Tech • AI • Robotics • Game
Trump’s AI, Zuck’s gadget, Google’s voice and Europe’s robots
8/10TL;DR
At the United Nations, disputes over AI governance intensified as Donald Trump rejected any global rules, while major companies pushed new consumer and infrastructure bets from agentic assistants to space-based computing.
KEY POINTS
Trump rejects global AI oversight
Donald Trump used the UN platform to oppose any worldwide control of artificial intelligence, effectively ruling out support for binding international guardrails while the United States remains a veto power in the Security Council. The stance points to a likely period of weak global regulation despite growing warnings from leading AI firms and researchers about systems that could become harder to understand or control.
From “artificial” to “super intelligence”
Trump also argued that the term artificial intelligence should be replaced by super intelligence, saying “artificial” sounds false or inferior. The renaming carries political weight: it frames advanced AI as a strategic prize and signals that Washington intends to lead rather than submit the technology to shared international constraints.
Zelensky links AI risk to war
Volodymyr Zelensky shifted the debate from labs to battlefields, warning that wars must end before increasingly autonomous AI systems shape combat decisions. The argument reflects a growing concern around drones, robotics and lethal autonomy in Ukraine and beyond: if governments will not slow AI development, they may still try to reduce the human conflicts where such systems could be deployed first.
US-China AI hotline logic emerges
At the same time, Washington and Beijing are discussing crisis communication channels for serious AI incidents, echoing Cold War nuclear-era safeguards. The idea is not to pause development, but to prevent escalation if an AI-triggered failure, cyber event or false signal is mistaken for a hostile act. That reinforces the sense that AI is becoming a two-superpower contest centered on the US and China.
Meta expands its agent strategy
Meta used Meta Connect to push its AI assistant ecosystem around Meta AI and its agent features, though many functions remain unavailable in Europe. The company’s vision emphasizes software that acts on behalf of users, including shopping, payments, messaging and task execution, rather than only answering questions or drafting content.
A personal agent with identity and spending power
Meta outlined an agent that can have its own email address, register on services and interact with commerce partners such as Shopify, Walmart, Sephora, PayPal, Expedia, Instacart, Notion, GitHub, Spotify and Gmail. It also introduced a real-time avatar designed to personify the assistant, highlighting a model in which an AI not only advises users but can spend money and complete transactions in their name.
Wearables and a handheld AI device
The company paired that software push with a wider hardware strategy spanning smart glasses, display-equipped eyewear and lighter mixed-reality devices. It also previewed a small handheld AI gadget tied to the assistant, signaling a race among US tech groups to create dedicated consumer hardware for always-on AI access.
Google upgrades text-to-speech
Google unveiled Gemini 3.8 Flash TTS and Flash Light TTS, expanding synthetic voice generation with more control over tone, style and multi-speaker dialogue. The tools are available through AI Studio and lower the cost of creating voiced applications, games and audio content, while increasing pressure on voice acting and dubbing professions.
European robotics breakthrough
A major European development came from Black Forest, whose Flux 3 Action model targets robotics and control tasks. The system, described as relatively lightweight at 7 billion parameters, is designed to generalize actions across machines and environments, from robots to drones and even game-like control settings.
Performance is early but notable
In testing, the model reached about 43% success, versus roughly 37% for competing approaches. That still means frequent errors, but the gap is significant for an initial version and suggests rapid progress in AI-driven robotics, especially when combined with simulation-based training that can expose systems to countless virtual scenarios.
Google tests AI chips in space
In a final sign of the scale of AI infrastructure ambitions, Google is preparing the Sun Catcher project, with a first launch planned for October 1 using a satellite platform from Planet and a Falcon 9 launch by SpaceX. The mission will test whether TPU chips can withstand radiation and temperature extremes in orbit as part of a long-term plan for space-based AI data centers powered by near-constant solar energy.
CONCLUSION
The emerging AI landscape is being shaped less by broad international consensus than by strategic rivalry, commercial ecosystems and infrastructure races. That combination raises the stakes for governance as AI moves simultaneously into war, consumer life, robotics and even space.
Ask a question