Latest news with #Dębiak
Yahoo
2 days ago
- Yahoo
Human programmer beats OpenAI's custom AI in 10-hour marathon, wins World Coding Championship — Polish programmer might be the last human winner
When you buy through links on our articles, Future and its syndication partners may earn a commission. Przemysław 'Psyho' Dębiak, a 42-year-old programmer from Gdynia, Poland, made history by defeating OpenAI's custom AI model at the AtCoder World Tour Finals (AWTF) 2025 "Humans vs AI" contest in Tokyo. Considered one of the most prestigious coding tournaments in the world, the AWTF invites just 12 of the top-ranked human programmers—and, for the first time, an AI competitor—to tackle its grueling challenges. After a 10-hour coding marathon, Dębiak edged out the AI by roughly 9.5%, clinching first place while the OpenAI-built model settled for second. 'Humanity has prevailed (for now)!' Dębiak wrote on X, confessing he'd slept only around 10 hours over three days while pushing himself to the limit. OpenAI CEO Sam Altman aptly responded with, 'Good job, Psyho.' The AI model, branded OpenAIAHC, was widely expected to dominate the market. Still, Dębiak's innovative, heuristic-driven approach—using problem-solving shortcuts and educated guesses instead of brute-force calculation—secured the win. Contest administrator Yoichi Iwata praised his distinctive method, noting that while the AI excelled at raw optimization, it 'fell short of human creativity.' The AtCoder World Tour Finals (AWTF) is regarded as the ultimate stage for heuristic programming contests that focus on 'good-enough' solutions to complex, unsolvable problems rather than perfect ones. This year's challenge tasked competitors with plotting a robot's path across a 30×30 grid using the fewest possible moves, an NP-hard optimization problem with countless possible outcomes. With no access to external libraries or documentation, success relied on intuition, creativity, and adaptability—qualities where human ingenuity can still outthink the raw speed and precision of AI models. Dębiak, a former OpenAI engineer who helped develop OpenAI Five (the Dota 2 AI), competed using only Visual Studio Code with basic autocomplete and admitted the AI pushed him to his limits: 'I was close to the model's score, and that pushed me to give everything.' Near the end of the 10-hour marathon, he overtook OpenAIAHC to claim victory and a 500,000 yen prize. A veteran algorithmic competitor, Mensa member, and four-time TopCoder Open Marathon champion, Dębiak has never held a full-time job and once even joked about considering careers from DJing to professional poker. His win carries symbolic weight, a human triumphing over AI in a field where machines typically excel. All that being said, AI's steady progress is undeniable—Stanford's 2025 AI Index found coding benchmarks jumped from 4.4% AI success in 2023 to 71.7% in 2024. With tools like GitHub Copilot used daily by over 90% of developers, AI is reshaping workflows. However, this win highlights that creativity, endurance, and intuition —especially in long-form heuristic challenges —remain human strengths. Dębiak isn't naive about the challenge ahead, saying, 'It's easy to imagine a different problem where AI would win and humans would be far behind.'. Still, this feels like a human John Henry moment—a testament not just to sheer human will, but to the spark of creativity machines haven't replicated and (hopefully) will never be able to. Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.


Hans India
3 days ago
- Sport
- Hans India
Human beats AI in Tokyo coding contest!
Tokyo: Atthe 2025 AtCoder World Tour Finals' Heuristic Challenge held in Tokyo, renowned coder Przemysław Dębiak — widely recognized in the programming world as 'Psyho' and a former member of the OpenAI team — edged out the specially designed AI model in a tightly contested battle. Hosted in Tokyo by the esteemed Japanese competitive programming site AtCoder, this year's grand finale introduced a unique twist — a 'Humans vs AI' showdown. Considered the pinnacle of invitation-only programming competitions, this elite tournament welcomes just 12 of the highest-ranked coders annually, chosen through rigorous qualification 10 hours, the event tasked participants with tackling one exceptionally difficult optimization problem—completely unaided, with no access to libraries, documentation, or outside help. Though the AI took an early lead, Dębiak ultimately overtook it. Although AI initially pulled ahead, Dębiak eventually surpassed it by drawing purely on his instincts, creativity, and deep expertise. Fellow coder Stanisław Eysmont commented, 'Przemek succeeded without pre-built tools, without reference materials, and without any guidance.' 'I was so tired. I actually felt at some point that I should take a break,' Dębiak, who competed under the name 'Psyho,' told the reporters. 'But at the same time, I was very close to getting a score comparable to the model,' he added. In a historic first, the contest opened its doors to an AI competitor, with OpenAI not only sponsoring the event but also entering its custom AHC model as an official participant. To ensure fairness between human participants and the AI system built by OpenAI, AtCoder supplied standardized hardware to all contenders. The rules of the tournament allowed the use of any programming language supported on the AtCoder platform, and there were no penalties for incorrect submissions. Dębiak was also instrumental during his tenure at OpenAI, serving as one of the pioneering engineers behind OpenAI Five—the AI system that famously triumphed over professional Dota 2 players in 2019. In a post shared on the social media platform X, OpenAI CEO Sam Altman acknowledged Dębiak's win and wrote, "good job Psyho." OpenAI also publicly recognised its tool's performance on X, stating, "Our model took 2nd place at the AtCoder Heuristics World Finals! Congrats to the champion for holding us off this time."


