Good morning. It's Wednesday, April 8, and we're covering Meta’s AI token race, older workers pivoting to AI gigs, Intel teaming up with Elon Musk, and more.
If you enjoy this email, share it with a friend. First time reading? Sign up here to keep up with the future of tech.
Intel Joins Musk’s Terafab AI Push
Intel will partner with Elon Musk’s Terafab project to build chips for AI, robotics, and data centers. The effort targets massive compute output to support Tesla, SpaceX, and xAI ambitions. Intel’s role could strengthen its turnaround and foundry business despite ongoing losses. Investors reacted positively, sending shares up over 2%.
Uber Expands AWS Bet on AI Chips
Uber is expanding its AWS deal to use Amazon’s Graviton chips and test Trainium3 for AI workloads. The move shifts some focus away from prior partnerships with Google and Oracle. AWS is positioning its custom chips as cost-efficient alternatives to NVIDIA.
Anthropic Launches AI To Hunt Software Flaws
Anthropic unveiled Project Glasswing, using its Claude Mythos Preview model to detect critical software vulnerabilities. The system has already identified thousands of high-severity flaws across major operating systems and browsers. Backed by partners including AWS, Apple, and Microsoft, the effort aims to strengthen global cybersecurity defenses.
Chrome Adds Vertical Tabs, Reading Mode
Google is rolling out vertical tabs and a new reading mode in Chrome to improve browsing and focus. Users can move tabs to a side panel for easier navigation and title visibility. The reading mode removes page clutter for distraction-free viewing. Both features are accessed via right-click menus.
POWERED BY ACCIO WORK
Accio Work: Your Agentic Team for Real Business
Meet Accio Work - the agentic workspace for business owners and solopreneurs. Our specialized agent team manages operations for you—sourcing products, negotiating with suppliers, managing stores, and launching marketing—all on autopilot.
With verified skills and business tool APIs, Accio Work takes action while you stay in control. Powered by Alibaba.com’s 1B+ products and global supplier data, it handles product selection and execution seamlessly. No setup required—just results.
REAL OR AI
Think You Can Still Spot AI? ↓ Real or fake?
VIDEO
Matt Builds Something… Journey!
PRODUCTIVITY
Meta Employees Compete for AI “Token Legend” Status Internally
Meta employees are competing on an internal leaderboard that ranks workers by how many AI tokens they consume, with top users earning titles like “Token Legend” and “Session Immortal.” The system, dubbed “Claudeonomics,” tracked more than 60 trillion tokens used across roughly 85,000 employees in a recent 30-day period. Leadership has encouraged heavy AI usage as a proxy for productivity, with some executives suggesting engineers should spend hundreds of thousands of dollars annually on tokens.
The trend reflects a broader Silicon Valley shift toward “tokenmaxxing,” where AI usage is gamified and treated as a performance signal. Critics and uncertainties remain around actual cost efficiency and whether higher token usage directly correlates with better outcomes. → Read the full article here. (Paywall)
WORKFORCE
Older Workers Turn to AI Training As Fallback Employment
A growing number of older, skilled U.S. workers are turning to AI data training jobs after struggling to re-enter the workforce following layoffs or health setbacks. The work—labeling and evaluating outputs for models like ChatGPT and Google Gemini—offers flexible, contract-based income, but often at significantly lower pay and without benefits.
Workers over 60 take about 50% longer to find new jobs, and many never regain prior earnings, making AI training a “bridge job” to stay afloat. While some experts can earn over $100 per hour, many roles pay $20–$40 hourly and lack stability. → Read the full article here.
ROBOTICS
Generalist Unveils GEN-1 Robot Model With 99% Task Reliability
Robotics startup Generalist has introduced GEN-1, a physical AI model that achieves 99% success rates on repetitive, fine-motor tasks like folding boxes, packing devices, and repairing robot vacuums. The system is trained on over 500,000 hours of human “data hands” interactions—capturing detailed motion and visual input—and can adapt to new tasks with about an hour of additional robot-specific training.
Unlike earlier systems, GEN-1 can recover from disruptions and improvise actions not explicitly programmed, such as adjusting grip or retrying failed steps. → Read the full article here.
NEWS
What Else is Happening
Dimon Predicts 3.5-Day Workweek: JPMorgan CEO Jamie Dimon says AI could shrink workweeks within 30 years while boosting jobs, but warns rapid automation may cut roles and demands retraining efforts
Anthropic Secures Massive TPU Expansion: Anthropic partners with Google and Broadcom for multi-gigawatt TPU capacity starting 2027, fueling Claude amid $30 billion revenue surge and rapidly growing enterprise demand.
Samsung Profit Surges on AI Chips: Samsung forecasts 57.2 trillion won Q1 operating profit, up eightfold year-over-year, as booming AI memory demand drives shortages, higher prices, and renewed competitiveness in memory.
Rocket Launches AI Strategy Reports: India’s Rocket unveils a platform generating consulting-style product strategies from prompts, offering pricing, insights, and competitor tracking at far lower cost than traditional firms.
REAL OR AI
The Image Is Real. Tune in next week for another Real or AI test.
