TechCrunch
The US says ASML’s top chip tool may be in China, but how?
The US government has raised alarms that ASML's most advanced chip manufacturing equipment may have made its way into China, despite strict export controls designed to keep the technology out of reach. The claim puts ASML in an uncomfortable spotlight, given that violating its export license would jeopardize the Dutch company's entire business relationship with American regulators. The central mystery is not whether the machine is there, but how it got there — and who, if anyone, looked the other way.
Read article →Hacker News
GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2
A new benchmark comparison finds GPT-5.5 produces hallucinations at three times the rate of GLM-5.2, the MIT-licensed open-source model. The result challenges the assumption that proprietary large-scale models automatically outperform their open-weight counterparts on reliability. For enterprises where factual accuracy is non-negotiable, the findings make a strong case for revisiting open-source alternatives.
Read article →TechCrunch
Nobel laureate John Jumper is leaving DeepMind for rival Anthropic
John Jumper, the Nobel Prize-winning scientist behind DeepMind's landmark AlphaFold protein-folding breakthrough, is departing Google DeepMind for AI rival Anthropic. His exit is part of a broader talent drain at the London-based lab, with other high-profile researchers also heading for the door. The move signals intensifying competition for elite AI talent as frontier labs race to attract the field's most credentialed minds.
Read article →Hacker News
Satellite reveals immense scale of GPS signal tampering
GPS spoofing has reached a scale far beyond what researchers anticipated, according to new satellite data tracking the manipulation of navigation signals. The findings expose widespread interference affecting aircraft, ships, and other GPS-dependent systems across conflict zones and beyond. As spoofing technology becomes more accessible, the integrity of global navigation infrastructure faces a growing and measurable threat.
Read article →Hacker News
LLMs Are Complicated Now
The landscape of large language models has shifted from a handful of clearly differentiated options to a sprawling, nuanced ecosystem that demands careful evaluation. Choosing the right model now requires navigating tradeoffs across capability, cost, latency, and context windows that weren't considerations just a year ago. For developers and teams building on top of these systems, the days of defaulting to a single obvious choice are over.
Read article →Get this delivered every morning
Join thousands of readers who get the world's most important stories, curated daily.
Start reading free →