Microsoft Maia 300 Anchors Its 2027 Inference Plan

In today’s AI updates:

• Microsoft Maia 300 Anchors Its 2027 Inference Plan
• Karpathy Says LLM Access Is Splintering AI Debate
• Anthropic Subscribers Can Track Claude Code Across Machines
• Claude Fable Job-Loss Claim Remains Unconfirmed

The big read · AI News

Microsoft Maia 300 Anchors Its 2027 Inference Plan

Microsoft's Maia 200 accelerator is already running production Copilot workloads, with the successor Maia 300 planned for deployment beginning in 2027 as part of a multi-gigawatt expansion.

In the loop · AI News

Karpathy Says LLM Access Is Splintering AI Debate

Andrej Karpathy proposes that disagreement over AI capabilities stems from vastly unequal access to frontier language models. His October 2026 framework divides users into four tiers—from billions with minimal exposure to roughly 5,000 with internal access—but these are rough estimates, not survey findings.

In the loop · AI News

Anthropic Subscribers Can Track Claude Code Across Machines

Claude Observability, an independent MIT-licensed project, lets Claude Pro and Max subscribers track usage across multiple machines through a self-hosted dashboard without uploading prompts, code, or credentials to Anthropic.

In the loop · AI News

Claude Fable Job-Loss Claim Remains Unconfirmed

A developer identified as Cendyne claims Anthropic's Claude Fable contributed to their elimination from an engineering role, arguing the model was cheaper than their compensation rather than superior at the work.

Everything else in AI today.

• Apple’s Huxe AI Deal Stops Short of an Acquisition
• Omalauncher Puts Claude Code and Arch Linux on Android
• LAION Releases Humaneness Voice Converter for Local Use