They managed to cut the size of the AI reasoning model by more than half—and claim it can now answer politically sensitive questions once off limits in Chinese AI systems.

Technology Review, published by MIT, is a leading source of news, analysis, and commentary on emerging technologies, scientific advancements, and their societal impacts. Covering topics such as artificial intelligence, biotechnology, renewable energy, and more, Technology Review provides  reporting and  articles on the intersection of technology and society. Readers can stay informed about the latest breakthroughs and trends shaping the future of technology.

MIT Technology Review

Multiverse Computing, a Spanish quantum AI firm, created DeepSeek R1 Slim—a compressed version of the Chinese reasoning model that's 55% smaller while maintaining similar performance. Using tensor networks borrowed from quantum physics, researchers claim to have removed built-in censorship that prevented the original model from answering politically sensitive questions. The compression technique maps correlations in the model to selectively remove specific information with precision. While this demonstrates a novel approach to model compression and bias removal, experts caution that fully eliminating censorship from models trained on controlled data may be more complex than testing against a small question set suggests.

Quantum physicists have shrunk and “de-censored” DeepSeek R1