Bytes: Week in Review — Anthropic's new AI model, a referendum on data centers, and NASA livestreams journey to space

This week's Marketplace Tech Bites explores three major developments in technology and space. First, Anthropic unveiled Claude Mythos Preview, a powerful AI model capable of detecting deep security vulnerabilities, but chose not to release it publicly—instead limiting access to over 40 trusted organizations like Google and JPMorgan Chase under Project Glasswing. This marks a growing trend of AI companies self-editing their releases for safety, raising ethical questions about control, transparency, and the future of AI development. Second, Port Washington, Wisconsin, became the first U.S. city to pass a referendum requiring voter approval for large data center tax incentives, reflecting rising public pushback over environmental, energy, and community impacts of massive tech infrastructure. The episode examines how local resistance could shape the future of data center expansion. Finally, NASA’s Artemis II mission set a new record for distance from Earth, with astronauts sharing their journey through continuous livestreams on YouTube, using iPhones to capture stunning images and engaging social media content that has drawn millions. The episode highlights how accessibility and storytelling are revitalizing public interest in space exploration. The tone is cautiously optimistic, emphasizing both the promise and peril of emerging tech. Key takeaways include: AI companies are increasingly self-censoring powerful models for safety; local communities are gaining leverage over data center development through democratic processes; livestreaming space missions enhances public engagement; social media proficiency is becoming an unspoken asset for astronauts; and the balance between innovation and public accountability is more critical than ever in the AI and space eras.