Germany demands Apple and Google remove Chinese AI app DeepSeek from their stores, claiming it illegally ships user data to China. The move reflects growing global tensions over AI privacy and data sovereignty.
Meta poaches three OpenAI researchers with $100 million signing bonuses as Zuckerberg builds a "superintelligence" team. Sam Altman dismisses the blitz, but departures suggest money talks in AI's talent war.
Claude Code turns terminal-shy developers into design pros. This AI tool reads screenshots, builds JSON design systems, and creates professional websites that look like $10k custom jobs. But at $200/month, it's splitting the dev world in two.
Good Morning from San Francisco,
Meta dodged a copyright bullet Wednesday. A federal judge ruled the company legally used 13 authors' books to train its AI models. But Judge Chhabria delivered a warning disguised as a victory.
The authors—including Sarah Silverman and Ta-Nehisi Coates—sued Meta for pirating their books from shadow libraries. Their mistake? Wrong legal arguments.
Meta's AI could only spit out 50 words max from any book. The judge called that fair use. But he spotted a better attack: market dilution. AI could "flood the market with endless amounts" of content using "minimal creativity."
Translation: Next time, sue smarter. The judge practically wrote the playbook.
Stay curious,
Marcus Schuler
Meta won a legal fight Wednesday when a federal judge ruled the company's use of 13 authors' books to train its AI models was legal under copyright law's "fair use" rule. But Judge Vince Chhabria made clear this wasn't permission for tech companies to take whatever they want.
The case involved authors including Sarah Silverman and Ta-Nehisi Coates, who sued Meta in 2023 for using their books without permission to train its Llama AI models. Meta had downloaded the books from "shadow libraries"—pirate sites—after licensing talks failed.
"This ruling does not mean Meta's use of copyrighted materials to train its language models is lawful," Chhabria wrote. "It only means these plaintiffs made the wrong arguments."
The authors claimed Meta's AI could reproduce chunks of their books. But tests showed Llama could only produce 50 words max from any book, even when pushed to copy content. The authors also said Meta hurt the licensing market, but courts have long rejected claims about lost licensing fees for the exact use being challenged.
But the judge spotted a stronger argument the authors mostly ignored: market flooding. Chhabria warned that AI can "flood the market with endless amounts of images, songs, articles, books, and more" using almost no time and creativity.
Despite ruling for Meta, Chhabria predicted "plaintiffs will often win" in similar cases "with better records on market effects." The ruling only affects these 13 authors since it wasn't a class action. This follows a similar win for Anthropic this week.
Why this matters:
• Authors got a roadmap for future wins—the judge basically told them how to build stronger market flooding cases
• Tech companies shouldn't celebrate yet—this narrow ruling suggests other cases with better evidence could easily go the other way
Read on, my dear:
Prompt:
A surreal living room with an aquarium ceiling, numerous colorful fish swimming above and a colossal whale gracefully swimming overhead, warm sunlight filtering through the water creating dappled light and reflections on the walls, floor, a large television, a comfortable sofa, lush green plants, framed artwork on the walls, and various electronic appliances, an elegant wooden door with rippling light patterns, immersive underwater ambiance, clear and luminous lighting, tranquil atmosphere, dreamlike, highly detailed, photorealistic, wide-angle shot.
Anthropic just solved AI development's biggest headache. Claude users can now build and share AI-powered apps where usage counts against each user's own subscription, not the creator's bill.
Previously, viral AI apps meant financial disaster for creators. Popular tools could rack up thousands in API costs overnight. That killed most experimental projects before they started.
The new model changes everything.
When someone uses your Claude-powered app, their API usage draws from their own subscription. You pay nothing for their usage. No one manages API keys. The economics finally work.
Claude writes the complete code based on your description. You describe what you want, and Claude handles prompt engineering, error handling, and AI orchestration. Users authenticate with existing Claude accounts, and the system handles everything else.
Early users are pushing boundaries. AI-powered games where NPCs remember conversations and adapt to player choices. Learning tools that adjust to individual skill levels. Data analysis apps that let users upload CSVs and ask follow-up questions in natural language.
The variety suggests people see possibilities beyond typical chatbot interfaces. Writing assistants handle everything from scripts to technical documentation. Agent workflows orchestrate multiple Claude calls for complex tasks.
The system has boundaries. No external API calls yet—everything runs through Claude's completion API. No persistent storage means apps reset between sessions. The text-based API limits some interaction types.
These constraints might actually help. They force creators to focus on core functionality rather than getting lost in infrastructure complexity.
