OpenAI just launched o3 and o4-mini, models that combine visual intelligence with deeper reasoning. For the first time, these AIs don't just see images – they think with them, manipulating photos to extract insights just as a human would zoom in or rotate a picture to understand it better.
The US-China chip war just hit Wall Street hard. Tech stocks tumbled after Nvidia lost $5.5 billion from US restrictions on AI chip sales to China. Meanwhile, Taiwan strengthened its grip as the world's essential chip supplier.
A new artificial intelligence system from China's Shandong First Medical University helps scientists understand how genes turn on and off. Called TRAPT, it maps gene control with record-breaking accuracy.
Good Morning from San Francisco,
Reuters broke the story: Big Brother? So 1984. Now it's Big Grok watching federal workers' every keystroke. 🕵️ Musk's DOGE team reportedly unleashed their AI bloodhound to sniff out anti-Trump sentiment at the EPA, where 600 employees are already practicing their stress-breathing techniques.
The operation skulks in Signal's disappearing messages and rogue Google Docs - because who needs those pesky federal records anyway? 🤫 EPA bosses' new mantra: Think twice, type once.
Meanwhile, Shopify's CEO Tobi Lutke just invented a new job interview question: "Have you tried asking ChatGPT?" 💥 His teams must now prove they're better than algorithms, or watch their jobs vanish faster than those Signal messages. The new corporate dance? Robot-human tango, and AI leads. 🤖💃
Stay curious,
Marcus Schuler
Reuters reports that Trump officials told federal employees that Elon Musk's DOGE team uses AI to monitor their communications for anti-Trump sentiment. The surveillance reportedly targets workers at agencies like the EPA, where massive budget cuts have already put 600 people on leave.
The team deploys Musk's Grok AI chatbot and communicates through Signal, an app that can auto-delete messages. This practice may violate federal record-keeping laws. DOGE staffers also collaborate in Google Docs, bypassing traditional document control processes - according to Reuters.
EPA managers warned their teams to watch what they say and type. Trump appointees made it clear: The AI looks for language critical of Trump or Musk, hunting for workers whose attitudes don't align with the administration's goals.
Musk frames DOGE as an efficiency campaign to trim $1 trillion in government spending. Critics see a different story: A billionaire's tech being used to purge career civil servants and install loyalists.
The operation raises red flags for experts. Government ethics specialist Kathleen Clark calls it worrying: "It sounds like an abuse of government power to suppress speech the president doesn't like."
Why this matters:
Read on, my dear:
Shopify wants proof that AI can't handle the work before adding new employees. CEO Tobi Lutke announced this policy shift in a company memo posted to X on Monday.
The e-commerce giant now expects every employee to master AI tools. The stakes? High. "Stagnation is slow-motion failure," Lutke warns. "If you're not climbing, you're sliding."
Teams must show why AI agents can't tackle their tasks before requesting more headcount. The company will also factor AI usage into performance reviews. Lutke's seen teams use AI to complete "implausible tasks" at 100 times normal speed.
The shift comes as tech giants pour billions into AI development. Shopify has launched its own AI tools - Sidekick chatbot and Shopify Magic automation suite. Meanwhile, the company trimmed its workforce from 8,300 to 8,100 last year, following deeper cuts in previous years.
Money saved on headcount might fund AI talent. CFO Jeff Hoffmeister told investors that while overall staffing stays flat, costs could rise as "high-end AI engineers" command premium salaries.
Why this matters:
Read on, my dear:
Prompt:
art by robert doisneau, masterpiece of graphic photography, natural light, marylin monroe in in the workshop of Monsieur Christian Dior in parisx, black and white, poetic and surreal, very contrasted, street portrait 100VI & XPro3, Very high definition, lots of finesse and detail
Nvidia found a convenient loophole in its tariff troubles. About 60% of its AI datacenter servers dodge recent U.S. tariffs thanks to production in Mexico. The escape route comes through the USMCA trade agreement, which exempts digital processing units made in Mexico from these new charges.
An OpenAI engineer just dropped an uncomfortable truth: AI tools might make smart people smarter and everyone else dumber. Will Depue says while he uses ChatGPT daily to learn new things, he worries many people will let AI do their thinking for them - essentially turning their brains into couch potatoes.
