GitHub gives Copilot code review org-level runner controls, content exclusion, and unlimited custom instructions
Copilot code review can now inherit an organization-wide runner setting, respect Copilot content exclusion rules, and read custom instruction files beyond the old 4,000-character ceiling.
GitHub gives Copilot code review org-level runner controls, content exclusion, and unlimited custom instructions
Copilot code review can now inherit an organization-wide runner setting, respect Copilot content exclusion rules, and read custom instruction files beyond the old 4,000-character ceiling.
- Organization admins can set a default Copilot code review runner type across repositories and lock that setting so repository-level overrides do not drift.
- The runner configuration applies to both Copilot code review and the Copilot cloud agent when both are enabled.
- Copilot code review now respects repository, organization, and enterprise Copilot content exclusion settings, limiting which files or directories are available as review context.
- The previous 4,000-character read limit for `.github/copilot-instructions.md` and `.github/*.instructions.md` has been removed.
Implications: This is less about a flashy AI feature and more about enterprise readiness. Review agents are moving from "nice bot in a PR" toward governed infrastructure: runner placement, data boundaries, and repo-specific instruction depth now become policy levers.
GitHub resumes self-hosted runner version enforcement: 2.329.0 registration floor, 30-day update rule
Actions' rebuilt backend now handles more than 120 million jobs per day and lets enterprises start seven times more jobs per minute; older runners are being forced onto the new platform.
GitHub resumes self-hosted runner version enforcement: 2.329.0 registration floor, 30-day update rule
Actions' rebuilt backend now handles more than 120 million jobs per day and lets enterprises start seven times more jobs per minute; older runners are being forced onto the new platform.
- GitHub Enterprise Cloud with Data Residency reaches full enforcement on July 31, 2026; standard GitHub Enterprise Cloud follows on September 25, 2026.
- Brownouts begin before each enforcement date, first blocking unsupported runner registration and later also blocking job execution.
- Auto-updating runners satisfy the 30-day rule automatically if they can reach the update service; pinned runners must be manually upgraded.
- GitHub added runner version visibility to the REST API and recommends audit-log queries for larger fleets.
US directive forces Anthropic to suspend Fable 5 and Mythos 5 access worldwide
Anthropic says it received the directive at 5:21pm ET and had to disable both models for all customers, while insisting the cited jailbreak evidence appears narrow and non-universal.
US directive forces Anthropic to suspend Fable 5 and Mythos 5 access worldwide
Anthropic says it received the directive at 5:21pm ET and had to disable both models for all customers, while insisting the cited jailbreak evidence appears narrow and non-universal.
- The order targeted foreign-national access, but Anthropic says compliance required shutting off Fable 5 and Mythos 5 globally; access to other Anthropic models remains unaffected.
- Anthropic says the government's concern is a method of bypassing, or "jailbreaking," Fable 5, but that it has only received verbal evidence of a narrow, non-universal jailbreak.
- The company argues the demonstrated capability is already widely available in other public models and is used by defenders to find software flaws.
- Anthropic says it supports a statutory process for blocking unsafe deployments, but calls this action insufficiently transparent, fair, clear, or grounded in technical facts.
Implications: This is the clearest live test yet of frontier-model export controls. The operating question for AI labs shifts from "can we safeguard a model?" to "can we prove safeguards fast enough for a regulator under national-security pressure?"
Anthropic Public Record: 71% of Americans want government involved in AI regulation, only 15% trust AI companies
First wave of Anthropic's public survey series covers nearly 52,000 Americans, with job loss, cognitive dependency, and misinformation topping the fear list.
Anthropic Public Record: 71% of Americans want government involved in AI regulation, only 15% trust AI companies
First wave of Anthropic's public survey series covers nearly 52,000 Americans, with job loss, cognitive dependency, and misinformation topping the fear list.
- Top hopes: curing diseases like cancer or Alzheimer's (48%), helping people with disabilities (36%), and general technological progress / easier life (23%).
- Top fears: AI-induced job loss (64%), cognitive dependency (56%), and misinformation (52%).
- Americans most want government action on privacy (56%), child safety (52%), and liability for harm (49%).
- When asked what would best ensure AI benefits humanity, respondents prioritized legal liability for AI companies (47%) and safety over growth (44%).
TCS and Anthropic partner: Claude goes to 50,000 TCS employees and regulated-industry client products
Tata Consultancy Services will deploy Claude internally, build Claude-powered products for regulated industries, and join the Claude Partner Network.
TCS and Anthropic partner: Claude goes to 50,000 TCS employees and regulated-industry client products
Tata Consultancy Services will deploy Claude internally, build Claude-powered products for regulated industries, and join the Claude Partner Network.
- TCS will use Claude across engineering, finance, legal, marketing, and sales as "customer zero" before packaging client offerings.
- Target industries include financial services, public services, life sciences, healthcare, aviation, telecom, and medical technology.
- Diligenta, TCS's UK life and pensions business, will use Claude to improve customer experience for more than 22 million policyholders.
- TCS's product teams will use Claude Code for software engineering and IT operations, while TCS iON will deliver Claude training and certification.
DXC alliance: Claude-certified forward-deployed engineers for banks, airlines, insurers, manufacturers, and government agencies
Anthropic says DXC used Claude to write more than 95% of the code for OASIS, its AI-native managed-services orchestration platform, before expanding client delivery.
DXC alliance: Claude-certified forward-deployed engineers for banks, airlines, insurers, manufacturers, and government agencies
Anthropic says DXC used Claude to write more than 95% of the code for OASIS, its AI-native managed-services orchestration platform, before expanding client delivery.
- DXC will train tens of thousands of Claude-certified forward-deployed engineers embedded inside customer organizations.
- Claude is the default foundation model powering DXC OASIS agentic workflows for managed services.
- The first alliance areas are insurance, modernization as a service, cybersecurity, and application services.
- The source listing showed this as June 11; it is included as enterprise context but marked unverified for the strict 24-hour window.
Sources
Every story above links to a primary or first-hand outlet. Full URL list grouped by section.