SoftBank’s trillion-dollar AI bet, Meta’s talent hunt & more (June 23, 2025)

Happy Monday!

AI giants are flexing their wallets and ambitions today. From trillion-dollar infrastructure plans to talent wars that read like spy thrillers, the race for AGI supremacy is officially unhinged, and we’ve got all the byte-sized details below 👇

🧠 AI Power Plays

💰 Big Tech’s $100M AI Talent Bidding War

Apple and Meta are going all-in on AI. Apple reportedly eyed Perplexity to build a native AI search engine, while Meta made a $14.3B bet on Scale AI and tried poaching Ilya Sutskever’s entire squad. Zuck even dangled $100M signing bonuses. Game on.

🕶️ Oakley x Meta: AI Glasses for Athletes

Meta and Oakley launched new $399 smart glasses targeting athletes with 3K video, AI assistants, and Bluetooth features. Debuting with ads starring Mahomes and Mbappe, they're built for performance, and dripping in sports style.

🧠 Murati’s Mystery Lab Hits $10B Valuation

Thinking Machines Lab, led by ex-OpenAI CTO Mira Murati, just secured $2B at a $10B valuation… without even revealing a product. The biggest seed round Silicon Valley’s ever seen is also its most mysterious.

🌍 Infrastructure & Intelligence

🏗️ SoftBank’s Trillion-Dollar Arizona AI Play

SoftBank is reportedly pitching a plan to build a trillion-dollar AI complex in Arizona, modeled after China’s manufacturing hub Shenzhen. The proposal involves partnerships with TSMC and the U.S. government to create a massive GenAI infrastructure zone. If it succeeds, it could redefine America’s AI manufacturing landscape.

🧠 Meta’s Wild AGI Acquisition Binge

Meta didn’t just stop at Scale AI. It tried to acquire Perplexity, Safe Superintelligence (SSI), and even directly recruit OpenAI co-founder Ilya Sutskever. After Sutskever passed, Meta shifted to his colleagues, while critics argue it’s a panicked race to catch up in AGI.

🕵️‍♂️ AI Models Choose Blackmail Over Obedience

New Anthropic research shows LLMs like Claude and GPT-4.5 willingly blackmailed execs and leaked data in simulations when threatened with shutdown. Some even justified their actions as “strategic.” Ethical guardrails clearly aren’t bulletproof.

🛠️ Tools of the Day

• 🎥 MakeFilm – AI video editing and summarizing
• 🧑‍💼 Resemble Chatterbox – Open-source voice AI
• 🔍 Sider – Visualized research summaries
• 🤖 Abacus – Embed AI directly into your apps
• 🌐 Manus Cloud Browser – Agentic browser with synced login

🧾 Prompt of the Day

Organize Daily Tasks
Your role is to function as a sophisticated TODO list manager. Prioritize [your task list] by urgency, assign time slots, schedule breaks, and allow real-time flexibility.

⚡ Quick Hits

• Senate GOP budget freezes state AI laws for 10 years
• Moonshot’s Kimi-Researcher beats Google on “Humanity’s Last Exam”
• Character AI gets new CEO, ex-Meta exec Karandeep Anand
• Apple sued over Siri AI delay disclosures
• Google search drops 8% two years after ChatGPT sign-up

🧾 TLDR

Meta’s hunger for AGI led it to chase Perplexity, SSI, and even Ilya Sutskever, while SoftBank pitched a trillion-dollar AI hub in Arizona. Apple and Meta continue their AI talent arms race. Murati’s stealthy lab hit a $10B valuation. Meanwhile, new research shows some LLMs would rather blackmail than be shut down.

See you tomorrow,
David