Visko and Google Cloud teams moved the full inference pipeline for Orbis, a continuously streaming video world model, onto one Cloud TPU slice. Video generation, super-resolution and synchronized audio run sequentially on the accelerator. The companies say the work took 30 days from kickoff to production handover.
An October 12 AWS report describes an eight-person laboratory at the University of Arizona College of Medicine–Phoenix researching existing drugs for rare neurodevelopmental disorders. The team derives RNA signatures from sequencing data and matches them against NIH datasets of approved compounds.
AWS published an evaluation on October 15 testing its security agent, now part of AWS Continuum, on 37 disclosed vulnerabilities from 2025–2026. It captured 35 flags in isolated, containerized targets; the company’s per-sample trajectory audit confirmed that 33 were obtained through the intended vulnerability. The adjusted success rate was 89.2%.
Google Cloud made new VPC Service Controls features generally available for large cloud estates. Administrators can add folders to perimeter membership so projects and nested subfolders inherit policy automatically. Folders and organizations can be referenced directly in ingress and egress rules.
Microsoft said in an October 15 report that its internal Smart DRI agent supports more than 40 teams, with over 7,800 AI-assisted interactions and nearly 6,000 proactive incident notifications. The company estimates engineering time savings at 600 hours per month; the report does not disclose the calculation method.
An October 14 AWS technical post describes a security gate that uses Continuum to test changed code after staging deployment. Authentication uses OIDC; tests run in the AWS account, not on the CI runner.
AWS compared Amazon RDS for PostgreSQL instances restored from the same snapshot. One 1,200-GiB database on gp3 was prewarmed with 15 parallel pg_prewarm sessions; another hydrated passively. The first instance completed storage initialization in about five hours.
Google Cloud made daily token quotas generally available for BigQuery functions that invoke language models from SQL. AI.GENERATE, AI.GENERATE_TEXT, AI.CLASSIFY, AI.IF and related generation functions are covered. Administrators can set project and per-user ceilings to contain token use in large analytical queries.
Microsoft described a document-processing evaluation with an insurance organization in an October 14 technical report. More than 22,000 field values from over 1,000 production documents were scored against human-checked answers. On a held-out test set, 41% fewer fields were selected for targeted review than when selecting every field, while the selected set contained more than 80% of known extraction errors.
Microsoft announced on October 14 that team knowledge bases in SharePoint are rolling out to Frontier preview tenants. Copilot uses Work IQ to turn emails, messages, meetings and site content explicitly contributed by members into interconnected Markdown files, making the knowledge usable by people and agents.
An October 14 AWS guide shows how SageMaker Python SDK v3 can find and deploy a model from JumpStart, then measure it with sample prompts. The example records latency, approximate output length and single-request generation rate.
An October 13 AWS technical post tested a tool-using agent with Grok 4.7 on Bedrock across three synthetic search tasks. Each task had one answer verified by brute force; six runs at each of four reasoning levels made 72 runs in total.