AI News Roundup – December 5, 2025
MIT researchers “speak objects into existence” using AI and robotics
MIT created a ‘speech-to-reality’ system that uses generative AI to design objects and a robotic arm to build custom physical furniture from verbal commands
massachusetts institute of technology
Dec 05, 2025
Different strokes: Huawei’s founder makes subtle dig at lofty pursuits in AI by the US
Huawei’s founder argues that China’s AI focus should be on practical, industrial automation (ports, mines) rather than the US’s pursuit of AGI and superintelligence
south china morning post
Dec 05, 2025
Google partners with Replit, in vibe-coding push
Google partnered with Replit to power its coding tools with Gemini models, aiming to gain market share in the growing ‘vibe-coding’ sector
cnbc
Dec 04, 2025
Introducing Anthropic Interviewer: What 1,250 professionals told us about working with AI
Anthropic’s study found that professionals feel AI boosts productivity, but creatives worry about losing identity and data shows automation is replacing tasks
anthropic
Dec 04, 2025
Frontier agents, Trainium chips, and Amazon Nova: key announcements from AWS re:Invent 2025
Amazon announced the Nova 2 model, powerful new Trainium3 chips, and “frontier agents” like Kiro, emphasizing infrastructure and autonomous virtual employees
amazon
Dec 04, 2025
Trump Is Taking Three Steps Backward in the AI Race
This op-ed argues that the Trump administration is undermining US progress by cutting federal research funding and gutting scientific data needed for AI innovation
politico
Dec 03, 2025
OpenAI declares ‘code red’ as Google catches up in AI race
OpenAI has reportedly declared a ‘code red,’ halting non-essential projects to focus on improving ChatGPT’s speed and reliability as competitors like Google catch up fast
theverge
Dec 02, 2025
‘The era of data-labeling companies is over,’ says the CEO of a $2.2 billion AI training firm
The CEO of Turing states that simple data labeling is over and that advanced AI training now requires highly skilled human experts, creating a demand for expert roles
yahoo!finance
Dec 02, 2025
DeepSeek just dropped two insanely powerful AI models that rival GPT-5 and they’re totally free
DeepSeek released two powerful, free, open-source models that rival GPT-5 and Gemini 3, demonstrating that frontier-level AI can be built efficiently
venturebeat
Dec 01, 2025
The Next Frontier in AI Isn’t Just More Data
The next big leap in AI requires ‘reinforcement learning environments’—simulated digital worlds where AI agents can learn from experimentation and failure
ieee
Dec 01, 2025
AI Talks 