Anthropic cited OpenAI's use of Claude Code “ahead of the launch of GPT-5” but says it will provide API access to OpenAI for benchmarking and safety evaluations  —  Anthropic says it has revoked OpenAI's access to the Claude API after ChatGPT's engineers were found using Claude's coding tools.

SentinelOne researchers detail over 10 offensive cybersecurity patents filed by Shanghai Firetech, a company allegedly involved in Beijing's Silk Typhoon hacks  —  Researchers have discovered more than 10 patents for powerful offensive cybersecurity technologies filed by a prominent Chinese …

How VCs are changing the way people learn about the tech industry, through a sprawling media ecosystem consisting of podcasts, newsletters, and streaming shows  —  Podcasts hosted by venture capitalists set the narrative in tech, changing the way that insiders—and everyone else—learn about the world of startups.

China's World AI Conference struck a stark contrast to Trump's America-first, regulation-light vision, focusing instead on global cooperation for AI safety  —  Behind closed doors, Chinese researchers are laying the groundwork for a new global AI agenda—without input from the US.

Sources: the researcher who declined Zuckerberg's offer of a $1.5B+ package over six years is Thinking Machines co-founder and ex-Meta staffer Andrew Tulloch  —  The loyalties and larger-than-life figures prompting some people to turn down insanely lucrative offers in Silicon Valley right now

How Palantir has benefited since President Trump took office, securing new and expanded government contracts and becoming the S&P's top-performing stock of 2025  —  From warfighting and tracking deportations to writing State Department cables with AI, the company is benefiting …

Manus unveils Wide Research, an experimental feature that lets users on its Pro plan enlist dozens of parallelized AI agents for large-scale, high-volume tasks  —  Chinese AI startup Manus, which made headlines earlier this year for its approach to a multi-agent orchestration platform for consumers and …

Microsoft's latest 10-K shows that the company's committed datacenter construction is down YoY, with almost no new leases signed in the past 12 months  —  Microsoft datacenter strategy - still in⏸️mode 🧵 @JeremieEO Four months ago we flagged Microsoft's pause on more than 1.5 GW of self-build capacity — on top of the leasing freeze we called out in December. Microsoft's latest 10-K backs us up: committed datacenter construction [image]

Lina Khan cites Figma IPO as vindication of M&A scrutiny, saying letting startups grow independently, rather than be acquired, can generate “enormous value”  —  A surprising figure is celebrating Figma's successful IPO: Lina Khan, former chair of the Federal Trade Commission.

Anthropic details “persona vectors”, patterns of activity within an AI model's neural network that control its character traits, such as evil and sycophancy  —  Read the paper  —  Language models are strange beasts.  In many ways they appear to have human-like “personalities” …

Signatories to the EU's voluntary AI Code of Practice include Amazon, Anthropic, Google, Microsoft, and OpenAI, while Chinese AI companies remain absent  —  Nearly all dominant EU and US AI developers have signed up, but - so far - there are none from China.

Analysis: since April 2, President Trump has excluded over $100B in US smartphone and laptop imports from tariffs, highlighting conflicting administration goals  —  Exclusions reflect conflicting goals: to rip up the rules of global trade while cushioning US consumers.