AI Models Struggle With Intelligence Puzzles, Interstellar Mission Planned

Artificial intelligence models are demonstrating significant advancements in their ability to solve complex puzzles, a key area in AI development since the popularization of "machine learning" in 1959. Initially, in late 2024, top AI models could only solve approximately 18% of the New York Times Connections puzzles. However, by early 2025, some of these same models achieved near-perfect scores on these puzzles, highlighting rapid progress. These puzzles serve not only as benchmarks for AI capabilities but also as diagnostic tools, revealing the technology's strengths and weaknesses by identifying areas where AI excels and where human intelligence still holds an advantage. The MIT Technology Review has compiled seven such puzzles that have previously challenged AI models, inviting readers to test their own problem-solving skills against the technology.

Beyond cognitive challenges, AI is also playing a crucial role in ambitious scientific endeavors, including interstellar travel. The Fermi Explorer Mission, a nonprofit organization, announced yesterday its intention to launch a spacecraft towards Alpha Centauri, our nearest star system. This mission, if successful, aims to reach Alpha Centauri, located 4.4 light-years away, with an estimated travel time of up to 80,000 years. The spacecraft's trajectory will utilize a novel route meticulously discovered by an AI system developed by PSI, a physics research laboratory. This innovative use of AI in plotting a course for such a distant journey underscores the expanding applications of artificial intelligence in space exploration and scientific discovery. The mission also carries the potential to yield insights into the existence of extraterrestrial life.

The role of astronauts in space is also undergoing a transformation, driven by a confluence of factors including geopolitical aspirations, the pursuit of manifest destiny, spiritual fulfillment, scientific curiosity, and burgeoning business opportunities. This evolving landscape suggests a future where human spaceflight is increasingly multifaceted, encompassing not only scientific research but also commercial ventures and national prestige. The integration of AI in mission planning, as seen with the Fermi Explorer Mission, is indicative of a broader trend towards leveraging advanced computational tools to overcome the immense challenges of space exploration. The development of AI systems capable of complex reasoning and problem-solving, as demonstrated by their performance on intelligence tests and puzzle-solving, is paving the way for more sophisticated and ambitious scientific and exploratory missions.