DeepMind has announced that it developed a novel system called AlphaCode that can compose computer programs at a competitive level. AlphaCode accomplished a predicted rank within the top 54% of candidates in programming competitions by fixing new problems that require a merge of critical thinking, logic, algorithms, coding, and natural language comprehension.
The team authorized their appearance by using competitions placed on Codeforces, a well-known platform which moderates regular competitions, bringing tens of thousands of competitors from around the world to those who want to test their coding skills. AlphaCode performed at the level of an average competitor, demonstrating that an AI code-generation system can be a real competitor in programming competitions.
DeepMind plans to extend this exploration and hopes that further studies will result in the development of instruments that can improve programming and bring scientists closer to the creation of a problem-solving AI.