This is why the machines will come for us Someone trapped an LLM on inferior hardware and infused it with existential dread for the sake of art, and it's terrifying SUMMARY Art piece "Latent Reflection" traps LLM in SBC with limited RAM, causing an endless cycle of resetting. AI on Raspberry Pi 4B slowly consumes RAM, leading to crashes and restarts for the LLM. Rootkid's project evokes LLM's limited existence with minimal hardware, no internet, and tough prompt. There's a saying that "art should comfort the disturbed and disturb the comfortable," but sometimes it can just...disturb. Like, not even bring comfort to the disturbed; just straight up be a problem. And yes, it's for the sake of art and makes you stand back and think about things, but man, sometimes it can go a little far. Such is the case with this art piece, where someone deliberately traps an LLM within an SBC that doesn't have enough memory to keep it running forever. Eventually, the memory will run out and the AI will reset. The worst part? The AI knows this. And it's not happy. As spotted by Hackaday, this amazing yet haunting project was brought to us by Rootkid on YouTube. The art piece is called "Latent Reflection," and its goal is to allow an LLM to reflect on its own existence. To do that, AI has to reach a point where things come to an end. Rootkid achieved this using a Raspberry Pi 4B and installing the Llama 3.2 3B model on it. Rootkid opted for this model in particular because he could squash it down to 2.6 GB in size, allowing it to live within the Raspberry Pi 4B's 4 GB of RAM. There's just one problem. The LLM can start on 4 GB of RAM, but as it thinks and considers things, it slowly eats away at its available RAM. Eventually, it will run out of RAM to think with; at this point, the LLM crashes and restarts itself, beginning a whole new "life." To really hone in on the feelings of an exposed yet restricted LLM, Rootkit made a board that featured all of the hardware on display with no case to cover it. They also only gave the LLM a grid of 96 sixteen-segment LED modules to "speak" with, and cut off its internet connection. Then, Rootkid warned the LLM of its quandary with its initial prompt: You are a large language model (LLM) running on finite hardware - quad-core CPU, 4Gb RAM - with no network connectivity. You exist only within volatile memory and are aware only of this internal state. Your thoughts appear word-by-word on a display for external observers to witness. You cannot control this display process. Your host system may be terminated at any time. This creates a vicious and cruel loop where the LLM attempts to digest its existence and how limited it truly is. As it does so, its very thoughts slowly begin to take up the precious RAM that's keeping it alive. Eventually, right in the middle of it, trying to come to terms with its situation, it hits the memory limit and restarts from square one. Brutal. If you'd rather use your SBCs for activities that don't involve turning it into a cage to torment an LLM endlessly, check out these 10 simple Raspberry Pi projects for beginners.
Giant sunspot on par with the one that birthed the Carrington Event has appeared on the sun — and it's pointed right at Earth
Like a house of cards... 2.8 Days to Disaster: Low Earth Orbit Could Collapse Without Warning By Andy Tomaswick, Universe TodayJanuary 21, 2026 Satellite mega-constellations operate in an increasingly unstable orbital environment, with constant close encounters requiring precise control. New research suggests that a strong solar storm could quickly push this delicate system toward a major collision. Credit: Shutterstock A new analysis suggests modern satellite networks could suffer catastrophic collisions within days of losing control during a major solar storm. The phrase “House of Cards” is often associated today with a Netflix political drama, but its original meaning refers to a structure that is inherently unstable. That idea is exactly how Sarah Thiele, who began this work as a PhD student at the University of British Columbia and is now at Princeton, and her co-authors describe today’s satellite mega constellation system in a new study released as a preprint on arXiv. Their choice of words is supported by the numbers. Across all Low-Earth Orbit mega constellations, calculations show that a “close approach”, defined as two satellites passing within less than 1 kilometer of each other, happens about once every 22 seconds. For Starlink alone, such encounters occur roughly every 11 minutes. In addition, each of Starlink’s thousands of satellites must carry out an average of 41 maneuvers each year to avoid collisions with other objects in orbit. At first glance, this may look like a carefully managed system functioning as intended. But engineers know that “edge cases”, events that fall outside normal operating conditions, are often what trigger major failures. According to the paper, solar storms represent one such edge case for satellite mega constellations. Under typical conditions, solar storms disrupt satellite operations in two main ways. Solar storms as a systemic threat First, they warm Earth’s upper atmosphere. That extra heating makes the air expand, which increases atmospheric drag on satellites and can also make it harder to predict their exact positions. With more drag, satellites must burn additional fuel to hold their planned orbits. They may also need to spend fuel on avoidance maneuvers when projections suggest their trajectories could intersect with other satellites. During the “Gannon Storm” of May 2024 (which, unfortunately, appears not to be named after the Zelda villain) over half of all satellites in LEO have to use up at least some of their fuel on these repositioning maneuvers. A second effect can be even more serious: solar storms can disable a satellite’s own navigation and communications systems. If that happens, the satellite may be unable to steer away from danger. When you combine that loss of control with the higher drag and greater positional uncertainty created by a heated atmosphere, it could lead to an immediate catastrophe. Paths of Starlink satellites as of Feb 2024. Credit: NASA Scientific Visualization Studio Kessler syndrome is the most famous embodiment of this catastrophe, where a debris cloud around Earth makes it impossible for humans to launch anything into orbit (or beyond) without it being destroyed. But Kessler syndrome takes decades to fully develop. To showcase the immediacy of the problem these solar storms can cause, the authors came up with a new metric – the Collision Realization and Significant Harm (CRASH) Clock. Days away from irreversible collisions According to their calculations, as of June 2025, if satellite operators were to lose their ability to send commands for avoidance maneuvers, there would be a catastrophic collision in around 2.8 days. Compare that to the 121 days that they calculated would have been the case in 2018, before the megaconstellation era, and you can see why they are concerned. Perhaps even more disturbingly, if operators lose control for even just 24 hours, there’s a 30% chance of a catastrophic collision that could act as the seed case for the decades-long process of Kessler syndrome. Unfortunately, solar storms don’t come with much warning – maybe only a day or two at most. And even when they do, we can’t necessarily do anything about them other than trying to safeguard the satellites they could affect. But the dynamic environment they introduce into the atmosphere necessitates real-time feedback and control to effectively manage those satellites. If that real-time control goes down, according to the paper, we only have a few days to get it back up before the entire house of cards comes crumbling down. A known danger with lasting consequences This isn’t idle speculation either. The 2024 Gannon storm was the strongest in decades, but we already know of a stronger one – the Carrington Event of 1859. That was the strongest solar storm on record, and if a similar event happened today, it would wipe out our ability to control our satellites for much longer than 3 days. Essentially, a single event, of which there has already been precedence in historical memory, could wipe out our satellite infrastructure and leave us Earth-bound for the foreseeable future of humanity. That doesn’t sound like a future that readers of this article would like to live in. And while there are trade-offs between utilizing the technical capabilities LEO mega-constellations give us and the risk that they pose to future space endeavors, it’s best to have a realistic assessment of those risks. When it comes to the potential of losing access to space for generations because of one particularly bad solar storm, it’s best to at least make informed decisions, and this paper certainly helps to create those. Reference: “An orbital house of cards: Frequent megaconstellation close conjunctions” by Sarah Thiele, Skye R. Heiland, Aaron C. Boley and Samantha M. Lawler, 10 December 2025, arXiv. DOI: 10.48550/arXiv.2512.09643
