The article argues that artificial general intelligence (AGI) is already upon us, evident through the emergence of coding agents and long-horizon agents that can autonomously tackle tasks. While defining AGI remains a challenge, the authors provide a functional definition: the ability to figure things out. This means having baseline knowledge, reasoning capabilities, and the ability to iterate to solutions. They highlight how recent advancements have allowed AI to operate autonomously, such as an agent that efficiently sourced a developer relations lead in just 31 minutes, navigating complexities and ambiguities without a pre-set script. The piece traces the evolution from basic reasoning models to more sophisticated long-horizon agents that can take actions over extended periods. These agents benefit from two main approaches: reinforcement learning, which helps them maintain focus, and agent harnesses, which are designed around their limitations. The authors note that the performance of these agents is improving rapidly, with predictions that they will soon handle tasks currently requiring human expertise, drastically reducing the time needed to complete complex assignments. The implications for founders are significant. Current AI tools are mostly conversational, but the next wave will be proactive, acting like colleagues. This shift will transform how work is done, as users transition from individual contributors to managers of multiple AI agents. The article emphasizes that businesses need to rethink their workflows, identifying tasks that require sustained attention and persistence—areas where these advanced agents can excel.