Why this matters:
• AI app development just became accessible to anyone who can describe their idea clearly—no coding or deployment expertise required
• The cost barrier that killed most experimental AI apps disappeared, potentially unleashing a wave of creative applications we haven't seen before
Read on, my dear:
Trump Mobile scrubbed its "made in America" marketing from its website after launching the T1 phone earlier this month, replacing bold manufacturing claims with softer language like "brought to life right here in the USA." The company still insists the phones are "proudly being made in America" despite removing the original banner that promised devices would be "MADE IN AMERICA."
Getty Images dropped its primary copyright claims against Stability AI in London court Wednesday, though the image company still pursues other allegations and a separate U.S. lawsuit. The move highlights the murky legal landscape around AI training on copyrighted content, coming a day after a U.S. judge sided with Anthropic in a similar book copyright dispute.
British hacker Kai West, known as "IntelBroker," was charged by US authorities for leading a group that stole data from dozens of companies and sold it online for over $2.4 million. West, 25, was arrested in France and faces extradition to the US, where he could receive 20 years in prison for breaches that allegedly included AMD, Cisco, and Hewlett Packard Enterprise.
Creative Commons launched CC signals, a framework that lets dataset owners specify how AI systems can use their content for training. The nonprofit aims to balance internet openness with creators' rights as companies increasingly wall off data or change policies to restrict AI access.
Meta hired three OpenAI researchers who set up the company's Zurich office, marking CEO Mark Zuckerberg's latest move to fix his company's AI struggles after its recent model launch disappointed. The trio—Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai—previously worked together at Google DeepMind before joining OpenAI, where they established the Swiss operation last year.
Scale AI's work training Google's Bard chatbot was flooded with "spam" from unqualified contractors who submitted gibberish and used ChatGPT to complete tasks requiring advanced degrees, according to internal documents. The problems persisted for 11 months between 2023 and 2024, with many contractors getting paid despite lacking credentials and producing shoddy work that evaded detection systems.
People using Anthropic's Claude chatbot for emotional support tend to become more positive as conversations continue, according to new company research. The study found that 2.9% of Claude interactions involve "affective use" - personal exchanges driven by emotional needs - though most users still rely on the bot for work tasks rather than companionship.
Gauth: AI Homework Helper Conquers Global Classrooms
ByteDance's stealth education play transformed a Singapore math app into a 200-million-user study companion that's quietly eating Chegg's lunch. The company rebranded from Gauthmath to Gauth in 2023, expanding beyond math to tackle every subject from chemistry to essay writing.
The Founders
• Founded 2020 by educator Curry Z in Singapore 📍
• Now operates as ByteDance subsidiary GauthTech Pte Ltd
• 200+ million users globally, employee count undisclosed
• Born from frustration with traditional tutoring limitations
The Product
• AI-powered homework solver with 24/7 human tutor backup
• Snap photo → get step-by-step solutions across all subjects
• Hybrid model: AI handles routine problems, experts tackle complex queries
• Free core service disrupts $20/month competitors like Chegg
• Chrome extension, mobile apps, web platform
The Competition
• Battles Google-owned Photomath (220M+ downloads) in math
• Duels Chinese rival Question.AI for US market dominance
• Crushes legacy players: Chegg stock crashed 50% after ChatGPT emergence
• Outflanks community-driven Brainly with instant AI responses
• Established market leadership in education app downloads
Financing
• Backed by $220B-valued ByteDance (TikTok parent)
• No traditional VC rounds - internally funded tech giant project
• Leverages ByteDance's AI expertise and global distribution muscle
• Strategic pivot after China banned for-profit K-12 tutoring 2021
The Future ⭐⭐⭐⭐
Gauth rides the AI education wave with deep pockets and proven execution. ByteDance connection creates regulatory risk if US-China tensions escalate. Competition intensifies as every tech giant builds homework helpers, but Gauth's hybrid AI-human model and free pricing maintain competitive moats. 🚀
Students will always seek easier homework solutions - Gauth positioned to capture that eternal demand.
Good Morning from San Francisco, 🌉 A federal judge just handed AI companies their first major copyright win. The catch? It
Good Morning from San Francisco, Former OpenAI CTO Mira Murati raised $2 billion for her six-month-old startup. That broke every
Good Morning from San Francisco, 🌉 OpenAI scrubbed its $6.5 billion Jony Ive acquisition videos after a judge sided with
Good Morning from San Francisco, Musk's xAI burns $1 billion monthly 🔥 while dodging pollution lawsuits in Memphis. The
Get tomorrow's intel today. Join the clever kids' club. Free forever.
