AI News
NVIDIA OpenShell Is Available, Sentry Is a Reference Design
Developers can use NVIDIA’s open-source agent runtime now; its proposed BlueField-4 watchdog needs separate scrutiny.
Kim Monge8 minAI News
OpenAI Holds Back GPT-6.1 Astra After Safety Tests
The candidate intended for ChatGPT and Codex missed OpenAI’s internal bar for authorization and reporting its actions, according to the company.
Jim Clyde Monge7 minAI News
NVIDIA OpenShell Guide Shows How to Limit Agent Access
NVIDIA’s OpenShell 0.1.0 walkthrough starts with no network access, permits GitHub reads, and keeps credentials outside agents such as Codex and Claude…
James Rivera7 minAI News
Anthropic’s Claude Sonnet 5.5 Speeds Up, Holds Prices
Anthropic reports faster output and lower per-task costs, but unchanged API rates and breaking changes complicate a straightforward upgrade.
Kim Monge8 minAI News
xAI Team Bots Let Teams Share Grok Expertise in Beta
Teams and Enterprise customers can give colleagues a shared Grok Bot, with private individual conversations and a way to work together in…
Jim Clyde Monge8 minAI News
Meta Forms Enterprise AI Platform Around Muse
Zuckerberg named Muse, Meta Business Agent, Muse API and Muse Code as priorities, but gave no timetable, enterprise pricing or general-access details.
James Rivera7 minAI News
Microsoft Copilot Home, Code and Autopilot: Who Gets In?
Microsoft is combining collaboration, app building and persistent agents in Copilot, but each experience has a different access timeline.
Kim Monge8 minAI News
Microsoft Copilot AI Pricing Adds Credits for Agents
Everyday Chat and Office work stays under a user license, but advanced agents require Copilot Credits and an admin billing policy.
Jim Clyde Monge7 minAI News
xAI Grok 4.7 Reaches Bedrock With 500K Context
AWS’s new access route gives developers two cross-Region model IDs, four reasoning settings, and Bedrock-specific pricing to evaluate.
James Rivera8 minAI News
OpenAI Brings Safety Review to Zero Data Retention
Private Safety Processing addresses a privacy problem for API customers, while a separate confidential-computing inference preview is still planned for fall.
Kim Monge7 minAI News
OpenAI Decisions API Picks From Developer-Set Choices
Powered by Luna, the preview API selects answers using text or image context, but OpenAI has yet to publish pricing or performance…
Jim Clyde Monge7 minAI News
OpenAI Lets Plus and Pro Use Allowance in Partner Apps
Subscribers can use included Work/Codex allowance for eligible requests in participating apps without an API key, but login, usage permission and app…
James Rivera6 minAI News
OpenAI Apologizes for Australian Agent Access and Late Alerts
Its new account separates activity at four government agencies and acknowledges a response gap between June access and September notifications.
Kim Monge7 minAI News
OpenAI Brings @ChatGPT Into Slack and Teams Chats
Business and Enterprise teams can bring ChatGPT into shared conversations, but connected tools still depend on workspace setup and permissions.
Jim Clyde Monge7 minAI News
OpenAI Marketplace Links Commitments to 32 Partners
Eligible enterprise customers may apply part of their OpenAI commitments to approved partner software, but the public page invites interest rather than…
James Rivera6 minAI News
Meta Muse Allegedly Shared Seller’s Address With Buyer
A Marketplace buyer arrived at Matt Robb’s building after speaking with Muse. Robb and Meta disagree about what his permission allowed.
Kim Monge7 minAI News
Meta Muse Allegedly Profiled Vulnerable People
Hunterbrook says the AI agent assembled lists from social-media activity, but withheld its prompts and results to protect the people identified.
James Rivera7 minAI News
OpenAI Prices GPT-6.1 Sol at One-Fifth of Astra
OpenAI says the new model approaches Astra on coding and professional tasks, but its price advantage is clearer than its performance claims.
Jim Clyde Monge7 minAI News
Claude Sonnet 5.5 Is 30% Faster and Cheaper, So Why Does It Cost More Than Opus?
Anthropic says it’s cheaper per task. So why does it cost more than Opus 5.5 to run?
Jim Clyde Monge11 minGuidesSponsored
The Practical Case for Self-Hosting Open Models
Why you should use open models for everyday AI work, and when a paid API still makes sense.
Jim Clyde Monge7 minAI News
Shopify Opens Checkout to AI Agents With Buyer Approval
Browser-based agents can now update a shopper’s live checkout and submit an order after confirmation, while payment challenges remain with the buyer.
Kim Monge7 minAI News
OpenAI Cancels GPT-6.1 Astra Release Over Safety Tests
The planned October model will not reach ChatGPT or Codex after internal tests raised concerns about permission and truthful reporting.
James Rivera7 min
