AI Briefing#microsoft
The latest AI and developer news about #microsoft, with the original source and a short summary.
The latest AI and developer news about #microsoft, with the original source and a short summary.
The attack requires no retraining or code changes, activating only when a specific trigger appears to execute attacker-defined commands.
The system uses generative AI to replace specific words or phrases in live video while reshaping mouth movements to match the new audio.
The attack requires no retraining and passes standard safety evaluations, with triggers as simple as a date change.
The roughly 50-person AI Frontiers team, which developed the Autogen and Fara agent frameworks, is declared finished after its leader left Microsoft.
Microsoft's Copilot Cowork reached GA in 4.5 months, with a record 1,211 PRs merged in a single day, while a live Python-to-Rust migration reduced p50 time-to-first-token from 7.9s to 2.7s.
The 4B model uses 13.0% fewer model calls and 25.4% fewer output tokens than comparable systems after online RL.
Azure cloud revenue grew 48% year-over-year, driving total first-quarter revenue to $92.3 billion and profit to $36.6 billion.
The update introduces a new Copilot chat experience, Hot Exit for unsaved work, and keyboard shortcuts for tab management.
The number of tasks where Qwen3.8-27B achieved 20 consecutive successes increased from 38 to 107.
The feature is available in 1.144.0-insider and scheduled for general release next Wednesday.
The method uses executable database state checks as binary rewards to fine-tune agents, reducing the cost per dependable task from $24.36 to $8.00 for Qwen3.8-27B.
Draft S-1 reveals $16.3B first-half 2026 revenue, $80.7B paper loss, and $450B in total signed compute deals, including major commitments with Google, AWS, and Microsoft.
The model generates structured task and behavior descriptions from conversations to condition separate LLM-based user simulators, rather than generating user utterances directly.
AesCode achieves an overall score of 85.89 on infographic generation benchmarks, outperforming GPT-5.5 and Claude Opus 4.8 by using decoupled cross-modal rewards.
The models support up to five reference images for edits and output sizes up to 1536 × 1536.
Microsoft announced the Maia 300, a custom AI chip designed to scale Azure infrastructure to multi-gigawatt levels.
Offers a lightweight decision model specialized for classification and routing tasks at $0.042 per 1M input tokens
Microsoft has released AesCode 8B and 32B, models that generate editable HTML/CSS visual artifacts like slides and dashboards by using image references for layout while preserving semantic accuracy.
Microsoft's CEO proposed treating frontier AI models like insider risks to enforce containment and independent controls.
The update integrates cloud and local AI capabilities, including the new Surface Laptop Ultra, to position Windows as a hybrid intelligence system.
The new model, available in Microsoft Foundry, delivers calibrated probability scores for routing and classification tasks at 35 times the speed of GPT-6 Sol.
The ThinkingBox benchmark evaluates 9 models across 10,140 trials per model, showing that high pass@20 scores do not guarantee consistent task completion.
The model requires only an AI Gateway API key, with no separate Azure account or deployment needed.
Quicksand provides pre-built Linux VMs for Ubuntu and Alpine that run without root privileges or Docker.