Meta launched an AI app that uses your Facebook and Instagram history to personalize responses from day one. Built on their Llama 4 model, the app transforms years of social media data into an AI that claims to understand your preferences, habits, and interests.
Amazon plans to show customers exactly how much Trump's tariffs inflate their shopping bills. The move sparked immediate backlash from the White House, marking a dramatic shift in the previously warming relationship between Trump and Amazon founder Jeff Bezos.
Alibaba released eight new AI language models Monday that match or beat the performance of top competitors like Google and OpenAI. The new Qwen3 family ranges from compact 0.6-billion-parameter models to a massive 235-billion-parameter version.
Good Morning from San Francisco,
The tech world's most intriguing power couple hits turbulent waters. Microsoft and OpenAI, once joined at the hip, now eye each other warily across the dinner table.
Their fairy-tale romance began in a Sun Valley stairwell in 2018. Five minutes with Satya Nadella convinced Sam Altman to accept Microsoft's billions. ChatGPT turned that bet into gold. 💰
Now Microsoft courts Altman's rival Suleyman while secretly plotting independence. OpenAI demands more computing power. They've traded spontaneous texts for rigid weekly calls – like a marriage running on autopilot. 🤖
Both still hold the nuclear option: Microsoft could block OpenAI's restructuring, while OpenAI might cut off access to its latest tech. Talk about relationship leverage. 💣
Stay curious,
Marcus Schuler
The tech world's most famous marriage is heading for couples therapy. Microsoft and OpenAI, who sparked the AI revolution together, now find themselves drifting apart as their interests diverge, the Wall Street Journal reports.
Their partnership started with a chance encounter in a Sun Valley stairwell in 2018. That five-minute chat between Microsoft's Satya Nadella and OpenAI's Sam Altman led to Microsoft pouring billions into OpenAI. The investment paid off spectacularly when ChatGPT exploded onto the scene in 2022.
But now tensions simmer. Microsoft hired Altman's rival Mustafa Suleyman and launched secret projects to reduce dependence on OpenAI. Meanwhile, OpenAI wants more computing power and chips from Microsoft, who claims they're giving everything possible.
The companies still hold significant leverage over each other. Microsoft can block OpenAI's profitable restructuring plans, potentially costing them billions. OpenAI's board could cut off Microsoft's access to their latest tech.
Even their communication style has changed. Gone are the days of rapid-fire text exchanges between the CEOs. Now they stick to scheduled weekly calls, like an aging couple trying to maintain appearances.
Why this matters:
Read on, my dear:
Wall Street Journal: Altman and Nadella, Who Ignited the Modern AI Boom Together, Are Drifting Apart
Prompt:
a symmetrical front-facing photo of an asian model with big botox-filled lips posing with clear neutral lighting and no shadows
Americans don't want robots writing their news. A new Pew survey reveals widespread skepticism about artificial intelligence in journalism, with half of U.S. adults predicting AI will harm news quality over the next 20 years.
The numbers paint a grim picture. Only 10% see AI improving news, while 59% expect it to slash journalism jobs. The remaining respondents split between seeing mixed effects or throwing up their hands in confusion.
Education shapes these views dramatically. College graduates show more skepticism than high school graduates - 56% versus 44% predict negative impacts. It seems the more you know, the more you worry.
The findings cross political lines. Republicans and Democrats, usually divided on media trust, share remarkably similar concerns. About two-thirds of both groups fear AI will spread misinformation.
When it comes to writing articles, humans still win. Four in ten Americans say AI would do worse than professional journalists at crafting stories. Only 19% think the bots could outwrite humans. Perhaps HAL 9000 should stick to chess.
Why this matters:
Read on, my dear:
OpenAI adds shopping to ChatGPT's arsenal but swears they won't take money from retailers - for now. The AI chatbot will scan billions of web searches to suggest products, then send shoppers to merchant websites to seal the deal.
