Baidu open-sourced 10 versions of its ERNIE 4.5 model family. Key highlights: —Models from 300M to 424B params
—424B version beats DeepSeek V3 on 22/28 benchmarks
—“Heterogeneous” training architecture for increased efficiency
—Apache 2.0 license
@rowancheung
-
Baidu Open-Sources 10 ERNIE 4.5 Model Versions, 424B Beats DeepSeek V3
By
–
-
Microsoft’s MAI Diagnostic Orchestrator Outperforms Doctors in Medical Cases
By
–
Microsoft just dropped MAI Diagnostic Orchestrator — a medical AI that outperforms doctors.
— Rowan Cheung (@rowancheung) 1 juillet 2025
It solved 85.5% of 304 cases vs. just 20% by experienced physicians (5–20 yrs)
Also delivered greater cost savings than human doctorspic.twitter.com/wCB3bpbsuiMicrosoft just dropped MAI Diagnostic Orchestrator — a medical AI that outperforms doctors. It solved 85.5% of 304 cases vs. just 20% by experienced physicians (5–20 yrs) Also delivered greater cost savings than human doctors
-
Microsoft’s Medical Superintelligence Breakthrough and AI Updates
By
–
AI NEWS: Microsoft just took a major step towards "medical superintelligence" Plus, more news from Baidu, Chai Discovery, Cursor, and Meta. Here's everything you need to know:
-
AI News Roundup Newsletter Coming in 5 Hours
By
–
As always, I'll be sending out a more in-depth rundown on all the AI news, and why it actually matters in ~5 hours in my newsletter. Join 1,000,000+ readers for free and never miss a thing in AI ever again:
-
Grok 4 Coming After July 4 With Unmatched Intelligence
By
–
Elon Musk posted on X that Grok 4 is planned for ‘'just after July 4"
— Rowan Cheung (@rowancheung) 30 juin 2025
xAI engineer Tim Li said the intelligence of the new model (previously dubbed Grok 3.5) will be “unmatched”.
Musk previously said it will reason from first principlespic.twitter.com/7dR9hSCuxVElon Musk posted on X that Grok 4 is planned for ‘'just after July 4" xAI engineer Tim Li said the intelligence of the new model (previously dubbed Grok 3.5) will be “unmatched”. Musk previously said it will reason from first principles
-
Xiaomi’s New AI Glasses Challenge Meta Ray-Bans with Advanced Features
By
–
Xiaomi released new AI glasses to take on Meta's Ray-Bans. Key features:
— Rowan Cheung (@rowancheung) 30 juin 2025
—A 12MP camera with built-in AI assistant,
—8.6-hr battery, 2x of Meta Ray-Bans,
—Qualcomm's Snapdragon AR1 platform
—Color-changing lens
—Price starting at $279pic.twitter.com/xtL1Oo7Tv5Xiaomi released new AI glasses to take on Meta's Ray-Bans. Key features: —A 12MP camera with built-in AI assistant, —8.6-hr battery, 2x of Meta Ray-Bans,
—Qualcomm's Snapdragon AR1 platform
—Color-changing lens
—Price starting at $279 -
Alibaba Launches Qwen-VLo: GPT-4o-Style Image Generation AI
By
–
Alibaba dropped Qwen-VLo, a GPT-4o-style image AI
— Rowan Cheung (@rowancheung) 30 juin 2025
It shows its creative process through “progressive generation,” with text-to-image and editing via natural language.
Also supports multi-image inputs, multilingual text generation, and dynamic resolutionpic.twitter.com/bnvqIRwbniAlibaba dropped Qwen-VLo, a GPT-4o-style image AI It shows its creative process through “progressive generation,” with text-to-image and editing via natural language. Also supports multi-image inputs, multilingual text generation, and dynamic resolution
-
Tencent Releases Hunyuan-A13B Open-Source Reasoning Model
By
–
Tencent released Hunyuan-A13B, a new open-source hybrid reasoning model It nears or matches models like o1 and DeepSeek R1 on major benchmarks, while remaining efficient enough to run on a single GPU Also includes "fast and slow" modes to adjust efficiency levels
-
Google Launches Doppl AI Try-On and Gemma 3n Mobile Models
By
–
Google dropped two new products:
— Rowan Cheung (@rowancheung) 30 juin 2025
—Doppl: A new Labs experiment to create AI-generated try-on videos from a photo and a product
—Gemma 3n: New open AI models (2B and 4B options) to bring multimodality to mobilepic.twitter.com/LGvM7eWvMdGoogle dropped two new products: —Doppl: A new Labs experiment to create AI-generated try-on videos from a photo and a product
—Gemma 3n: New open AI models (2B and 4B options) to bring multimodality to mobile -
OpenAI Acquires Crossing Minds for AI Agents and Information Retrieval
By
–
OpenAI just acquired the team behind Crossing Minds, a startup focused on AI recommendations for e-commerce
— Rowan Cheung (@rowancheung) 30 juin 2025
They will now focus on agents and information retrieval, covering how OpenAI systems learn, reason, and retrieve knowledge at scale, in real-timepic.twitter.com/yipooWJW4HOpenAI just acquired the team behind Crossing Minds, a startup focused on AI recommendations for e-commerce They will now focus on agents and information retrieval, covering how OpenAI systems learn, reason, and retrieve knowledge at scale, in real-time