Happy Friday! Becoming an influencer takes work. At least it used to. A partner at a16z wanted to find the bare minimum needed. So she created Janie, a fictional redheaded teen, and spent $100 plus about 30 mins a day making fake sorority rush videos for her. That was enough. A week later, “Janie” had over 1,000 followers and was racking up hundreds of thousands of views per video. TikTok labeled almost half of them as AI generated, but it didn’t seem to matter. At that price, a creator can run an entire cast of synthetic influencers and let the algo decide who gets famous.
In today’s dose:
- Vibe coding with friends
- China’s data backed loans
- Agents chase lost sales
- Claude cancer drugs
Slack wants you to vibe code with friends. Its new collaborative coding tool lets you call on Claude, Devin or Copilot directly inside Slack. Ask an agent to build a feature, and Slack creates a dedicated code channel around the job. Everyone can follow the work, compare changes, preview the result, and give feedback before anything ships. Once the team approves it, the channel archives itself and leaves behind an audit log. Slack is turning vibe coding from a private chat with an agent into a team sport. Coding by committee sounds like a token bonfire with receipts. (Gizmodo)
China has made data a national strategy. It now runs the world’s largest government backed market for turning company data into money. Say a robotics company has years of factory data showing how machines move and work. China lets it put a price on that data, add it to the balance sheet, and use it to secure a bank loan. That loan buys more robots, which create more data the company can borrow against later. An American rival may collect the same data, but accounting rules treat it as an expense, so the value stays off the balance sheet. China’s market for investments backed by data is now worth $3 billion, 4x what it was in 2025. China is building a loop where data makes AI smarter, then finances the companies collecting more of it. (Tech Times)
The Work AI Index 2026 from Glean’s Work AI Institute surveyed 6,000 digital workers and found that AI saves time, but much of it goes back into cleanup. Read the report to see where AI time savings go and what high AI achievers do differently.
👀 closer look
AI is learning how to close the sale. Flipkart built an agentic sales team that goes after lost business. Say you search for “best gaming phone,” see nothing you like, and close the app. The agents take it from there, figuring out what you wanted and searching the web for better options. They match those to phones in stock at your price, then message you on WhatsApp to bring you back. Over 23 days, the company sent 15,000 of these messages and got nearly 4x as many clicks as older campaigns. More importantly, some people came back and bought the phones the agents recommended. At 2 to 3 cents per search, every lost sale is worth another try. (arXiv)
Claude has entered the race to cure cancer. Scientists handed Claude VEGF A, a protein tumors use to grow new blood vessels, and asked for something that would stick to it. Claude chose where to grab the protein, picked the scientific tools, ran experiments, and sent its best designs to two independent labs. The labs built exactly what Claude sent them, and 54 of its 90 designs worked. Claude repeated the process on other proteins tied to cancer, Alzheimer’s, and inflammation. Across the full campaign, as many as 35% of its designs worked, compared with the drug industry’s usual 10% to 15%. The real benchmark is whether these designs make cancer treatment cheaper. (Anthropic)
fun stats
🦾 80%. Tasks humanoid robots need to handle from simple voice or text commands before physical AI hits its “ChatGPT moment.” Unitree’s CEO thinks that tipping point could be just 2 to 3 years away.
💳 1 in 5. Enterprises that can’t stop a runaway AI agent from spending money in real time. Once it starts swiping the company card, there’s no kill switch.
🧃 52%. S&P 500 profit growth last quarter, juiced by Alphabet and Amazon’s paper gains on AI investments. Take those away and it falls to 33%.