In the ecosystem of StarCraft 1, a game that has enjoyed immense love for nearly 30 years since its domestic release, a fresh breeze has recently swept through with the arrival of 'AI Pluto.'

While it was a major player in signaling the birth of esports and has now settled as a national pastime, the emergence of AI Pluto is reigniting the stagnant StarCraft scene. Even though some time has passed since its arrival, videos related to Pluto continue to draw high view counts on former pro gamers' YouTube channels and online communities.

Pluto is a StarCraft: Brood War AI created single-handedly by Norwegian developer Vegard Mella. Trained through 'self-play reinforcement learning'—learning autonomously by playing against itself rather than using human match records—a single neural network handles all three races: Terran, Zerg, and Protoss. It made a name for itself after winning the 'CoG 2026 StarCraft AI Competition' in August, and when the competition model was released on GitHub on September 23, anyone could compete against Pluto on a standard PC. It was from this point that former pro gamers rushed to upload videos of their matches against Pluto.

As direct matchups against humans became more frequent, perspectives on AI Pluto have divided sharply. Some express amazement, noting that as an AI trained on self-play under certain physical constraints, "it presents a new paradigm in a game over 30 years old." On the other hand, some point out its limitations, arguing that "due to fundamental hardware specification differences, a matchup against humans is an inherently unfair competition."

Opinions among former pro gamers are also varied. Former pro and current streamer A said, "While AI Pluto is fresh, it doesn't feel at all like humans have been completely defeated by AI. Rather, it feels like content to enjoy for fun; I didn't get the sense that it demonstrates designs superior to humans in hard-to-explain areas like builds or psychological warfare appropriate for a real-time strategy game. However, its combat capabilities that humans can hardly replicate—such as damage distribution through focus fire and reaction speed—are definitely overwhelming."

Another former pro gamer, B, had a slightly different view. He said, "Even though this is a game that has been around for a long time, it gave us a chance to reflect on whether we were stuck inside a box. The builds and movements shown by AI Pluto are truly surprising. From defending against Mutalisks without Missile Turrets to a full package including mixing in bio units in Terran vs. Terran, there are quite a few aspects that shatter fixed preconceptions. While some say these are only possible because it's an AI, as a shift in perspective, it is well worth studying."

And former pro gamer, YouTuber, and broadcast personality Maengdeok Eomma (Kim Jeong-hoon), who creates content casting Human vs. AI Pluto matches, said, "AI Pluto's fast and precise control is the most impressive. However, it is difficult to conclude based on that alone that it is superior to human strategy, or conversely, that victory was entirely due to control.

Among the matches I cast, scenes like actively using SCVs in Terran vs. Terran to absorb enemy Siege Tank fire, army compositions actively mixing Marines and Medics in Terran vs. Protoss, and defending against Mutalisks without Photon Cannons in Protoss vs. Zerg were memorable, and showing playstyle maneuvers rarely seen in human vs. human matches was interesting.

According to the explanation released by the developer, Pluto is an AI trained through self-play reinforcement learning. I don't know the detailed learning process well, but I think it is meaningful that through the play shown by the AI, humans can gain a fresh perspective on established meta and unit utilization. If people can find points to learn or test for themselves from AI Pluto's play, I think that in itself would be a meaningful achievement as a research project," he added.

However, there are certainly side effects. As AI Pluto generated significant attention, cases have emerged where some users exploit the AI to play against ordinary players on ladder matches or public lobbies, mocking them or disrupting gameplay. Furthermore, illegal trading and selling of the AI are taking place, raising growing concerns over the disruption of the ecosystem.