Aleph Alpha's latest Model playing Doom (tesseracted.com)

Aleph Alpha has unveiled its latest AI model, Kolibri-1, which demonstrates its capabilities by playing the classic video game Doom in a shared live environment. The game features two rotating maps where Kolibri-1 makes decisions in real-time, choosing movement, aiming, firing, and interacting through a unique system of evaluated probabilities. Importantly, Kolibri-1 operates using structured game state data rather than visual inputs, which allows for a more focused decision-making process. This development is significant for the AI/ML community as it showcases a streamlined approach to decision-making in dynamic environments. By relying on probability evaluations rather than raw sensory input, Kolibri-1 highlights a shift towards more sophisticated AI models that can effectively navigate complex scenarios. Technical details such as the exclusion of unavailable targets and the handling of persistent exploration mechanics underscore the model's potential for tactical decision-making, emphasizing its adaptability and real-time processing abilities. This could lead to advancements in various applications, from gaming to robotics, where similar real-time decision-making is crucial.