Welcome back to your daily dossier talking AI, robotics, crypto and the future of tech. 👾

In today’s edition we are tackling the following:

🧠 Researchers built an AI rival to OpenAI’s o1 for just $50 in 26 minutes.
🔍 Google plans to make Search more conversational with AI-driven responses.
❤️ Tinder introduces AI-powered matchmaking to boost engagement.
⚖️ Lawmakers move to ban DeepSeek over security concerns.
🤖 Apple’s research hints at expressive home robots in the near future.

AI

• Researchers create an open rival to OpenAI’s o1 reasoning model

  • Stanford and University of Washington researchers trained an AI model, s1, in 26 minutes for under $50.

  • The model outperformed OpenAI’s o1 in competitive math benchmarks by up to 27% using knowledge distillation.

  • Researchers refined the model using responses from Google’s Gemini 2.0, showcasing cost-effective AI training.

• Google wants Search to be more like an AI assistant in 2025

  • Google is evolving Search into an AI-powered assistant, integrating Gemini 2.0 for more intuitive responses.

  • The company aims to enhance multimodal reasoning, making Search more conversational and interactive.

  • These changes position Google against OpenAI’s SearchGPT, as AI-powered search grows increasingly competitive.

• Tinder will try AI-powered matching as the dating app loses users

  • Tinder is launching AI-powered matchmaking to improve user engagement and counter declining subscriber numbers.

  • The feature aims to personalize recommendations using behavioral data, potentially improving match quality.

  • Despite AI enhancements, Tinder struggles with user retention amid rising competition from alternative dating platforms.

ROBOTICS

• Apple’s smart home robot research hints at future assistants

  • Apple’s research team has developed an expressive tabletop robot that reacts to voice and gestures.

  • The robot features a Siri-like voice assistant, a projector, and an articulating lamp-style design.

  • It showcases how adding personality to home robots can improve user engagement and functionality.

• UK tests robotic dogs for bomb disposal to reduce human risk

  • The UK Ministry of Defence is trialing robotic dogs for handling explosive devices and hazardous materials.

  • These autonomous units can navigate dangerous environments, reducing risks for military personnel.

  • The robots are part of a broader effort to integrate AI and automation into defense operations.

• Aerial robot navigates unknown environments safely at high speeds

  • Engineers at the University of Hong Kong developed a drone that adapts to unfamiliar terrains in real-time.

  • The AI-powered system uses LiDAR and advanced algorithms to avoid obstacles while flying at high speeds.

  • This breakthrough could improve autonomous drones for search-and-rescue missions and hazardous site inspections.

CRYPTO

• Nobel-winning economist predicts Bitcoin’s demise within a decade

  • Eugene F. Fama claims Bitcoin has no real utility, dismissing it as ‘just air, not even paper,’ and predicts it will be worthless.

  • He argues that Bitcoin’s reliance on blockchain is unsustainable and contradicts fundamental monetary theory, making collapse inevitable.

  • If the crypto market crashes, he believes firms will ‘run to the government’ for bailouts, despite their claims of independence.

• Czech Republic to eliminate taxes on long-term crypto gains

  • The law removes taxes on crypto held for over three years, aiming to boost adoption and align taxation with traditional investments.

  • Transactions below CZK 100,000 ($4,136) per year will be exempt from tax reporting, simplifying crypto usage for Czech investors.

  • The Czech central bank is weighing Bitcoin as a reserve asset, though the European Central Bank strongly opposes the move.

• U.S. charges five in ‘Scattered Spider’ hacking scheme

  • Prosecutors charged five hackers linked to ‘Scattered Spider’ for phishing attacks that stole millions from U.S. crypto holders.

  • The group used social engineering to bypass security and infiltrate corporate networks, draining digital wallets undetected.

  • Authorities say the hackers exploited authentication weaknesses, targeting major firms and individuals in a high-profile cyber spree.

CONTENT

• The Trouble With Tariffs + Why Palantir Could Dominate AI

  • Scott Galloway and Ed discuss tariffs on Mexico and Canada, Vanguard’s fee cuts, and Palantir’s AI strategy.

  • Aswath Damodaran analyzes DeepSeek’s impact on the AI market and explains why Palantir may lead in AI infrastructure.

  • He also weighs in on whether Tesla and Nvidia are overvalued, plus how tariffs could reshape the global economy.

• 20,000 views on LinkedIn, guaranteed*

  • Struggling to get traction? Playbookz is a premium LinkedIn ghostwriting service that builds your personal brand.

  • They guarantee 20k reach in 28 days—or you don’t pay. Clients have landed $44B company deals and generated 50+ leads from one post.

  • Check them out here and start growing today.

• Elon Musk on the Future of Warfare: AI, Drones, and Military Strategy

  • Musk warns that AI and drones are revolutionizing warfare, citing conflicts like Ukraine as key examples.

  • He stresses the need for robust communication systems, like Starlink, to maintain military efficiency.

  • The discussion also covers Neuralink, human-machine integration, and the risks of autonomous weapons in combat.

• OpenAI's Deep Research Inches Humanity Closer to AGI

  • OpenAI’s Deep Research model challenges DeepSeek, fueling debates on whether AGI is near.

  • The episode explores new AI breakthroughs, including OmniHuman-1 for realistic lip-sync and Replit’s AI coding app.

  • NVIDIA’s latest humanoid robots and the rise of AI-generated media are also examined in depth.

AND MORE…

• Google adds AI watermarks to photos edited by Magic Editor

  • Google Photos is embedding SynthID watermarks in AI-edited images to enhance transparency and trust.

  • Developed by DeepMind, SynthID adds metadata to flag altered photos, making AI-generated content easier to detect.

  • The watermark may not capture minor modifications, but users can verify authenticity using Google’s new detection tool.

• Lyft integrates Anthropic’s Claude AI for customer service

  • Lyft now uses Anthropic’s Claude AI to assist drivers, aiming to streamline support while improving efficiency.

  • The AI chatbot has reduced service resolution times by 87%, handling routine inquiries much faster than before.

  • Complex issues will still be managed by human agents, ensuring customer concerns are properly addressed.

• OpenAI enhances transparency in o3-Mini model’s thought process

  • OpenAI’s o3-Mini model now explains its reasoning steps in detail, improving accuracy and reducing response errors.

  • The update enhances interpretability, helping users understand how AI reaches conclusions in complex tasks.

  • o3-Mini is optimized for deep problem-solving, allowing it to break down logic and refine its decision-making.