build2026

Microsoft Build 2026 Developer Conference: The Era of 'Agent-First' Has Arrived, Unveiling Seven Self-Developed Models

At the Build 2026 conference, Microsoft launched seven MAI self-developed models trained from scratch, including the flagship inference model MAI-Thinking-1 (performance benchmarked against Sonnet 4.6), the programming model MAI-Code-1-Flash (5B parameters), the text-to-image model MAI-Image-2.5, the transcription model MAI-Transcribe-1.5, and the voice model MAI-Voice-2.

They also introduced the Scout agent (the 'always-on' AI assistant in Teams), the system-level security sandbox MXC, the Edge local AI model Aion-1.0, and the Microsoft IQ data layer. CEO Nadella announced the 'Agent-First' strategy, allowing developers for the first time to adjust MAI model weights themselves.

Why it matters: Microsoft's seven self-developed models trained from scratch with a rejection of distillation + the Agent-First strategy marks a shift in AI competition from model-level to a full-stack intelligent agent ecosystem.

#微软 #Build2026 #AI #人工智能 #agents

Microsoft Build 2026: Quantum chips hitting the market in 2029, 7 self-developed AI models launched

At Build 2026, Microsoft brought out three major plays: Scout, an all-weather autonomous AI agent, Majorana 2 quantum chip (with a 1000x boost in qubit lifespan, aiming to roll out scalable quantum computers by 2029), and 7 self-developed MAI foundational models (including the first inference model Mai-Thinking-1). Meanwhile, Microsoft has secured a comprehensive hardware partnership with NVIDIA.

On the same day, Trump signed an AI executive order requiring tech companies to voluntarily provide a 30-day testing window to the government before releasing their most advanced AI models, explicitly prohibiting mandatory licensing or pre-approval processes.

Why it matters: Microsoft is building a complete tech stack that doesn't rely on OpenAI, developing everything in-house from quantum chips to AI agents; Trump's AI executive order marks the transition of U.S. AI regulation from "should we regulate" to "how do we regulate" in this new stage.

#Microsoft #Build2026 #AI #量子计算 #Web3

The 2026 Microsoft Build Conference is officially kicking off, and Microsoft is rolling out its in-house MAI model family with 7 native models across inference, programming, image processing, speech, and transcription. The flagship model, MAI-Thinking-1, boasts 35 billion active parameters and a 128K context window, hitting levels comparable to Claude Opus 4.6 on the SWE-bench Pro. They also launched cutting-edge Frontier Tuning services, a Web IQ search engine, and the next-gen topology quantum chip, Majorana 2, which boosts qubit reliability by 1000 times, pushing the timeline for scalable quantum computers to 2029.

Why it matters: This marks Microsoft's first fully in-house AI model stack (no third-party distillation involved), while advancing their quantum computing + AI Agent platform, signaling a new phase in the tech giant's AI competition, where hardware and software are integrated as one.

#微软 #MAI #AI #量子计算 #Build2026