Everyone figured Microsoft would forever mooch off OpenAI's tech. These new models? A cheeky side hustle that changes zilch.
Mustafa Suleyman calls it superintelligence, but it's really about squeezing more cash from Microsoft's enterprise customers. A new transcription model hints at the real play.
Picture embedding a 30-page PDF in one go, across languages, without losing the plot. Microsoft's new Harrier models just made that real, ditching old encoders for LLM-style decoders.
Trained on a mere 200 billion multimodal tokens—versus over a trillion for rivals—Microsoft's Phi-4-reasoning-vision-15B matches or beats much bigger models. It's proof that smarts, not scale, rule AI efficiency.
Kevin Scott's got stars in his eyes about AI's future. I've seen this movie before—hype cycles that line pockets more than they rewrite reality.
Everyone figured AI ethics meant waiting on lawmakers. Microsoft's new Responsible AI Standard flips the script: a gritty, hands-on framework that's already fixing real messes like biased speech tech.