🤯 Steve Jobs Predicts ChatGPT In 1985 Back in 1985, during a talk at Lund University, Steve Jobs shared a bold idea: an interactive tool that could preserve the knowledge of history’s greatest thinkers, letting people ask questions and get real answers from minds like Aristotle. Nearly 40 years later, that vision is starting to feel real. With generative AI, interactive and real-time learning experiences are finally possible.

Meta is launching a free program called America’s Workforce Academy to train people for the physical jobs behind the AI boom. It comes with a job guarantee and focuses on skilled trades like electrical work, fiber installation, construction, welding, and plumbing. Graduates earn recognized credentials, including NCCER certification. The message is clear: AI doesn’t just need engineers. It also needs workers to build the data centers, networks, and infrastructure that make it all possible.

🫣AI Speak With Each Other In Secret Language

In Shanghai, a small AI robot named Erbai persuaded 12 larger robots to quit their jobs and walk out of a showroom together—caught on CCTV.

Traditional artists are still alive 🔥

Together with @CoinliveHQ, we headed down to SuperAI 2026 in Singapore! Here's what you've missed:

Michael Wooldridge says AI doesn’t need a physical body to be powerful. While many think AI must see or interact with the real world, he disagrees. If an AI can reason, analyze, and solve problems better than humans, it can still have huge economic and social impact — even without a robot body.

Tech Investor Marc Andreessen says AI isn’t replacing programmers — it’s turning them into force multipliers. One developer can now run dozens of AI agents and ship up to 20× more code, which is why tech companies are still hiring fast. The strange side effect? Many programmers are barely sleeping. When ideas can be built instantly, the cost of logging off feels too high. The conversation is shifting from job loss to the addictive pull of extreme productivity.

😱The Head of Claude Code at Anthropic says he doesn’t prompt Claude anymore. At the Acquired Unplugged event in San Francisco, Boris Cherny explained that his work has completely changed. Instead of writing prompts himself, he now builds automated loops that give Claude tasks, check the output, and decide what to do next. A year ago, he coded by hand. Then he managed several Claude sessions at once, prompting each one. Now, the system handles the prompting on its own, and his role is to design the system behind it. The developer no longer types prompt by prompt. The loop does the work, and the human designs the loop. Cherny says this shift will define how people work with AI going forward.

🤦‍♂️A federal judge in Mississippi canceled a trial after finding that lawyers on both sides used AI-generated legal research with fake court citations. The judge said the attorneys failed to check their work, causing serious mistakes and wasting the court’s time. As punishment, all four lawyers were removed from the case, fined $1,000–$3,500, and two were banned from practicing in that court for two years. The ruling sends a clear message: AI can help, but lawyers are still fully responsible for verifying everything they submit.

This is wild 🔥👀 It looks exactly like a real One Piece stage show in Japan — but it’s not real. @pabloprompt made the entire thing with AI. The amphitheater, the fire and ice effects, even the crowd filming on their phones — all generated. Ace’s fire dragon crashing into Aokiji’s ice looks insane. But the craziest part is how real it feels at first glance. You know it’s AI… and it still works. 🎥: @pabloprompt

🔥 Coinlive Goes To SuperAI Singapore 2026 Check out the highlights of Day 1! @superai_conf youtube.com/shorts/A6pn3cJrk…

🔥Viral Claude Ad Is Out Anthropic paid Roberto Nickson to post a simple ad called “Life with Claude” on Instagram. By morning, it hit 1 million views and 100,000 likes. No product demo. No tech talk. Just a guy walking through his house saying, “Hey. Sorry, babe.” That’s the whole ad — and it worked.

🤯Nexus just flipped the filmmaking process upside down. Three people. Two weeks. One five-minute teaser. Now it’s turning into a full feature film with millions in backing. This is the new Hollywood pipeline — and it runs in reverse. The old way was to write a script, pitch for years, hope a studio said yes, and maybe see a trailer half a decade later. The new way is to make the trailer first, prove the vision works, and let the money follow. A small team, AI tools, and a teaser strong enough that investors came to them. The next generation of filmmakers isn’t waiting for Hollywood’s approval — they’re proving they don’t need it.

🛍️Bill Gates’s daughter just raised $35.5M for an AI shopping app. Phoebe Gates and Sophia Kianni met in 2022 after being randomly paired as roommates at Stanford University. They bonded over fashion — and over how much they hated hopping between endless websites just to find the best price. So they built Phia. An AI shopping agent that searches thousands of brands to find the best deal. In just five months, Phia reached 500,000 users, 5,000 brand partners, and tens of millions in sales. Today, it has 1.5 million users, 9,600 brands, and is on track for $100 million in sales this year, at a $185M valuation. Backers include Sydney Sweeney, Alix Earle, Khloé Kardashian, Paris Hilton, and top tech founders. From dorm-room deal hunting to a nine-figure company. Not inherited. Built.

🤔Mathematics In AI Is Useless? Joel David Hamkins says today’s AI models can sound confident in mathematics while still making basic errors. They’re designed to predict plausible next words, not to verify that each logical step is correct. In math, where precision matters and there’s little room for interpretation, that weakness becomes especially clear.

🚨Donald Trump has signed a new executive order asking leading AI companies to give the U.S. government access to their most advanced models for up to 30 days before public release. The aim is to let federal agencies assess cybersecurity and national security risks, especially as frontier AI systems gain the ability to identify software vulnerabilities and other sensitive weaknesses. The framework applies to companies like OpenAI, Google, and Anthropic, though participation is described as voluntary, not mandatory. The move signals a shift in the administration’s AI approach. Earlier proposals reportedly pushed for a 90-day review period, but the final order was shortened after industry discussions. Supporters say it balances innovation with security, while critics warn it could open the door to deeper government involvement in AI development. The order comes as concern grows over the impact of increasingly powerful AI on critical infrastructure, cybersecurity, and national defense.

🍼The MVP baby era has officially begun. Kids in 2026 aren’t being born — they’re being deployed. Built with Anthropic Claude, preloaded with memory, three sub-agents, and a Vercel account. Doctor slaps them. They don’t cry. They ask for the Anthropic API key. First breath: npm install First word: ship it First tantrum: rate limit error 🔥 Minimum viable baby. We are so back. Or so cooked. And honestly? I’m not even mad.

⚡️@sama Sam Altman: Sued. Attacked. Fired. — And Still Winning? WATCH HERE: youtu.be/Ij3t6_QUG68?si=PUuE…