Welcome back to your daily memorandum talking tech, business, AI, markets, and more. 🗞️

In today’s edition, we are tackling the following:

🧠 Anthropic says Claude 4 outperforms humans at reasoning and hallucinations.
🎥 Microsoft fired protesters over Gaza AI contracts during Nadella’s Build keynote.
⚡ AI now uses 20% of data center power, and could reach 50% by 2025.
🤖 Tesla’s Optimus robot handles chores, Musk calls it ‘the biggest product ever.’
💰 Bitcoin breaks $111K as ETF demand surges and risk assets rally.

TECHNOLOGY

• Anthropic CEO claims AI models hallucinate less than humans

  • CEO Dario Amodei claims AI models hallucinate less than humans but acknowledges AI errors can appear more surprising or misleading.

  • He insists hallucinations are not a barrier to AGI, with models improving in accuracy, reasoning, and contextual understanding.

  • Amodei emphasizes that while AI presents false information confidently, this is manageable and not unique compared to humans.

• Introducing Claude 4

  • Anthropic launched Claude Opus 4 and Claude Sonnet 4, models designed for coding, reasoning, and handling complex tasks.

  • Claude Opus 4 is touted as the world’s best coding model, capable of sustaining performance on deep engineering challenges.

  • Both models offer enhanced memory and tool use and are available via Anthropic API, Amazon Bedrock, and Google Cloud Vertex AI.

• Disney is suing YouTube for poaching a key media and sports executive

  • Disney sued YouTube and former exec Justin Connolly for breach of contract after Connolly became YouTube’s head of media and sports.

  • Connolly previously led Disney’s streaming and media networks, holding key roles in ESPN and Disney’s distribution deals.

  • Disney claims his move jeopardizes sensitive negotiations, including Disney’s ongoing license renewal discussions with YouTube.

BUSINESS

• Microsoft fires employee who interrupted CEO's speech to protest AI tech for Israeli military

  • Microsoft fired engineer Joe Lopez after he interrupted CEO Satya Nadella’s keynote to protest AI contracts with the Israeli military.

  • Lopez, part of the ‘No Azure for Apartheid’ group, accused Microsoft of supplying AI tools used in military operations.

  • Microsoft confirmed its contract with Israel but stated its services are not used for operations harming civilians in Gaza.

• Klarna CEO and Sutter Hill take victory lap after Jony Ive's OpenAI deal

  • Klarna CEO Sebastian Siemiatkowski’s Flat Capital invested $3.6M in io, which converted into OpenAI shares after the $6.5B acquisition.

  • Sutter Hill Ventures also profited from its early stake in io but quickly deleted public posts following the deal announcement.

  • OpenAI paid $5B for the remaining IO shares, adding to its existing 23% stake from a prior investment.

• AI is eating data center power demand—and it’s only getting worse

  • AI now accounts for up to 20% of global data center energy use, potentially hitting 50% by the end of 2025.

  • Forecasts suggest AI could consume 82 terawatt-hours in 2025—equivalent to Switzerland’s annual electricity usage.

  • Companies like Google report emissions spikes, with AI-driven growth complicating efforts to meet net-zero climate targets.

MARKETS

WORLD

• US House passes 10-year moratorium on state AI laws

  • The U.S. House passed a bill blocking states from enacting AI regulations for 10 years to avoid fragmented state-level rules.

  • The bill, part of the ‘One Big, Beautiful’ package, now moves to the Senate, where it faces bipartisan resistance.

  • Supporters say it allows time for federal standards, while critics argue it prioritizes Big Tech over consumer protections.

• UK agrees to hand the Chagos Islands to Mauritius in a deal it says protects a key US base

  • The UK agreed to transfer sovereignty of the Chagos Islands to Mauritius while retaining a 99-year lease on Diego Garcia.

  • The deal includes £101 million in annual payments to Mauritius but has sparked legal challenges from displaced Chagossians.

  • Critics argue it ignores Chagossians’ rights, while the UK stresses the importance of securing the U.S. military base.

• Trump-Putin call yields no breakthrough on Ukraine cease-fire

  • A call between Trump and Putin failed to advance a Ukraine cease-fire, with both sides remaining entrenched in their positions.

  • Trump told European leaders that Putin believes Russia is winning and has no incentive to negotiate peace at this stage.

  • The call highlights the ongoing diplomatic deadlock as Russia continues its military operations in Ukraine.

