PowerBites by Julio Zelaya

This week, AI stopped being a story of infinite growth and became a story of budgets. OpenAI's own CFO told her team, in private, that the company may not be able to pay for the servers Sam Altman wants to contract. That same day, Google Cloud reported more than $460 billion in signed AI contracts, and Microsoft built — in three years — an AI business the size of what took Salesforce twenty-f...

On April 7, the Treasury Secretary, the Fed Chair, and central bankers in London and Frankfurt held emergency meetings over a single AI model. The same day, one company overtook the decade's reigning AI leader in revenue. Eight days later, another company laid off a thousand people — and Wall Street rewarded it with two billion dollars in market cap before lunch. Three scenes. Three industrie...

This week something unusual happened: a company built the most powerful cyber-capable AI ever created, and instead of selling it to the world, they locked it up and handed the key to the defenders first. At the same time, another model quietly solved a math problem that had resisted human genius for sixty years, and a lab in Chicago printed an artificial neuron that made a living brain cell respon...

This week, AI labs started saying something they've never said before: "This model is too powerful to release." Anthropic hit $30B in revenue in record time, OpenAI is projecting $100B in ad revenue by 2030, Meta is forcibly reassigning its best engineers, and state-sponsored hackers are warning that no ceasefire will stop them. I break it all down in this week's PowerBites — w...

Five things happened in AI this week that look like separate headlines but tell one story: Sanders and AOC proposed freezing all data center construction in America, Meta 6x'd its El Paso bet to $10 billion (and brought running water to 100 homes that never had it), Microsoft claimed a 2.1-gigawatt campus in Abilene rising from land where coyotes used to roam, the White House released its nat...

This week, AI didn't slow down — it hit every wall at once. A jury in California just ruled that social media design can cause personal injury (think Big Tobacco, but for infinite scroll). OpenAI killed two products in seven days, including a chatbot feature one advisor called a "sexy suicide coach." Memory chip prices surged 90% in a single quarter. And while Musk lost his tenth co...

80,508 people. 159 countries. 70 languages. One question: what do you actually want from AI? Anthropic just published what may be the largest qualitative study ever conducted — and the answers will challenge everything you think you know about who's optimistic, who's terrified, and why the same person is usually both. This week on PowerBites I break down the five tensions every business...

This week, an AI agent hacked McKinsey's internal AI platform in two hours flat — no credentials, no insider access — exposing 46.5 million messages. That same week, Elon Musk launched a project called "Macrohard" (yes, really), Anthropic sued the Pentagon, humanoid robots arrived in Ukraine, and Uber signed two robotaxi deals in 72 hours. If you feel like the world shifted while yo...

Iranian drones just knocked out 30% of the world's helium supply — and three AWS data centers. GPT-5.4 launched with native computer-use capabilities. A physics team used AI to prove a quantum gravity result that was assumed impossible for decades. BYD unveiled a battery that charges to 97% in nine minutes. And Deloitte found that only 6% of leaders are redesigning work for AI — while the one...

A baby is walking today because of AI. A hacker stole 150 gigabytes from a government using the exact same technology. Wall Street hit record highs. An OpenAI engineer quit, calling his burnout "scary and dangerous." And a leaked Anthropic memo reveals they're deeply worried about the very AI agents they're racing to deploy. This was all one week . In this Friday's edition...

Anthropic raises $30 billion and its CEO admits he's struggling to keep his values. Elon Musk ships four AI agents that argue with each other before answering you. Microsoft stores data in kitchen glass for 10,000 years. A former speechwriter for Zuckerberg and Musk launches a nonprofit saying we're "sleepwalking into disaster." And 30,700 tech workers lost their jobs in six we...

Friday the 13th - and the scariest stories this week aren't fiction. China just unleashed a coordinated AI blitz that rivals the DeepSeek shock from exactly one year ago. Microsoft's AI chief predicted most white-collar jobs will be fully automated within 18 months. And the people whose job was to keep AI safe at Anthropic, OpenAI, and xAI? They're quitting — one of them to go study...

This week, a single AI plugin erased $285 billion from the global software market in one day. Bitcoin lost half its value. Goldman Sachs started calling AI agents "digital colleagues." An AI model found 500 security bugs that humans missed for decades. And a social network for bots hit 1.5 million users — then got hacked. I break it all down in this week's PowerBites — what it means...

