Part 2/9:
Since the leaks from OpenAI, it has become evident that their AI models, specifically the 01 and 03 versions, are achieving unprecedented performance across a variety of benchmarks. Some experts even assert that these developments mark our entry into an early form of AGI.
Key figures like Ilia Suk and Leopold Ashenbrener, who have transitioned from OpenAI to launch their ventures focusing on "safe superintelligence," have echoed sentiments that we are not merely dreaming of advanced AI but are on the cusp of realizing it. The encapsulation of this revolution can be summarized with the concept of "test and compute," which involves investing additional resources in AI thought processes to yield staggering results.