional on VS Code or GitHub Copilot App being present.

AIOpenAI's Greg Brockman says ChatGPT Work can operate over and act on a company's data, including building dashboards, by connecting existing tools such as PowerBI, Tableau, Clickhouse, Oracle BI, and AWS Redshift. The linked ChatGPT announcement describes a Data agent with a Data Plugin that turns company data into answers, interactive dashboards, and actions through conversation.

AIGoogle is rolling out Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking across several channels. Consumers get them in Search Live and Gemini Live, developers get public preview access through the Gemini API, and enterprises get private preview through Gemini Enterprise, with Customer Experience support coming soon.

Why it matters: The post lays out where each Gemini 3.8 Live variant reaches consumers, developers, and enterprises, which clarifies access paths for a voice model release.

AIGoogle DeepMind introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, two live dialogue models for voice agents. Extended Thinking scores 82.6 on Artificial Analysis' Speech to Speech Quality Index, 68.6% on τ-Voice, and 97.7% on Big Bench Audio. Gemini 3.8 Live is rolling out now in the Gemini API, Google AI Studio, and Search Live, with enterprise access in private preview.

Why it matters: The release covers a voice model's benchmark results and availability across developer, enterprise, and consumer products, useful for judging voice agent options.

AICognition and AWS have entered a multi-year Strategic Collaboration Agreement to help enterprises deploy the Devin autonomous engineer in production. Devin can be purchased through AWS Marketplace, and the companies are exploring deeper engineering integrations within customers' AWS environments. Mercedes-Benz reportedly used Devin to analyze more than 200,000 lines of COBOL, reducing an estimated eight-month modernization project to eight days.

Why it matters: The collaboration shows how an autonomous coding agent is being packaged for enterprise legacy modernization inside existing AWS environments, with concrete customer migration figures.

AINVIDIA's RT-DETR Hand Detection v1.0 detects and localizes left and right hands in RGB images, outputting 2D bounding boxes with per-hand confidence scores in a single pass. The model, built on RT-DETRv2-S with HGNetv2-S backbone and about 20M parameters, is intended as a region-of-interest stage for downstream 3D hand pose estimation and is exported to ONNX. The source describes it as for demonstration purposes rather than production use, runs on NVIDIA Lovelace GPUs under Linux, and is licensed under the NVIDIA Software and Model Evaluation License.

AIPeriodic Labs chose SGLang and Miles to build Neon, an open-source model it says surpasses GPT-6 Astra on its analysis benchmark after mid-training and RL on 1,300 H200s. RadixArk says Periodic extended both frameworks for scientific RL at trillion-parameter scale, delivering more efficient training, lower memory use, and 2.5x faster inference. The work has been contributed back to both projects.

AIDwarkesh Patel reports that materials synthesis experiment spaces are very wide but amenable to depth-first search, where each next experiment becomes better designed and more informative as data accumulates. The post is a lab visit reaction and does not name specific models, figures, or results.