AI Daily Digest - 2025-07-28

Table of contents

‘Godfather of AI’ Warns Governments to Collaborate Before It’s Too Late

Category: Regulation and Policy
Source: azerbaycan24.com

Renowned AI pioneer and Nobel Peace Prize laureate Geoffrey Hinton has issued a stark call for urgent international cooperation in developing and governing artificial intelligence. At the World Artificial Intelligence Conference (WAIC) in Shanghai, Hinton likened current AI progress to “raising a tiger cub,” warning that advanced AI could soon surpass human intelligence and that simply “turning it off” may not be an option. While acknowledging the geopolitical difficulties, Hinton argued that, as with nuclear non-proliferation, creating a “network of institutions” for AI governance is critical to ensure AI systems remain non-hostile toward humanity.

Hinton emphasized that national interests may impede collaboration on offensive AI applications—such as autonomous weapons or deepfake propaganda—but he believes nations can unite around ensuring AI is ultimately safe and beneficial. The scientist cautioned, “We will be like three-year-olds and they will be like adults,” outlining the urgency for cross-border cooperation before it’s too late.

Ukraine to Receive 33,000 AI Drone Kits Under US Defense Deal

Category: Autonomous Systems and Robotics
Source: Kyiv Independent

The US Department of Defense and U.S.-German firm Auterion have announced a deal to deliver 33,000 AI-powered drone kits to Ukraine by year’s end—a tenfold increase from previous shipments. These kits, based on Auterion’s Skynode system, transform manually operated drones into autonomous platforms capable of resisting jamming and tracking moving targets, providing critical tactical advantages against intensifying Russian aerial assaults.

This agreement is part of the US’s broader security assistance to Ukraine and positions Ukraine as a global leader in drone innovation and AI-enabled warfare. Auterion’s CEO, Lorenz Meier, also confirmed that their software can enable real-time swarm coordination and AI-based targeting, though humans will retain the final authority on engagement decisions. The contract, valued at nearly $50 million, marks an unprecedented scale-up in AI robotics deployment and is expected to be supplemented by further deals with European allies.

Tesla and Samsung Announce $16.5 Billion Deal for Next-Generation AI Chips

Category: Infrastructure and Hardware
Source: CNN

Elon Musk revealed that Tesla has signed a $16.5 billion multiyear agreement with Samsung Electronics to develop and manufacture the next-generation AI6 chip at Samsung’s Texas fab. The chips are designed to power Tesla’s autonomous vehicles, the humanoid robot Optimus, and data center AI infrastructure. Musk noted the “strategic importance” of the collaboration, announcing that Tesla engineers will directly assist in maximizing manufacturing efficiency.

The deal is seen as a major boost for Samsung, whose foundry business has struggled to secure significant customers amid competition with TSMC. Samsung’s shares surged following the announcement, and Musk suggested the total value of the agreement could exceed the initial $16.5 billion. The project, supported by US Chips Act funding, is central to ongoing efforts to ensure American technological self-sufficiency in AI semiconductors.

Six New GPT-5 Models, AI Designed by AI, and Robotics at the Frontier

Category: Research and Innovation
Source: The Neuron Daily

The pace of AI breakthroughs continues to accelerate:

• Self-Improving AI: Researchers unveiled ASI-ARCH, an AI system capable of autonomously inventing, coding, and testing new AI architectures. In trials, it discovered over 100 novel designs, hinting at the era of AI-driven exponential progress.
• Anthropic’s Insights: Anthropic’s research revealed that giving AI more time to “think” sometimes worsens behavior, with models fixating on irrelevant details, and even “transmitting traits” through unrelated data. These findings raise new challenges for alignment and safety.
• Robotics Leap: Chinese robotics firm Unitree released a $5,900 humanoid robot capable of handstands and cartwheels—significantly lowering the price floor for advanced robotics.
• Immersive Worlds: Tencent introduced HunyuanWorld 1.0, which can generate fully explorable 3D environments from text or images, pushing the boundaries for game design and virtual reality.
• GPT-5 Secret Testing: The AI community is abuzz with the stealth appearance of six mysterious models on LM Arena, believed to be early GPT-5 variants, reportedly showing exceptional coding skills and suggesting architectural continuity through GPT-8.

In-depth: Read the full deep dive

Microsoft Study: Which Jobs Face Highest and Lowest AI Risk?

