🗞️ THE WEEKEND RECAP
Top Stories You Might Have Missed
🤔 Grok 4 Leans on Musk to Tackle Controversy: xAI’s latest AI model appears to reference Elon Musk’s X posts and opinions when answering sensitive topics like immigration and abortion, blurring the line between “truth-seeking” AI and founder-aligned chatbot.
⏲️ OpenAI Delays Open Model: The much-anticipated open-source model has been pushed back with no new timeline, citing safety reviews and alignment challenges as the holdup.
💼 AI Recruiters Are Taking Job Interviews Now: From McDonald’s to Club Pilates, companies are using bots like “Alex” and “Olivia” to screen applicants—glitches, ghosting, and privacy worries included. Critics say the tech lacks empathy and clear disclosure rules.
🤖 Moonshot AI Open-Sources Kimi K2 to Reclaim Lead: China’s Moonshot unveiled Kimi K2, a powerful open-source model focused on coding and task automation, hoping to outshine rivals like DeepSeek and Anthropic and regain footing in the fast-shifting AI race.
🩺 Apple’s New AI Model Detects Health Shifts: Trained on 2.5B hours of Apple Watch behavioral data, the WBM model flags conditions like pregnancy or infection with up to 92% accuracy—often beating raw sensor data by focusing on long-term behavior patterns.
👤 Goldman Sachs Tests AI Coder Devin as ‘New Employee’: The bank plans to deploy hundreds of Cognition’s viral agent to assist its 12,000 devs—part of a hybrid AI-human model aimed at boosting productivity without replacing staff. Thousands more may follow.
💸 SpaceX to Pour $2B into xAI: Musk’s rocket company is backing his AI startup xAI with a $2 billion investment—tightening ties between his space and AI ventures as xAI ramps up.
💰 Robinhood CEO’s AI Math Startup Hits $875M Valuation: Harmonic AI raised $100M to build “mathematical superintelligence” with Aristotle, a verified reasoning model aiming to outsmart humans at math and avoid hallucinations. Public release expected soon.
🦾 Intel Spins Off AI Robotics Arm With $50M Funding: RealSense, now an independent company, aims to meet booming demand for humanoid robots. Backed by MediaTek and Intel Capital, it’ll focus on 3D perception tech as automation investments accelerate.
Enjoying our newsletter? Forward it to a colleague—
it’s one of the best ways to support us.
☝️ POWERED BY CAI CREATIVE
Text to Fully Editable MIDI Music
Turn text into fully editable MIDI with precise control over notes, rhythms, and feel, using natural language to describe exactly what you want. Whether it’s a syncopated bassline in 5/4 or a moody prechorus with rising strings, Cai understands musical intent and delivers accurate, customizable MIDI that fits seamlessly into your creative workflow.
Sign up and join our early community of beta testers!
👾 FORWARD FUTURE INTERVIEW
Dylan Patel on the Superintelligence Race, Why Apple Hates NVIDIA, and Who's 'Cooked' in AI
He's the guy the chip industry reads before making a move. Dylan Patel, founder of SemiAnalysis, joins Matthew Berman to break down the stark realities of the AI landscape. From the billion-dollar talent war to the corporate grudges shaping the future of technology, Dylan delivers unfiltered analysis on who’s winning, who’s losing, and why. → Read the full article here.
🏁 AI RACE
Apple Stalls on AI as Rivals Race Ahead, Wall Street Grows Impatient
Apple’s WWDC 2025 failed to convince investors or developers that it’s catching up in the AI race. Instead of bold moves, Apple delivered a cautious batch of updates to “Apple Intelligence,” including features like live translation and image generation powered by OpenAI—many of which are already available on rival platforms. The company’s most notable step was opening its smaller and server-based AI models to developers, but it didn’t provide updates on its promised Siri overhaul or rumored partnerships with Google or Perplexity.
Compared to recent high-impact launches from Google, Microsoft, and Anthropic, Apple’s announcements felt underwhelming and incomplete. Analysts criticized the lack of ambition, suggesting Apple’s slow pace in generative AI could threaten its market position. Apple’s stock has slumped 15% this year amid mounting concerns over its AI strategy, and critics argue its “better late than buggy” approach may no longer be viable in a fast-moving AI world. → Read the full article here.
⚠️ MISINFORMATION
Chatbots and Chaos: AI Stumbles Amid Texas Flood Misinformation Surge
When deadly floods hit Texas, AI chatbot Grok spread confusion by falsely blaming Trump’s NOAA budget cuts—then denying it made the claim. The episode spotlighted how chatbots, trained on flawed or biased data, can confidently deliver inaccurate or contradictory information during crises. Experts say tools like Grok, built on social media content and vulnerable to manipulation, are especially risky as more Americans turn to AI for news.
Despite some success debunking conspiracy theories, chatbots also reinforce user bias and can amplify falsehoods at scale. With 7% of Americans using AI for news weekly, experts urge users to verify sources and remember: these bots are not truth machines, just tech that still needs a fact-check. → Read the full article here.
🌍️ GLOBAL NEWS
China Plans 39 AI Data Centers with 115,000 NVIDIA GPUs
China is ramping up its AI infrastructure with plans for 39 new data centers equipped with 115,000 NVIDIA Hopper-based chips—H800 and H20 models that skirt U.S. export restrictions. These “compliant” GPUs are less powerful than their U.S.-banned counterparts but are still capable enough for large-scale AI workloads.
The move highlights China’s aggressive strategy to maintain AI momentum despite tightening U.S. trade controls. Experts say this raises serious questions about the long-term effectiveness of U.S. export bans, as China appears increasingly adept at sourcing restricted tech through loopholes and tailored chip variants. The data centers, part of China’s “Eastern Data, Western Computing” initiative, reflect a broader push to secure digital sovereignty and accelerate AI development at national scale. → Read the full paper here.
📽️ VIDEO
Grok 4 Is Really Smart... Like REALLY SMART
Grok 4 pushes the AI frontier with multi-agent reasoning, reinforcement learning, and record-breaking benchmarks—scoring 50.7% on humanity’s hardest exam. Get the full scoop in Matt’s latest video! 👇
🧰 TOOLBOX
Smarter Hiring, Sharper Studying, and Face-First Intelligence
👨💼 Phenom: AI-driven platform to streamline hiring, boost retention, and personalize candidate and employee experiences.
📚 QuickTakes: AI study buddy for notes, guides & 24/7 help—record lectures, upload files & prep smarter with organized class folders.
🪪 FaceMex: Real-time AI facial recognition & emotion analysis for secure ID checks, behavior insights & scalable, privacy-first solutions.
