Aschenbrenner AGI interview highlights the rapid progress in artificial intelligence, driven by exponential increases in compute cluster sizes. The interview suggests that Artificial General Intelligence (AGI) could be achieved by around 2027 and Artificial Super Intelligence (ASI) shortly thereafter. Aschenbrenner argues that AI labs will likely be nationalized due to the significant power of these tools, while also noting that tech startups are not yet secure enough to handle state-level spying on their models. Critics argue that his extrapolations of current trends into the future are overly optimistic, but Aschenbrenner's points about the continued trend of exponential improvement in model performance are supported by past successes such as GPT2 and GPT4. The interview also touches on concerns about energy and chip supply constraints, which may limit the growth of enormous compute clusters.