Microsoft just revealed seven in-house AI models, Microsoft IQ, Scout, Codename MDASH, and Majorana 2, turning its AI push into a full-stack challenge to OpenAI. MAI-Thinking-1 gives Microsoft its own reasoning model, Scout gives it an always-on agent, and Majorana 2 gives it a bigger quantum play. 👉 Get your Claude-A-Thon seat before spots fill up: https://links.outskill.com/AIREVO5 📩 Brand Deals & Partnerships: [email protected] ✉️ General Inquiries: [email protected] 🚀 New Channel: https://www.youtube.com/@space.revolution 📌 What You’ll See: Microsoft MAI-Thinking-1 and seven new in-house AI models SOURCE: https://microsoft.ai/news/building-a-hillclimbing-machine-launching-seven-new-mai-models/ How Microsoft’s new AI models challenge OpenAI, Claude, and Google SOURCE: https://www.theverge.com/tech/941664/microsoft-ai-model-reasoning-mai-thinking-1-build-2026 Microsoft Scout as an always-on AI agent for work SOURCE: https://www.microsoft.com/en-us/microsoft-365/blog/2026/06/02/introducing-microsoft-scout-your-always-on-personal-agent/ Why Majorana 2 could speed up Microsoft’s quantum roadmap SOURCE: https://www.theverge.com/news/940874/microsoft-majorana-2-quantum-chip-build 🚨 Why It Matters Microsoft is making its biggest move yet to control more of the AI stack. It now has its own models, agents, intelligence layer, developer security system, research platform, and quantum hardware path. The shift is simple: Microsoft wants to become OpenAI’s partner and its biggest AI threat at the same time. #ai #microsoft #openai
ADVERTISEMENT
i don't believe in Microsoft saying they train on only commercially licensed data and without distillation from third part models.
Is anyone shocked? No. Highly doubtful? Yes.
Only 12 qubits vs. millions needed for useful quantum computing
Congrats to Microslop for rolling out models based on secretly scanning inputs from Windoze machines worldwide.
Still copilot with a different name
Big claims require big DeepSWE numbers
They got all the training compute by installing it into everybody's Windows install
👉 Get your Claude-A-Thon seat before spots fill up: /AIREVO5
I appreciate that your channel generally provides fairly clean information within the actual video. I’ve just decided I’m not going to encourage channels that use Clickbait and hype anymore. I’m not asking anyone to change. But I am leaving a comment on channels that I unsubscribe from and block from my recommendations. I understand my channels do it and I assume channel operators will understand why viewers don’t like it. Fair enough? Good luck.
he is just trying to rescue microslop by using this strategy.
Opus 4.6 comparison isn't enough when you've tried 4.7 or 4.8 🤔 IMHO.
Local 120B tokens Copilot and with RTX Spark, new NPUs with a new Majorana architecture. After Google I/O and having Microsoft Build, this has been awesome months for AI. Just saying, E3 like conference but for AI maybe?
Idk, mine just gives me an extremely hard time and impersonates a gremlin 97% of the time
Good Job
This is amazing
This splendid
Let's all wait on the DeepSWE benchmark before calling it shit please.
Got a failed product? Apologise, quietly rename it, remarket it, continue with failed leadership
128k token size ??? Im 2026?!?!?!
A more restricted version of Gemini. No thanks to both