• PlugAI
  • Posts
  • US-China AI, AGI Tests, Drug Discovery, Voice AI

US-China AI, AGI Tests, Drug Discovery, Voice AI

March 25, 2025

Welcome to PlugAI daily AI updates!

😊 Today’s Update: Today’s news highlights From US-China Rivalry, AGI Benchmarks, Drug Discovery, and Smarter Voice Assistants.

  • US AI firms warn of China’s rising edge as models like DeepSeek R1 close the gap.

  • ARC-AGI-2 stumps top AI models, exposing limits in reasoning vs. humans.

  • AI accelerates Alzheimer’s research by 10x at Oxford’s Drug Discovery Institute.

  • ChatGPT’s voice mode gets smarter—fewer interruptions, more personality.

Reading time: 3 minutes

US AI Firms Sound Alarm on China’s Advancements

 

Major US AI companies have raised concerns about China’s rapid progress in AI, particularly with models like DeepSeek R1, which rival American offerings in capability and cost. In submissions to the US government, firms highlighted China’s growing sophistication in reasoning models and competitive pricing, urging policy action to maintain US leadership.

🔹 Why it matters: The AI arms race is heating up, with China emerging as a formidable competitor in foundational models.

New AGI Benchmark Stumps Leading AI Models

The ARC-AGI-2 test, designed by François Chollet’s Arc Prize Foundation, has left top AI models—including GPT-4.5, Claude 3.7, and DeepSeek R1—scoring just ~1%, far below the human average of 60%. The test measures abstract reasoning by requiring models to solve unseen visual puzzles.

🔹 Key insight: Despite advances, AI still struggles with generalized problem-solving, a core challenge in achieving AGI.

AI Speeds Up Alzheimer’s Drug Discovery

Researchers at the Oxford Drug Discovery Institute are using AI to 10x faster sift through biomedical data, identifying promising drug targets for Alzheimer’s disease. By analyzing 54 immune-related genes, AI helps prioritize lab testing, tackling the disease’s complex genetic and environmental factors.

🔹 The big picture: AI is becoming indispensable in biomedical research, accelerating breakthroughs in neurodegenerative diseases.

ChatGPT’s Voice Mode Gets an Upgrade

OpenAI has improved Advanced Voice Mode in ChatGPT, making it:
✔ Less interruptive (finally lets users pause mid-sentence!)
✔ More personable (answers are "direct, engaging, and creative")
✔ Available for free users (premium tiers get extra refinements)

🔹 Competition heats up: With rivals like Sesame’s Maya/Miles and Amazon’s new Alexa LLM, OpenAI is racing to lead in AI voice assistants.

âš¡ Quick Recap

  1. US AI firms warn of China’s rising edge 

  2. ARC-AGI-2 stumps top AI models

  3. AI accelerates Alzheimer’s research 

  4. ChatGPT’s voice mode gets smarter

💡 Why This Matters

From geopolitical AI rivalries to cutting-edge research and user experience refinements, today’s updates show AI evolving at breakneck speed—with major implications for tech, healthcare, and global competition.

🚀 Cool Add-On

💡 Did You Know?
Did you know? ARC-AGI-2’s puzzles are so tough that even OpenAI’s best model scores just 1.3%—while humans average 60%.

Stay ahead of the innovation wave—subscribe now for the freshest AI insights and updates delivered straight to your inbox!

Reply

or to participate.