MIT Technology Review's daily newsletter covers several tech stories: Goodfire's new Silico tool uses mechanistic interpretability to let developers debug and adjust LLM parameters during training; Elon Musk admitted xAI trained Grok on OpenAI models via distillation; a new Christian cell phone network launching in the US with mandatory porn blocking and optional gender-content filters; Trump's mass firing of NSF scientists; China's open-source AI strategy gaining developer goodwill after DeepSeek's R1 release; and Huawei set to overtake Nvidia in China's AI chip market.
Table of contents
This startup’s new mechanistic interpretability tool lets you debug LLMsWith mass firing, Trump deals a fresh blow to American scienceSort: