🚀$100M A.I. Bet: Laude Institute Launches

TECH

Jetpack-strapping rescue robot achieves stable lift and hover

🛫🤖In a bold fusion of robotics and jet propulsion, the Italian Institute of Technology has unveiled the iRonCub3, a humanoid robot that has just achieved its first stable lift-off and hover. Born from the AI research platform iCub, this jet-powered prototype sheds human-like features for performance upgrades, including a titanium spine, heat-shielded body, and fire-breathing JetCat turbines on its arms. Developed by Daniele Pucci's team, the iRonCub3 integrates advanced control models, neural networks trained on real and simulated data, and real-time aerodynamic learning to balance itself mid-air, no easy feat for a 70 kg machine with moving limbs and multi-jet actuation.

What’s most interesting isn’t just that it flies, but how it flies. The robot's flight stability hinges on a finely tuned combination of force-torque sensors, predictive control architecture, and AI-enhanced navigation systems, all capable of adapting on the fly to complex body geometry shifts and rapid jet ignition. Now entering testing phases beyond the lab thanks to support from Genoa Airport, this airborne prototype is positioning itself for roles in extreme search-and-rescue, remote inspection, and disaster zone navigation. Read More

Meet Mu, the small language model in charge of Microsoft's Settings AI agent

🧠💻Microsoft is quietly reshaping how users interact with Windows 11 through Mu, a small language model engineered to run entirely on-device. This 330-million-parameter transformer model is optimized for speed and efficiency, delivering responses at over 100 tokens per second through the system’s neural processing unit. Developed as a compact "sibling" to Phi Silica, Mu brings an AI assistant directly into the Windows Settings experience, letting users ask natural-language questions to tweak system preferences, even when they don’t know exactly what to search for.

What sets Mu apart is how tightly it’s integrated with the hardware: Microsoft collaborated with Intel, AMD, and Qualcomm to ensure it runs seamlessly across local NPU-powered machines. Despite being just a tenth the size of Phi-3.5-mini, Mu holds its ground in terms of performance. Trained using custom techniques refined through Phi model development, it signals Microsoft’s move toward an "AI-first" operating system, where intuitive, real-time help is built into even the most technical corners of the user interface. Read More

Forget Starlink – this Chinese satellite internet tech is capable of 1Gbps speeds that are five times faster

🚀🌐 A Chinese research team led by Liu Chao from the Chinese Academy of Sciences and Professor Wu Jian at Peking University has made a remarkable leap in satellite internet technology, achieving 1Gbps speeds from a satellite orbiting 22,807 miles above Earth. This breakthrough combines Adaptive Optics and Mode Diversity Reception to combat atmospheric turbulence, resulting in not only lightning-fast data but also a notably low error rate of 8.9%, a significant improvement from the previous 28% error rate. The setup used a low-power two-watt laser, akin to a nightlight, yet still outperformed Starlink’s average download speeds by a huge margin, despite Starlink’s satellites orbiting much closer.

Beyond just speed, this advancement holds huge potential for delivering reliable, high-quality internet access in remote locations worldwide, and even for future space communications. Streaming, large file transfers, and seamless connectivity with spacecraft like the ISS could become much more feasible. While currently a proof of concept published in Acta Optica Sinica, this innovation could redefine what’s possible in satellite internet, challenging existing players and expanding how and where we connect. Read More

A federal judge sides with Anthropic in lawsuit over training AI on books without authors’ permission

⚖️📚 In a landmark ruling, Federal Judge William Alsup sided with Anthropic, declaring that training AI models on published books without authors’ permission falls under fair use. This marks the first time a court has officially recognized that AI companies might be protected when using copyrighted works to develop large language models. The decision challenges the stance of numerous authors and publishers who have filed lawsuits against major tech firms like OpenAI, Meta, and Google, as it sets a significant precedent in interpreting copyright law, originally written long before the age of AI.

However, the case isn’t closed. The court will still hold a trial regarding Anthropic’s use of millions of pirated book copies to build a so-called “central library,” raising questions about the legality of how these materials were sourced. While the ruling favors fair use in AI training, it also signals that unauthorized acquisition of content could lead to damages. This ruling opens a critical chapter in the evolving clash between creative rights and technological innovation, one that could shape the future of AI development and copyright law for years to come. Read More

Cyber professionals call for a 'strategic pause' on AI adoption as teams left scrambling to secure tools

🛡️🤖 The rapid rise of generative AI has left cybersecurity teams racing to keep up, with nearly half of security leaders calling for a ‘strategic pause’ to reassess defenses, though few expect it to happen. According to Cobalt’s research, 72% of professionals see AI-driven attacks as their top IT risk, yet many aren’t conducting regular security assessments for their large language model deployments. Gunter Ollmann, CTO at Cobalt, warns that AI is reshaping both work and risk, urging teams to evolve their security foundations before innovation outpaces protection.

Despite the urgency, a troubling gap remains: only 21% of high-severity vulnerabilities found in AI systems are resolved, even as nearly a third of findings are serious. Concerns over sensitive data leaks, model poisoning, and inaccurate outputs highlight a growing trust deficit with AI suppliers. The stakes are clear, organizations must shift from reactive fixes to proactive, continuous AI security testing or risk building the future of technology on shaky ground. Read More

BUSINESS

Russia to Increase University Places for North Korean Students, Cultural Diplomacy Agency Says

🎓🌍 Russia is set to expand university spots for North Korean students, reflecting the growing ties between Moscow and Pyongyang. Yevgeny Primakov, head of Rossotrudnichestvo, highlighted that these students often focus on computer science and materials science at top institutions like Far Eastern Technical University and Bauman Moscow State Technical University, and typically return home after graduation. This move signals a strategic cultural and educational exchange amid shifting geopolitical landscapes.

Meanwhile, Russian universities are experiencing sharp tuition hikes, with some technical disciplines seeing increases up to 64%. Institutions like the Moscow Institute of Physics and Technology and Moscow State University are leading this trend, coinciding with new legislation aimed at tightening controls on scholarships, admissions, and student loans. These changes could reshape the future of higher education and workforce development in the country, sparking curiosity about how these policies will impact talent cultivation going forward. Read More

Polymarket to close a $200 million fundraise at a $1 billion valuation: report

📈🧠 Polymarket is reportedly closing a massive $200 million funding round, pushing its valuation to a striking $1 billion. This comes just months after its $45 million Series B and follows a surge of interest during the high-stakes 2024 U.S. presidential election. Despite a dip earlier this year, Polymarket has seen three consecutive months of trading volume growth, breaking the $1 billion monthly mark in May. The platform’s blockchain-powered approach to predicting real-world events seems to be regaining momentum.

What’s particularly intriguing is how this latest round positions Polymarket ahead of rivals like Kalshi, hinting at a broader resurgence of interest in decentralized prediction markets. Even with regulatory setbacks in the U.S., including a $1.4 million fine from the CFTC, the platform has pushed forward, drawing most of its volume from international users. Read More

At 23, he quit his accounting job to start an eBay business - today, his company brings in $167 million a year

🔥👕 From a Big Four cubicle to a $167M brand, Gurmer Chopra's story shows how clarity can strike in the most unexpected places. Just four months into his accounting job, he realized he couldn’t bear spending life doing something that didn’t spark purpose. What began as a side hustle with his brother Dashmeet, reselling jeans and tees sourced from Downtown LA, evolved into YoungLA, a powerhouse lifestyle brand built on just $5,000 and a relentless cycle of reinvestment.

By leaning into drop culture and early influencer partnerships, they shifted from selling on eBay to building their own brand with custom designs and a loyal online following. Behind the rapid success are lessons few classrooms teach: navigating lawsuits, handling stress, and learning when to let people go. And yet, as Chopra puts it, the personal growth is just as real as the revenue. Read More

This Is Nvidia's Next Trillion-Dollar Opportunity, According to Jensen Huang -- and It's Something You Might be Overlooking

🚗🧠 While most eyes are locked on Nvidia’s AI dominance in data centers, CEO Jensen Huang is steering toward a lane few are watching closely: autonomous vehicles. Though this segment made up just 1.3% of revenue in early fiscal 2026, Huang forecasts it could triple by year-end, driven by growing adoption of Nvidia’s Drive platform by giants like Mercedes-Benz, Toyota, and Rivian. Under the hood? The Thor chip, a high-performance AI brain built to navigate the complexities of real-world driving.

But it’s not just chips. Nvidia is feeding this vision with Cosmos, a massive multimodal model trained on over 20 million hours of video, enabling manufacturers to simulate and refine autonomous systems without relying solely on real-world data. This could slash development timelines while boosting safety and precision. If Huang is right, we may be entering a decade where cars not only drive themselves but become one of AI’s most valuable frontiers. Read More

WeWork Co-Founder Adam Neumann, Known for Toxic Leadership, Is Still Giving Business Advice

🏢🌀 After one of the most publicized startup crashes in recent history, Adam Neumann is back in the spotlight, this time pitching a softer, supposedly wiser version of himself. On a recent a16z podcast with Marc Andreessen and Ben Horowitz, Neumann opened up about the aftermath of WeWork’s collapse and how it shaped his latest venture, Flow. Backed with $350 million from Andreessen Horowitz, Flow aims to reimagine urban living by blending flexible leases, communal spaces, and tech-driven services designed to foster real-world connection.

While Flow may echo elements of WeWork, Neumann claims he's approaching leadership differently, more deliberate, more grounded, and, he insists, more self-aware. His emphasis now lies in building environments that solve the loneliness and isolation many city dwellers face. Yet, the shadow of his past still lingers, raising a question that hangs quietly over the entire conversation: can a visionary known for chaos truly change, or is this just the next act in a very familiar playbook? Read More

GOOD TO KNOW

On This Day: On this day in 1950, North Korea unleashed an attack southward across the 38th parallel, after which the UN Security Council (minus the Soviet delegate) passed a resolution calling on UN members to assist South Korea.

Fun Riddle: Charles Darwin formulated his theory of evolution while observing which animal?

QUOTE

LEISURE READS

Reddit Turns 20: Where Are Its Founders and Early Employees Today? - Read

These Are the Spotify Life Hacks Everyone Should Know About - Read

Burial Vault Sealed For 400 Years Found at End of Long-Forgotten Staircase - Read

More people buying electric cars and heat pumps than ever before - Read

ANSWER TO RIDDLE

Question: Charles Darwin formulated his theory of evolution while observing which animal?

Answer: Finch 🐦️