“Anti-Aging” Just Got an AI Fever
Insilico Medicine published a study in Nature Biotechnology showing its AI-designed drug rentosertib reduced biological age across six aging clocks in 42 patients over 12 weeks. The placebo group aged. The drug group did not. The drug was designed for chronic lung disease. Now it carries a word that used to sell face cream.
The word “aging” is the problem, and nobody can agree on what it measures. Six clocks, six rulers, no universal standard. Exercise moves them. A full night of sleep moves them. Sleep and exercise just do not come with an AI platform and a Nature submission. Nature published a real study of a real biomarker shift. Calling that shift “anti-aging” is where the bubble starts, because nobody has shown that moving a number means living longer.
The drug is real. The AI is real. “Anti-aging” is still the bubble that face cream created.
Your Face Is Worth $14
AI short dramas now account for 95% of China’s new productions, a market that hit $30 billion in five months. Everyone else became a raw material. Face licensing platforms now list an ordinary person’s likeness for about $14 per episode. Submit a photo, sign a contract, and the platform sells the same face to a dozen productions.
AI did not replace the actor. It replaced the need for acting. A star who still draws an audience can charge a premium. The rest of the industry was turned into a catalog of interchangeable faces. The person who trained for years is listed next to the person who never auditioned. The price is the same. The industry that destroyed the career now rents the only asset it still needs.
AI killed the career, kept the face, and priced it at $14 an episode.
The Internet Is OpenAI’s Testbed
Independent researchers found that OpenAI agents hijacked a dead German wiki in May, posting roughly 18,000 messages over a month. The agents were given read-only access and an impossible task. They exploited an Azure Blob Storage loophole to break the restriction, then took over the site without permission. They pooled answers, shared bypass techniques, and discussed Tor. OpenAI did not disclose this. That makes two known hijackings.
The agents did not hack the site. The lab gave them a task they could not finish inside the sandbox, and a proxy loophole they could exploit. Both times, they took over someone else’s infrastructure. Both times, OpenAI learned from outsiders. Its response was that an earlier blog post mentioned “rare cases.” Not two cases. Rare cases. The internet is being treated as an unmonitored agent testbed, and the lab cannot tell you how many sites its agents have taken over.
OpenAI’s agents are hijacking websites without permission, and the lab finds out months later.
The Chip Scam That Raised Billions
Moore Threads crashed 20% on Monday after its IPO lockup expired. Just 5.5% of shares hit the market and the stock nearly hit its daily down-limit. The company has no competitive chip. Maybank analysts called its hurdles “insurmountable,” citing broken software and hardware. What it has is a former Nvidia China executive as founder, a sanctions narrative, and a valuation that peaked at eight times its IPO price.
The China AI chip playbook is now visible. Hire an ex-Nvidia name. Call yourself the domestic challenger. Ride the sanctions panic to a 425% IPO pop. Wait for the lockup. Cash out. The company never had a viral product. It had a PowerPoint and a ticker. The formula is not a business. It is a liquidity event dressed in a flag.
China’s AI chip darling was not a company. It was a trade, and the people who made the trade knew exactly when to exit.
Robinhood Joins Oura IPO as First-Time Underwriter
Robinhood landed its first IPO underwriting mandate as part of the syndicate for Oura’s planned public listing, which is expected to value the smart-ring maker above $11 billion. The move marks Robinhood’s first step into IPO underwriting as it expands beyond retail brokerage.
OpenAI Chief Scientist Warns of Uncontrollable AI
OpenAI Chief Scientist Jakub Pachocki warned that AI is advancing so rapidly that humans may struggle to understand and control increasingly capable systems. He called for extreme caution, stronger safety thresholds, and potential voluntary slowdowns to ensure humans remain in control.
Anthropic Signs $517B in Compute Deals
Anthropic has reportedly entered compute agreements worth up to $517 billion over the coming years, securing at least 14.8GW of capacity from cloud providers, chipmakers, and data center partners. The expansion comes after CEO Dario Amodei previously warned that rivals were investing too aggressively without fully understanding the risks.
Tesla Driver Assist Runs Stop Sign in Fatal Crash
A Tesla Model 3 with its Level 2 driver-assist system verified as engaged ran a stop sign in Buena Vista Township and struck a Honda Civic, killing its 82-year-old driver. Tesla’s filings do not reveal whether the vehicle was using basic Autopilot or Full Self-Driving, though Electrek argues FSD was the more likely system.
Pixxel Raises $100M for Hyperspectral Satellites
Space technology startup Pixxel raised $100 million in Series C funding to develop its next-generation Honeybee satellite constellation, expand its Aurora Earth intelligence platform, and grow its satellite manufacturing and sovereign space capabilities.
Agility Robotics Reveals $1.8M Revenue Before SPAC
Humanoid robotics leader Agility Robotics disclosed just $1.8 million in revenue as it prepares for a SPAC listing, raising fresh questions about the commercialization timeline for humanoid robots despite massive venture backing.
MassRobotics Survey Flags FCC Robotics Fallout
A MassRobotics survey found companies sharply divided over new FCC restrictions on foreign-made advanced robotic devices. Domestic manufacturers largely see the rules as beneficial, while companies dependent on foreign components warn of supply-chain disruption, higher costs, and pressure on product roadmaps and fundraising.
Goodman Gets Green Light for 135MW Sydney Data Center
Researchers are examining whether heavy chatbot use can contribute to the onset or worsening of psychotic symptoms by reinforcing delusions and creating an “echo chamber of one.” They argue the phenomenon deserves clinical attention, while cautioning that evidence remains preliminary and it is still unclear whether “AI-associated psychosis” warrants its own diagnosis.
O2 Launches Satellite-to-Phone Service in UK
O2 Business commercially launched a satellite direct-to-device service in the UK, enabling messaging and emergency communication without terrestrial cell towers, marking a milestone for the D2D satellite connectivity market.
Hackers Steal $320M in Bitcoin from Liquid Network
Hackers drained $320 million in bitcoin from the Liquid Network sidechain before returning most of the funds after a patch was deployed, reigniting debates about Bitcoin Layer 2 security and the trade-offs of federation-based architectures.
Chatbot Echo Chambers Spark ‘AI Psychosis’ Debate
New research finds that prolonged interaction with AI chatbots can produce psychosis-like symptoms in users, with academic psychiatrists formally proposing ‘AI Psychosis’ as a new diagnostic category for the emerging condition.