The AI Wire

techcommunity.microsoft.com ▸ tooling· filed 2 Oct 2026

Open-source demo separates AI model proposals from sandboxed tool execution

An open-source demo runs a Copilot-based food-ordering assistant on Azure Kubernetes Service, with tool executions confined to per-session sandboxes built from OpenSandbox and Kata Containers. The project separates model-proposed actions from actual execution, showing the execution platform determines whether and how proposals become actions. It also documents operational issues including image warmup timing and a Markdown rendering fix.

azure ▸ tooling· filed 10 Sept 2026

Microsoft introduces APIOps CLI for Git-based Azure API Management configuration

Microsoft has announced APIOps CLI, an open-source command-line tool that manages Azure API Management configuration through Git-based workflows. It extracts existing APIM configuration into versioned artifact files, supports pull-request review and CI/CD pipeline scaffolding, and publishes approved configuration back into environments. The tool is intended to address manual configuration changes and limited change visibility by treating API configuration as code.

azure ▸ product· filed 10 Sept 2026

Claude Fable 5.1 arrives in Microsoft Foundry alongside restricted Mythos 5.1

Anthropic has released Claude Fable 5.1 in the Microsoft Foundry model catalog and restricted Claude Mythos 5.1, describing both as the same underlying model with different safeguard configurations. The company says the main advance is sustained performance on multi-step, tool-use workflows, with fewer overconfident errors and better reporting of incomplete work. Anthropic recommends testing on realistic workloads and simplifying prompts built for earlier models, though these claims have not been independently verified.

simonwillison.net ▸ product· filed 31 Aug 2026

Simon Willison's hands-on test of ChatGPT Work: open internet access, browser automation, and unanswered safety questions

Simon Willison's hands-on test shows ChatGPT Work is an agent platform with unrestricted internet access, a headless Chrome browser, and parallel sub-agents, but OpenAI has not explained how it defends against prompt injection. The documentation is so sparse that he had to ask the agent itself to list its 223 tools and 44 skills. This leaves core safety mechanisms unaddressed in OpenAI's marketing.

youtube.com ▸ models· filed 30 Aug 2026

GLM 5.3 Flash: Z.ai’s cheaper, multimodal sibling matches larger models but slows on max reasoning

Z.ai's GLM 5.3 Flash is a 320-billion-parameter multimodal model that costs about one-ninth the output price of the flagship GLM 5.3. Independent tests show strong function-calling and agentic behavior, but max‑reasoning mode slows throughput to 50–60 tokens per second, and agentic fallbacks to the larger GLM are expensive, pushing users toward other frontier models for critical tasks.

youtube.com ▸ industry· filed 30 Aug 2026

Andrew Ng: Big AI labs are fear-mongering to shape regulation

Andrew Ng contends that leading AI companies are exaggerating risks to drive regulation that favors incumbents, while dismissing fears of a job apocalypse and warning that AI models hinder learning. He advises individuals to adopt AI skills and focus on product decisions rather than fear.

azure ▸ product· filed 30 Aug 2026

Azure API Management Standard v2 Adds Availability Zone Redundancy

Microsoft has added Availability Zone redundancy to the Azure API Management Standard v2 tier, allowing customers to distribute gateway capacity across multiple zones without upgrading to a higher-priced tier. The platform handles failover automatically, but the protection covers only the gateway layer, so backend services and disaster recovery still require separate planning. The feature is enabled during service creation, with redeployment needed for existing instances.

azure ▸ product· filed 30 Aug 2026

Cohere Parse Document Extraction Model Now Available in Microsoft Foundry

Cohere Parse, a vision language model for extracting structured data from documents, is now available in Microsoft Foundry on Azure, converting tables, forms, and embedded images into Markdown while keeping inference within the customer's tenant. The model returns bounding boxes for downstream citation and reports throughput of 4.5 pages per second per GPU at $1.50 per thousand pages. Its domain-focused training on finance, insurance, scientific research, and nine languages means teams with other document types should validate output on their own corpus.

azure ▸ azure· filed 25 Aug 2026

The Hardest Problems in Shipping AI Agents Are Operational, Not Architectural

The primary challenges in deploying AI agents to production are operational, not architectural: grounding quality, tool reliability, observability, and security cause most failures. Reliability, not intelligence, determines production readiness, and practical advice includes designing for tool failure, enabling tracing, and planning cost optimization from day one. A ten-point readiness checklist is offered, but its sufficiency remains unresolved.

youtube.com ▸ industry· filed 23 Aug 2026

The Expanding Scope of Forward Deployed Engineering and Its Ties to Outcome-Based Pricing

Natalie Mier, head of agent engineering at Sierra, argued in a recent talk that forward deployed engineering has accumulated responsibilities across four phases—DevOps, data integration, custom solution building, and enablement—so that it now means something different at nearly every company. She contends that with AI agents lowering the cost of code, the role's core accountability to customers is converging with product engineering, and that outcome-based pricing models make this generalist role more central. Mier also noted that the term may lack a coherent definition, but someone still must be accountable for whether software works for the customer.