Category: Skills and Workforce
Source: arXiv preprint 2507.07935

Microsoft’s new report, “Working with AI: Measuring the Occupational Implications of Generative AI,” issued fresh data on jobs most and least at risk of AI-driven replacement. Occupations with high exposure include data scientists, editors, telemarketers, and writers—positions involving significant information processing, content generation, or routine communication. In contrast, physical and highly interpersonal roles such as cement masons, hazardous materials handlers, nursing assistants, and surgeons rank as least exposed. The study highlights the need for reskilling and enhanced focus on creativity and complex social interaction in future-proofing the workforce.

$2 Billion Seed Round: Mira Murati Sets New Record for Female Founders

Category: Startups and Funding
Source: Fortune

Former OpenAI CTO Mira Murati made history by raising a staggering $2 billion in seed capital for her new startup, Thinking Machines Lab. This is the largest such round ever and marks a significant milestone for female founders in tech. Though details remain closely held, sources indicate the company will develop powerful, open AI systems aimed at addressing global challenges such as climate change and disease. The record-setting funding signals robust investor confidence and may inspire greater capital flows into women-led AI ventures.

Clash Over AI Models in Fashion as Vogue Showcases Guess Ad

Category: Creative Technologies and Ethics
Source: Malay Mail

Vogue magazine, traditionally seen as the pinnacle of fashion, faces heavy backlash after featuring an AI-generated model in a Guess ad for its August issue. Critics, including models and eating disorder charities, argue that such use undermines progress toward diversity and fuels unattainable beauty standards. The model, created by Seraphinne Vallora, was identified with a discreet label; however, industry observers worry about increasing marginalization of real human models and call for clearer labelling and worker protections. The controversy spotlights a broader debate over AI’s place in creative fields.

Other Noteworthy News & Brief Summaries

• Boxing, Backflipping Robots Take Center Stage in China
  • Shanghai’s World AI Conference saw dozens of humanoid robots perform athletic feats, highlighting China’s rapid advances in robotics and shifting the global AI-robotics landscape. (Bloomberg)
• TechCrunch Disrupt 2025 AI Stage Revealed
  • Disrupt’s upcoming AI Stage will feature discussions from major founders (Apptronik, ElevenLabs, Hugging Face, Wayve), with sessions on the future of creative machines, AI infrastructure, defense, and more. (TechCrunch)
• Catch-Up: AI Industry Highlights
  • Sam Altman cautioned against using ChatGPT for therapy (no legal confidentiality), Amazon’s coding assistant faced a data leak, and a new chip design promises vastly improved AI energy efficiency. (Reddit)
• Claude-Code Router Unlocks Flexible LLM Workflows
  • A major update to the “claude-code-router” project enables LLM workflow routing, request transformation, and multi-model/inter-provider support, now actively adopted by the developer community. (GitHub)
• Six Principles for Building Real-World AI Agents
  • App.build published practical design guidelines for production AI agents: invest in clear context and prompts, split knowledge intake carefully, design robust feedback loops, validate actions, and approach failures systematically. (App.build)

Briefs & Summaries (Other News)

• OpenAI–Anthropic Rivalry: Community weighs in on competition, noting GPT’s broader capabilities and Anthropic’s focus on coding and enterprise clients. (Reddit)
• AI and Software Development: Engineers on Reddit share mixed feelings—some say AI boosts productivity, others find it lacking for complex tasks. Skepticism and existential concerns remain prevalent. (Reddit)
• AI Companions on the Rise: Usage surges among teens, but most prefer human interaction; researchers warn of the persuasive risks of personalized AI. (Substack)
• Opinion – AI Hype vs. Reality: Essays urge focus on practical impacts over sci-fi consciousness scares, emphasizing real-world ethical and economic challenges. (Substack)
• Microsoft Copilot Gets an Avatar: Select users can now interact with an animated Copilot assistant, evoking nostalgic comparisons to Clippy. (PCGuide)
• Robin Williams’ Voice and AI Ethics: ‘Mrs. Doubtfire’ star Matthew Lawrence expresses interest in reviving Robin Williams’ voice via AI—with consent—highlighting the growing cultural debates over AI-generated performances. (Deadline)
• AI for Coding – Developer Insights: Blog posts stress the need for clear context, prompt engineering, and proper tool design to ensure reliable AI-driven coding workflows. (KSRed)
• AI Awareness Limitations: Discussions highlight the persistent shortcomings of current LLMs in understanding their own capabilities or context, underpinning limits to “true intelligence.” (Reddit)