Palmer Luckey got fired from Facebook for backing Trump. Now Meta needs his defense company to win a $22 billion military contract. The reunion changes everything.
DeepSeek released an AI update without fanfare that now ranks just behind OpenAI's best models. The Chinese startup's quiet approach achieved what billion-dollar marketing campaigns couldn't: performance gains that forced competitors to cut prices.
Altman's latest venture asks Americans to stare into an orb for $42 in crypto 👁️💰. The OpenAI CEO calls it human verification. Critics call it surveillance with a payout.
The white spheres scan your iris, encrypt the data, then keep it forever 🔮. Seven thousand five hundred devices will hit gas stations by year's end. Meanwhile, several countries banned the project outright 🚫.
The irony cuts deep. Altman warns about AI agents flooding the internet, then builds tools to help bots use human credentials 🤖. He profits from the problem and sells the solution.
Anthropic caught up fast 🏃♂️.
Claude now talks back with five voices after trailing ChatGPT for two years 🗣️. The company made web search free and added voice mode in the same week. Every major AI chatbot now offers spoken conversations. The race shifted from who can chat to who chats best 🎯.
Stay curious,
Marcus Schuler
The Orb Arrives: Scan Your Eye, Get Paid
Sam Altman has a new pitch. Stare into his beach ball-sized device and get $42 in cryptocurrency. The OpenAI CEO calls it proof you're human.
The Orb is rolling out across American cities. Austin, Miami, Los Angeles, and San Francisco now host the white spheres. By year's end, 7,500 devices will sit in gas stations and corner stores nationwide.
Here's how it works. You download an app, flash a QR code, then gaze into the Orb's camera. It scans your iris patterns and creates a unique digital code. Seconds later, Worldcoin tokens appear in your wallet.
Altman argues this system will help humans stay relevant as AI agents flood the internet. His company Tools for Humanity wants to verify 180 million Americans. The goal sounds ambitious for a project that has signed up just 12 million people globally since 2023.
The privacy angle gets murky fast. The company claims it deletes your iris image after scanning. But it keeps encrypted derivatives of your biometric data forever. European regulators called this approach fundamentally flawed and ordered full data deletion rights.
The timing feels odd. Altman spent years warning about AI creating fake content online. Now he's building systems to help AI agents operate with human credentials. Users can delegate their verified identity to bots, letting AI act on their behalf.
Several countries have banned or investigated the Orb. Spain, Argentina, Kenya, and Hong Kong cite concerns about excessive data collection. Critics note the irony of Altman solving problems his own AI company helped create.
The business model banks on cryptocurrency speculation. Tools for Humanity raised $244 million and holds tokens worth $1.2 billion. Most coins are reserved for human verification rewards. Early investors and staff, including Altman, control the remaining 25%.
South Korea serves as the testing ground. The company verified 55,000 people there and plans to reach 1 million next year. But adoption remains slow. In Seoul coffee shops, most visitors come for quick crypto, not digital identity.
The project faces a chicken-and-egg problem. Few platforms accept World ID verification, giving users little reason to scan beyond free money. Reddit, Telegram, and Shopify support it, but mainstream adoption lags.
Altman's car ride interview reveals shifting priorities. He downplays the anti-bot mission and emphasizes AI empowerment instead. The contradiction runs deep through a project that promises to preserve humanity while enabling its digital replacement.
Why this matters:
The solution creates the problem: Altman profits from AI agents while selling protection from them, a conflict that should make users pause.
Your iris is forever: Despite privacy claims, biometric data never truly disappears from this system, making it a permanent digital tattoo you can't remove.
Prompt: polaroid style photo of 30 year old Blonde woman in an alley at night from the 90s, she is smoking a cigarette, wearing grunge fashion, flash photography, color film, analog style, imperfect, slight double exposure
Anthropic’s Claude gets five voices and hands-free conversations
Anthropic rolled out voice mode for Claude's mobile apps this week. Users can now have spoken conversations with the AI assistant instead of just typing questions.
The feature puts Claude on equal footing with ChatGPT, Google's Gemini Live, and xAI's Grok. All major AI chatbots now offer voice interactions.
Claude's voice mode runs on the new Sonnet 4 model by default. Users pick from five voice options and can switch between text and voice mid-conversation. The app shows key points on screen while Claude speaks and saves transcripts afterward.
Free users get 20 to 30 voice conversations before hitting daily limits. Paid subscribers can connect Google Workspace to discuss emails and calendar events through voice. Enterprise customers get Google Docs access too.
The rollout happens gradually over several weeks. Voice mode works in English on iOS and Android devices. Users tap a sound wave icon next to the text input to start talking.
Anthropic also made web search free for all users this month. The feature was previously limited to paid subscribers. Combined with voice mode, Claude now matches the core features offered by its main competitors.