CONTENT

• Open-Source AGI = Human Extinction? Debate with $85M Backed AI Founder

  • Sentient’s Himanshu Tyagi debates whether open AGI boosts progress or dangerously increases humanity’s existential risk.

  • Discusses Sentient’s $85M raise, AI’s control over systems, and extinction risks tied to open-source frontier models.

  • Covers P(doom), AGI safety, superintelligent AI risks, and how society should govern open AI development.

• How does Claude 4 think? – Sholto Douglas & Trenton Bricken

  • Anthropic researchers explain how Claude 4 reasons, scales with RL, and tackles mechanistic interpretability challenges.

  • Topics include computing bottlenecks, autonomous agents, alignment difficulties, and how countries should prepare for AGI’s impact.

  • Highlights how Claude differs from past models and the future of AI architectures in solving complex tasks.

• Unconventional product lessons from Binance, N26, Google

  • Mayur Kamat shares lessons on fast experiments, high-leverage decisions, and how fintech sharpens product management skills.

  • Discusses AI’s impact on product development, hiring top talent, managing failure, and choosing or avoiding the C-suite.

  • Covers balancing career growth, creating strong experimentation cultures, and why strengths matter more than fixing weaknesses.

FUTURISM

• A microrobot system designed to float inside a stroke patient

  • Artedrone's SASHA microrobot navigates blood vessels using magnetic guidance and imaging to autonomously perform clot removal.

  • The system aims to democratize stroke treatment by enabling non-specialists to conduct mechanical thrombectomies outside centers.

  • Preclinical studies show promising results, with plans underway for first-in-human trials to expand access to stroke care.

• Robot navigates with the 5 senses

  • Duke University researchers developed WildFusion, an AI system combining vision, touch, and vibration sensing for robotic navigation.

  • This multisensory approach allows robots to operate effectively in complex, unstructured environments like forests and disaster zones.

  • WildFusion enhances robotic perception, enabling more human-like interaction with diverse terrains and obstacles.

• Altman says the world isn't ready for the 'humanoid robots moment'

  • OpenAI CEO Sam Altman warns that society is unprepared for the imminent integration of humanoid robots into daily life.

  • He predicts scenarios where multiple robots perform tasks in public spaces, creating a 'sci-fi' atmosphere.

  • Altman emphasizes the need for societal readiness as these technologies rapidly advance and become commonplace.

EXTRAS

• Open social web browser Surf makes it easier for anyone to build custom feeds

  • Flipboard's new app, Surf, enables users to create personalized feeds from sources like Mastodon, Bluesky, Threads, blogs, and YouTube.

  • The introduction of ‘Starter Sets’ offers curated categories with recommended sources, simplifying the feed-building process for users.

  • Surf aims to empower users by providing control over their social media experience, moving away from algorithm-driven timelines.

• Tech CEOs are using AI to replace themselves

  • Klarna's CEO, Sebastian Siemiatkowski, utilized an AI-generated avatar to present the company's Q1 2025 earnings, highlighting AI integration.

  • Zoom's CEO, Eric Yuan, employed a digital twin via Zoom's AI Companion for the Q1 2026 earnings call.

  • These instances reflect a growing trend of tech leaders leveraging AI avatars in executive communications and presentations.

• Feds Charge 16 Russians Allegedly Tied to Botnets Used in Ransomware

  • The U.S. Department of Justice indicted 16 Russian nationals linked to DanaBot, a malware operation infecting over 300,000 systems globally.

  • DanaBot evolved from a banking trojan to a tool for credit card theft, ransomware, and espionage against military and government targets.

  • The indictment marks a significant move in disrupting cybercriminal networks that blur lines between criminal and state-sponsored activities.

AND MORE

• Google faces DOJ probe over Character.AI deal, raising antitrust concerns.

• US business activity improves in May, but inflation is poised to accelerate sharply.

• AI-Media and Lightning International join forces to make FAST channels accessible to everyone.

• Elf Labs is set up for a huge 2025, with new franchises, tech, and intellectual property in the pipeline.*

• Alphabet hits near three-month high on new AI updates.

• Walmart to cut about 1,500 corporate jobs to boost operational efficiency.

• CISA releases AI data security best practices to protect sensitive AI models.

• Up to 40% off a MasterClass subscription and unlock access to 200+ classes taught by the world’s best.*