While it may appear that AI models are swiftly surpassing human capabilities, there remains a glimmer of hope for humanity. This hope is represented by Przemysław Dębiak, a former OpenAI employee who recently emerged victoriously against one of the company’s models in a coding championship.

Dębiak participated in the AtCoder World Tour Finals 2025 Heuristic contest in Tokyo last week, competing not only against numerous human contestants but also against OpenAI’s custom model named ‘OpenAIAHC’. This coding challenge required participants to solve a complex optimization problem within a 10-hour timeline, followed by a scoring system based on their performance.

“Humanity has prevailed (for now!) I’m completely exhausted. I figured I had 10 hours of sleep in the last 3 days and I’m barely alive.” — Przemysław Dębiak, on Twitter

Despite being allowed to use any programming language at AtCoder, participants were restricted to identical hardware and faced a mandatory five-minute wait between submissions.

In this competition, Dębiak, competing under the name Psyho, achieved a remarkable final score of 1,812,272,588,909, placing him at the top of the leaderboard, well ahead of the AI’s score of 1,654,675,725,406, which took second place. This event marks a noteworthy point in history as it is believed to be the first instance of an AI model directly competing against human programmers in a coding contest.

While it is evident that coders continue to enhance their skills, the rapid growth of AI technology suggests that it may soon dominate leaderboards in similar competitions. For now, however, the victory belongs to Dębiak, reminding us that humanity can still hold its ground against emerging technologies.