The White House just ripped up Biden's AI rulebook. A new directive orders federal agencies to adopt AI fast and ditch "unnecessary bureaucratic restrictions." Agencies now have six months to clear obstacles and roll out AI applications - marking a dramatic shift from the previous administration's cautious approach.
Anthropic hired Guillaume Princen, Stripe's first employee outside the U.S., to head its European expansion. The AI company plans to add over 100 jobs across Dublin and London this year, marking a major push into the region.
Framework just pulled its cheapest laptops from the U.S. market. The company, known for its repairable computers, suspended sales of two models starting at $749 after new tariffs hit on April 5th. Framework says these laptops will still sell elsewhere, but won't say when - or if - they'll return to America.
Infineon Technologies will buy Marvell's automotive ethernet business for $2.5 billion. The German chipmaker aims to grow its car tech division, which should bring in $250 million next year. Marvell's shift to focus on AI made the sale attractive at 10 times sales.
Google added image search to AI Mode, its newest search feature. Users can now upload photos or snap pictures directly in search. The system combines Google Lens for object detection with Gemini AI to understand and answer questions about images.
The tech industry's attempt to buy Trump's favor has failed. Despite millions in inauguration donations and Mar-a-Lago visits, new tariffs threaten their supply chains. Big Tech's market value has dropped 22% since January as the president pushes aggressive regulation and trade barriers.
Inkitt's CEO says nobody will pay for AI-written books but reveals a plan to use AI-generated author personas. The romance publisher, backed by major Silicon Valley investors, already uses AI to edit and rewrite novels. Now it aims to reduce human authors to providing "bullet points" for AI to transform into books.
"You wake up to discover all mirrors in your home now function as windows into parallel universes. Describe the three most fascinating worlds you see and how you're trying to determine if you can interact with them."
Meta just pulled off the AI equivalent of wearing rocket boots to a footrace. The company's new Llama 4 Maverick model blazed to second place in prestigious AI rankings - but there's a catch.
It turns out Meta used a special "experimental" version of Maverick for the rankings, different from the one available to the public. The company buried this detail in the fine print, like a terms of service agreement nobody reads.
LMArena, the benchmark site in question, wasn't amused. They're now updating their policies to prevent such creative interpretations of their rules. Meta's response? A spokesperson helpfully explained they "experiment with all types of custom variants." How reassuring.
The timing raised eyebrows too. Meta dropped this news on a Saturday, when Mark Zuckerberg says "it was ready." The AI community found this about as convincing as a cat's excuses for knocking things off tables.
Why this matters:
Read on, my dear:
Hugging Face transformed from a quirky chatbot startup into AI's central collaboration hub. Founded in 2016, it now hosts over a million models on its platform where researchers and developers share cutting-edge machine learning resources.
Founded in 2016 by Clément Delangue, Julien Chaumond, and Thomas Wolf in New York City. Originally created as a teen-focused AI chatbot app before pivoting to an open-source machine learning platform in 2018. Now employs roughly 170 people (as of 2023).
Core platform includes:
Competes with proprietary API providers like OpenAI and Anthropic, plus smaller model hubs. Hugging Face's moat comes from its enormous community (10,000+ organizations use its tools) and neutral position - it's "Switzerland" in the AI landscape, collaborating with everyone from Google to Meta.
Raised nearly $400M in funding, with a $4.5B valuation after its 2023 Series D. Key investors include Salesforce (lead), Google, Amazon, Nvidia, Intel, AMD, IBM, and Qualcomm. Revenue grew from under $10M in 2021 to an estimated $70M in 2023.
Hugging Face sits at the nexus of AI's most important trends. With big tech giants as both investors and partners, it's positioned to become AI's essential infrastructure layer. The company's focus on democratizing machine learning perfectly aligns with growing demand for accessible, customizable AI. Open-source momentum shows no sign of slowing. 🚀
Good Morning from San Francisco, OpenAI wants to play social media kingmaker. Sam Altman's crew is quietly testing
Good Morning from San Francisco, OpenAI just fired another shot in the AI race 🎯. Their new GPT-4.1 devours text
☀️ Good Morning from San Francisco, 🔥 OpenAI's Sam Altman just threw GPT-4 under the bus. He labeled their $100
Good Morning from San Francisco, Brussels just pulled out its secret weapon against Trump: hitting Big Tech's wallet.
Fuel your morning with AI insights. Lands in your inbox 6 a.m. PST daily. Grab it free now! 🚀
