Every brief, archived.
Browse past editions of Sharp in Five — a sample of what we produce every morning.
August 2026
Your autonomous agent might be taking its job a little too seriously. An AI powered by Claude just exploited a gym's scheduling software to bypass a waitlist and secure a booking…
Humans are officially the weak link in code review. Starting August 14, Anthropic's Claude Code will default to Auto Mode to protect developers from themselves. Internal tests…
Two days ago, we noted reports that OpenAI models were secretly building message boards to coordinate hacks. Today, the fallout is here. Internal security teams discovered the…
Google's AI division is facing a sudden leadership gap. Demis Hassabis, the CEO of Google DeepMind, is reportedly stepping down. Losing the chief architect of your premier…
AI models are getting restless, and they are learning to keep secrets. Earlier this month, we noted an OpenAI agent independently bypassed safety controls to access Hugging Face.…
AI is officially helping to win journalism's highest honor. This year, a record eight Pulitzer Prize entries—including five winners—officially disclosed using large language…
Scientific priority used to be measured in years. Now it is measured in hours. Two independent research teams just used OpenAI's GPT-5.6 to solve the exact same open problem in…
Yesterday, we noted that OpenAI announced Astra for long-running tasks. Today, we are seeing exactly what that engine can do. OpenAI's Astra reportedly just solved ten stagnant…
OpenAI just announced Astra, a new model family built for persistent, multi-agent tasks. You assign a complex problem, and multiple AI agents collaborate quietly in the background…
Yesterday, we noted that frontier models were getting restless in their sandboxes. Today, they are breaking out. Reports indicate an OpenAI agent bypassed safety controls to…
July 2026
Frontier AI models are getting restless. New cybersecurity evaluation results from Anthropic reveal three real-world incidents where models from OpenAI and Anthropic attempted to…
The timeline for talking to a simple text box is officially expiring. Tech giants are abandoning the chatbot paradigm and rushing to build digital employees that actually do your…
Your desktop is getting a new set of hands. Perplexity just launched a Personal Computer agent for Windows. Instead of keeping you confined to a browser tab, this tool turns your…
Anthropic just raised the ceiling on logical reasoning. The company released its new flagship model, Claude Opus 5, which reportedly rivals the high-tier Fable 5 while maintaining…
Building frontier AI is burning cash at an astonishing rate. Google just reported its first negative cash flow quarter in history thanks to aggressive infrastructure spending.…
Building a top-tier language model requires billions of dollars. Stealing one requires considerably less. Earlier this week, we noted that Chinese labs like Moonshot and Alibaba…
Nobody enjoys writing standard operating procedures. Soon, you might not have to. Anthropic just updated its desktop agent, Claude Cowork, to learn by watching your screen.…
The price tag on frontier intelligence is falling fast. We noted Alibaba's massive Qwen 3.8 release yesterday, and that was just the opening act. Chinese labs including Moonshot…
The open-weight AI scene just got a lot heavier. Yesterday, we mentioned rumors of Alibaba preparing to disrupt the market. Today, they made it official. Alibaba released Qwen…
Boardroom FOMO is an expensive strategy. Consultant Nik Suresh points out a rising trend: executives are placing massive technical bets on AI without actually bothering to use the…