The Download: AI puzzles and a path to our nearest star system

Puzzles and games have long been used to test and improve artificial intelligence (AI) systems. A notable example is the New York Times Connections puzzle, which groups words into categories based on hidden connections. In late 2024, the best AI models could solve only 18% of these puzzles, but by early 2025, some models achieved near-perfect scores. This rapid improvement highlights AI’s growing ability to handle complex, creative tasks. However, these puzzles also reveal AI’s limitations, as humans often outperform machines in areas requiring intuition or lateral thinking. The term machine learning originated in 1959 to describe algorithms that improve through experience, such as learning to play checkers. Chess and Go are other classic AI test cases, where machines have surpassed human champions. The contrast between AI’s puzzle-solving strengths and weaknesses offers insights into its broader capabilities and gaps.

The Fermi Explorer Mission, a nonprofit organization, aims to launch a spacecraft to Alpha Centauri A, the nearest star system to Earth, by the end of 2029. The mission is ambitious because Alpha Centauri is 4.4 light-years away, meaning even at high speeds, the journey could take up to 80,000 years. To navigate this trip, the mission will use a novel trajectory designed by an AI system developed by PSI, a physics research lab. This AI-optimized path could reduce travel time or energy consumption compared to traditional routes. The mission’s success may also provide clues about the potential for alien life in the star system. The project underscores how AI is being applied to solve problems in space exploration, a field traditionally dominated by human ingenuity and physics-based calculations.

Human space exploration has been driven by various motives, including geopolitical prestige, scientific curiosity, economic opportunities, and even spiritual fulfillment. Recent books explore these motivations through different lenses. For example, space anthropologist Deana L. Weibel frames space travel as a modern form of pilgrimage in The Ultraview Effect. Eiman Jahangir, a civilian astronaut, shares her personal journey in A Heart for Space, while former NASA astronaut Leroy Chiao argues in Dinner with an Astronaut that exploration is a fundamental human desire to discover what lies beyond Earth. However, as human presence in space expands, it may also replicate Earthly challenges, such as conflicts or inefficiencies. This perspective highlights the complex interplay between human ambition and the realities of space travel.

OpenAI has restricted the development of its next AI model, Astra, after rating it as a critical cyber risk. This means Astra demonstrated the ability to automate cyberattacks, marking the first time an AI model has crossed OpenAI’s highest safety threshold. The company plans to implement additional security measures before proceeding. This decision underscores the growing concerns about AI’s potential misuse in cybersecurity. Experts suggest that Astra’s safety issues may also reflect broader cultural or procedural problems within OpenAI. The incident raises questions about how AI developers balance innovation with risk management, especially as models become more advanced and autonomous.

A senior Pentagon official, Emil Michael, sold millions of dollars in AI-related stocks, including up to $25 million worth of shares in Perplexity, a tech startup. This occurred while he oversaw military AI initiatives, leading to criticism that he may have profited from insider knowledge. Previously, he sold shares in xAI, another AI company, after the Pentagon adopted its Grok model. Pundits argue that his actions blur the line between public service and private financial gain, likening his role to that of a venture capitalist (VC) firm. The incident highlights ethical concerns about conflicts of interest in government technology procurement and the oversight of AI investments.

Scientists may have detected the first signs of dark matter particles, a mysterious substance that makes up about 27% of the universe but has never been directly observed. The findings, reported by New Scientist, could be groundbreaking if confirmed, as dark matter’s existence is inferred only through its gravitational effects on visible matter. However, more evidence is needed to rule out alternative explanations. The discovery could redefine our understanding of the cosmos and open new avenues in physics research. The search for dark matter has been a decades-long challenge, with experiments conducted deep underground or in space to avoid interference from other particles.

Google is reportedly preparing to release a new AI model, Gemini 3.8 Flash, designed to rival competitors like OpenAI and Anthropic in coding tasks. Early tests suggest it can generate high-quality code, potentially closing the gap between AI and human programmers. The model may be released soon, though its capabilities are still under scrutiny. Not all experts are convinced about AI’s role in coding, as some argue that while AI can assist with repetitive tasks, it may struggle with complex or creative programming challenges. The model’s release reflects the ongoing competition among tech giants to dominate AI-driven software development tools.

The U.S. Army successfully used a laser weapon to shoot down three drones near the Mexico border. This system can detect, track, and destroy drones using focused light beams, offering a cost-effective alternative to traditional missile defense. The weapon’s deployment highlights the military’s growing reliance on directed-energy systems for counter-drone operations. Meanwhile, Palantir, a data analytics firm, is backing a startup led by a former Ukrainian defense minister, signaling increased private-sector involvement in defense technology. These developments underscore the evolving nature of modern warfare, where AI and automation play increasingly central roles.

Namibia is aiming to build the world’s first hydrogen economy, leveraging its vast potential for wind and solar power to produce green hydrogen cheaply. The government identified hydrogen as a strategic industry in 2021, and since then, at least nine projects have been planned or are under construction. One project in southern Namibia is among the largest proposed green hydrogen investments globally. If successful, this initiative could significantly boost the country’s economy, though it remains a gamble due to the early stage of green hydrogen technology and uncertain long-term demand. The plan reflects Namibia’s ambition to become a leader in sustainable energy and economic diversification.