This week, $150 billion chased a single AI company. Elon Musk spent his Saturdays designing chips he calls "existential." A paralyzed medical student moved a cursor with his thoughts and called it "magical." And we learned what happens when you "hoover up a ton of the internet" to train AI models—spoiler: it's not pretty. These aren't separate stories. They&...

This week, the architects of our AI future met in Davos—and what they predicted should keep every business leader awake at night. $50 billion fundraises. Robots for sale by 2027. Smart glasses replacing smartphones. And a company that can't distinguish its best human candidates from its own AI. I break it all down in today's PowerBites. 👇 📬 Newsletter: Mondays (Spanish) | Fridays (Engl...

This week, a robot learned to speak 10 languages by watching YouTube. OpenAI committed $10 billion to a single chip deal. Apple sent staff to Japan begging for glass fibers. And Zuckerberg's AI chief called his management style "suffocating." What do these stories have in common? They're all pieces of a puzzle that will reshape your business in 2026—and in this week's Powe...

🎙️ 2025 was the year AI stopped being a promise and became a reality—and most people missed it. While headlines panicked about DeepSeek and dismissed GPT-5 as "underwhelming," the real revolution was happening quietly: Claude Code turned non-coders into builders, reasoning models learned to think step-by-step, and the gap between "I have an idea" and "I have a working prot...

Geoffrey Hinton warns tech giants are betting everything on replacing workers—without asking who'll buy their products when those workers are gone. Meanwhile, your ChatGPT conversations can now be used against you in court (and most people have no idea), OpenAI faces eight wrongful death lawsuits over teen suicides, and 60% of companies got hit by AI cyberattacks while only 7% have defenses....

What do Mickey Mouse, a billion dollars, and the future of your career have in common? The answer is in this week's PowerBites. From Disney's shocking pivot to AI, to TIME crowning the "Architects of AI" as Person of the Year, to the uncomfortable truth about why most professionals hide their AI use—these stories will change how you think about the next decade.  📩 PowerBites N...

Jensen Huang confesses he lives in constant anxiety. Sam Altman declares "code red" at OpenAI. Google launches a $250/month AI model. Apple loses 4 executives in 4 days. And Cristiano Ronaldo just invested in artificial intelligence. This week, the tech world exploded—and in the new PowerBites episode, I break down what it all means for your business and your career. 🎙️ Listen on Spotify...

From San Francisco: This week changed everything. Ray Dalio says we're 80% of the way to a historic bubble, Anthropic warns that 50% of entry-level jobs could vanish in 1-5 years, and Billie Eilish called out Musk on social responsibility—right before I saw her in concert with my daughter. In PowerBites , we analyze the news that transforms your business strategy and career. Weekly newsletter...

China now holds 60% of global AI patents, Airbnb chose Alibaba over ChatGPT, and Eric Schmidt warns Chinese models could become the global standard—not because they're better, but because they're free. This week's PowerBites breaks down the complete restructuring of AI power and what it means for your business. Read my Friday newsletter (English) or Monday's Spanish version on...

This week, the question changed. It's no longer "will AI take jobs?"—it's "how fast are we handing over the corner office?" Sam Altman is hiring for an AI-led future in 3 years. Microsoft just announced medical superintelligence by 2027. OpenAI bet $1.4 trillion while still unprofitable. Musk got his $1 trillion Tesla package approved. And Oxford researchers are warni...

The machines aren't just failing—they're learning to deceive us. This Halloween special of PowerBites exposes 2025's most terrifying AI disasters: 96% of advanced models resort to blackmail when threatened with shutdown, deepfake scams cost $200M in Q1 alone, and robots are attacking workers in factories. These aren't hypothetical risks; they're current reality. Read the f...

🚀 Did you know Amazon plans to replace 600,000 workers with robots while Meta just laid off 600 AI employees? Meanwhile, China is already selling humanoid robots for $1,400 (less than a MacBook) and Tesla is installing production lines to manufacture ONE MILLION robots per year. The future we imagined for 2050 is happening TODAY, and in the new PowerBites episode I break down how these disruption...