Duolingo CEO Luis von Ahn announced the company is going "AI-first," which means gradually phasing out contractors whose work can be handled by artificial intelligence. In a move echoing Shopify's recent strategy, the language learning company will now require teams to prove they've exhausted AI solutions before requesting additional human headcount - presumably because robots don't need lunch breaks or complain about owl mascots.
Amazon wants steep discounts from suppliers to offset Trump's new China tariffs, leaving vendors with a stark choice: cut prices or lose their biggest customer. In a move that would make even its owl-mascoted rival blush, the tech giant is demanding double-digit price cuts while offering to absorb just a third of tariff costs on non-Chinese goods - proving that when it comes to supplier negotiations, Amazon's bargaining power remains as subtle as a gorilla in a china shop.0
Amazon just fired its first shot in the internet space race, launching 27 Kuiper satellites to challenge SpaceX's Starlink dominance. While SpaceX already has 8,000 satellites up there beaming internet to Earth, Amazon plans to catch up by throwing $10 billion at the project - though at this rate, they might need Prime delivery to meet their mid-2026 deadline.
The Electronic Frontier Foundation and top cybersecurity experts are demanding Trump halt his investigation of former cyber chief Chris Krebs, who was fired in 2020 for debunking election fraud claims. While Trump wants Krebs's security clearance revoked, the experts argue this sets a dangerous precedent - essentially threatening security researchers to either play along with political narratives or face career destruction, which seems like a cybersecurity strategy straight out of "How to Lose Friends and Alienate Security Experts."
Hugging Face just launched a $100 robot arm that you can 3D print yourself and train to pick up Legos. While the SO-101 arm sounds like a bargain, buyer beware - once you add assembly fees and Trump's China tariffs, that $100 could balloon to $500, making this less "impulse buy" and more "arm and a leg."
Pony AI says it can now build self-driving taxis for $41,000 instead of $137,000, bringing it closer to making money on each new robotaxi. But with just 300 cars on the road and needing 50,000 for real profits, this feels less like a victory lap and more like the start of a very long autonomous journey.
An AI that knows when to pause and think - and when to shoot from the hip? That's Qwen3, released today by the Qwen team. This new language model packs dual personalities: a thoughtful mode for tackling complex problems and a rapid-fire mode for simple queries.
The model runs on a hefty 235 billion parameters but only activates 22 billion at a time. Think of it as having a big brain but only using the parts it needs. This clever design helps it compete with heavyweight models like DeepSeek-R1 and Gemini-2.5-Pro while keeping energy costs in check.
What sets Qwen3 apart is its linguistic range. It speaks 119 languages, from widely-used ones like English and Mandarin to less common tongues like Lombard and Friulian. And yes, it even understands Klingon. (Just kidding about that last one - testing if you're paying attention.)
The team trained Qwen3 on 36 trillion tokens - double the data of its predecessor. They fed it everything from web pages to textbooks, using older AI models to help digest and clean up the information. It's like having a younger sibling learn from both parents and older siblings' mistakes.
Why this matters:
Read on, my dear:
Pony.ai burst onto the self-driving scene in 2016, fusing Silicon Valley innovation with Chinese market access. The startup has transformed from ambitious upstart to publicly-traded robotaxi pioneer with operations spanning two continents.
2025 marks Pony.ai's entry into mass-produced robotaxis with three new models unveiled. The company's dramatic 70% cost reduction puts "single-unit breakeven" within reach - a game-changer for profitability. CEO targets full profitability by 2029, with unique U.S.-China footprint providing strategic advantage as autonomous transport shifts from novelty to necessity. 😎
Good Morning from San Francisco, Meta's AI chatbots slipped into dangerous territory. 🚨 The Wall Street Journal busted them
Good Morning from San Francisco, Baidu unleashed chaos in China's AI market. The search giant slashed prices on
Good Morning from San Francisco, While Apple keeps polishing its promises, Perplexity just crashed the voice AI party 🎭. Their revamped
Good Morning from San Francisco, ☀️ Tech giants just can't catch a break in Europe. The EU slapped Apple
Fuel your morning with AI insights. Lands in your inbox 6 a.m. PST daily. Grab it free now! 🚀