The company launched two new models last week - Opus 4 for complex coding tasks and Sonnet 4 for mixed reasoning work. Voice mode showcases Sonnet 4's ability to handle quick responses and longer conversations.
Safety measures prevent voice cloning. Claude uses preset voices rather than mimicking specific people. The same content policies apply whether users type or speak.
Why this matters:
Voice interactions make AI assistants more practical for busy people who need hands-free help
Anthropic completes its transition from text-only startup to full-featured competitor against OpenAI and Google
Telegram Teams Up With Musk's xAI for $300 Million AI Deal
Telegram struck a deal with Elon Musk's xAI to bring Grok AI to its billion users. The messaging app gets $300 million in cash and equity plus half of all subscription revenue from xAI services sold through Telegram. TON token jumped 18.5% as news leaked two hours before the official announcement, climbing from $3.28 to $3.55.
China's AI Upstart Tweaks the Model That Shocked Silicon Valley
DeepSeek rolled out a "minor trial upgrade" to its R1 AI model that rocked the tech world in January. The Chinese startup told users they can test the updated version through its official WeChat channel. R1 originally outperformed Western AI models at a fraction of the cost, triggering a global tech stock selloff and sparking China's AI arms race.
Meta's Future Hangs on Judge's Antitrust Decision
Meta's landmark antitrust trial wrapped up Tuesday after six weeks of testimony, including CEO Mark Zuckerberg defending his company's $1 billion Instagram and $19 billion WhatsApp acquisitions. The FTC argues Meta used a "buy or bury" strategy to kill competitors, while Meta claims it faces real competition from TikTok and YouTube. Judge James Boasberg will decide whether to force Meta to sell both apps, potentially reshaping Silicon Valley's power structure.
EU Gives Apple 30 Days to Fix App Store or Face Daily Fines
The European Commission rejected Apple's App Store changes and gave the company 30 days to comply with EU competition law. Apple's new rules still block developers from steering users to cheaper payments outside the App Store, despite earlier changes that reduced fees from 30% to 27%. The EU threatens daily fines until Apple fixes the violations.
Saudi Arabia Launches $10 Billion AI Fund to Challenge Silicon Valley
Saudi Arabia launched Humain, a state-owned AI company backed by its $940 billion sovereign wealth fund, aiming to build some of the world's largest data centers. The company will spend $10 billion on a venture capital fund and has already signed $23 billion in deals with Nvidia, AMD, and other US tech firms. Humain plans to handle 7% of global AI processing by 2030 through massive data center parks powered by hundreds of thousands of chips.
Getty Images Battles AI Firms Over Copyright Theft in Expensive Legal Fight
Getty Images CEO Craig Peters says his company is spending millions to fight AI firms that steal copyrighted images to train their models. The photo licensing giant is suing Stability AI over claims the startup copied 12 million images without permission to build its Stable Diffusion model. Peters calls this "unfair competition" and "theft," while AI companies argue that paying for content would kill innovation.
🚀 AI Profiles: The Companies Defining Tomorrow
Ocean
Ocean launches as a decentralized AI search engine that promises 10x faster responses than ChatGPT while staying free from Big Tech control. The startup bets on speed, privacy, and community governance to carve out space in the crowded AI assistant market.
The Founders Ahmad Shadid founded O.XYZ in 2024 after stepping down from his $4.5B Solana project Io.net. The Qatar-based quant engineer dropped $130M of his own cash to build "sovereign AI" outside corporate control. Based in Geneva with a software hub in Doha, the team pulls talent from Stanford, OpenAI, and NASA. 🚀
The Product Ocean runs on Cerebras wafer-scale processors with 900,000 AI cores, delivering responses up to 20x faster than competitors. Users can chat by voice or text without creating accounts - just say "Hi" and you're in. The platform routes queries through specialized AI models via its O Routing Intelligence system, tapping the best expert for each question. Privacy-first design means no data harvesting.
The Competition Ocean swims with sharks. ChatGPT dominates with 100M+ users and Microsoft backing. Google pushes Gemini integration across search. Perplexity raised $500M at a $14B valuation for cited AI answers. Ocean's edge: unmatched speed and decentralized architecture that no government can shut down. But trust remains unproven - former insiders flagged inflated tech claims in 2024.
Financing Shadid's $130M personal investment dwarfs the $4.5M seed round from crypto VCs. No official valuation exists, though the self-funding suggests high hundreds of millions. The project issued $OI governance tokens for community ownership instead of traditional equity.
The Future ⭐⭐⭐⭐ Ocean could redefine AI search if it maintains its speed advantage and builds user trust. The decentralized model appeals to crypto natives and privacy advocates globally. However, competing against trillion-dollar incumbents with unproven governance at scale poses real risks.