Hans India
4 days ago
- Science
- Hans India
Human Coder Triumphs Over OpenAI in 10-Hour AtCoder Showdown
In a landmark moment for human ingenuity, Polish programmer Przemysław Dębiak, better known in the coding world as Psyho, emerged victorious over an advanced OpenAI model in a grueling 10-hour programming challenge at the AtCoder World Tour Finals 2025. Despite the formidable pace and precision of AI, it was sheer human persistence that clinched the win — if only just. Shortly after the event, Dębiak took to social media platform X (formerly Twitter), posting: 'I'm completely exhausted. ... I'm barely alive. Humanity has prevailed (for now!).' Dębiak, who once worked for OpenAI himself, rose above exhaustion and intense competition to score a staggering 1,812,272,558,909 points, edging out the AI entrant 'OpenAIAHC,' which scored 1,654,675,725,406 — just 9.5% behind. The remaining ten human finalists, all top-tier programmers selected through a year-long ranking process, couldn't match the pace of either. Held in Tokyo by the renowned Japanese platform AtCoder, this year's finals featured an unprecedented 'Humans vs AI' special event. The programming community watched closely as twelve of the world's top coders took on a sophisticated AI designed by OpenAI — the same company behind ChatGPT. The stakes were high, not just in terms of points, but in what the outcome would symbolise in the ongoing conversation around human versus machine capabilities. The format of the event was intentionally designed to push limits. Spanning 600 minutes, the competition involved solving an extremely complex optimization puzzle — a hallmark of the Heuristic Contest category. These types of challenges are famously hard, often requiring approximate or creative solutions under time pressure. To ensure fairness, all participants — human and AI — operated on identical hardware. Submissions were spaced out by a mandatory five-minute cooldown. While the AI system outperformed the rest of the human participants, the fact that Dębiak bested the machine has quickly become a symbolic victory — a testament to human resilience in the age of automation. Even OpenAI acknowledged the significance. In a public statement, the company wrote: 'Our model took 2nd place at the AtCoder Heuristics World Finals! Congrats to the champion for holding us off this time.' OpenAI CEO Sam Altman added his voice with a brief but pointed message: 'Good job psyho.' Dębiak's win is not only a personal milestone but a reminder that human creativity and grit remain highly relevant — especially in arenas dominated by logic, speed, and algorithmic prowess. The result has sparked fresh reflection on how AI and humans will coexist in technical professions going forward. As AI tools continue to advance and close the gap, many wonder how long humans can hold their lead. Yet for now, the scoreboard — and perhaps something more — remains in favour of humanity.

Business Insider
5 days ago
- Business
- Business Insider
A programmer beat an OpenAI tool at coding. He says the AI made him push himself to win.
AI hasn't yet killed the coding star. On Wednesday, a programmer from Poland, Przemysław Dębiak, beat an OpenAI tool at the AtCoder World Tour Finals 2025 Heuristic Contest in Tokyo, which is sponsored by the company. "I was so tired. I actually felt at some point that I should take a break," Dębiak, who competed under the name "Psyho," told Business Insider over a video call on Friday. "But at the same time, I was very close to getting a score comparable to the model," he added. That pushed him to dig deep, using "all the remaining energy" to beat the AI. "I'm trying to give 100% of what I have and try to stay alive," he said, adding that without OpenAI's model, his score "would be much, much lower." Dębiak's victory drew the attention of OpenAI CEO Sam Altman. In a post on X on Wednesday, Altman wrote, "good job psyho." When reached for comment, OpenAI directed BI to a post on X, where it had written, "Our model took 2nd place at the AtCoder Heuristics World Finals! Congrats to the champion for holding us off this time." The contest is held annually and run by AtCoder, a Japan-based competitive programming site. Yoichi Iwata, the competition's administrator, told BI that OpenAI's model outperformed those who used similar approaches, but Dębiak "arrived at a completely different solution." "We expected a human to win, and were rather surprised that the AI was able to secure second place," Iwata added. "While the AI surpassed humans in terms of optimization ability, we believe it still fell short of human creativity." The competition's results page shows Dębiak prevailing over competitors from Japan, Georgia, and France. On Friday, Dębiak shared a results page on X that showed him as the competition's victor and OpenAI in second place. "The results are official now and my lead over AI increased from 5.5% to 9.5%," he wrote. Update: I'm alive and well The results are official now and my lead over AI increased from 5.5% to 9.5%😎 Honestly, the hype feels kind of bizarre. Never expected so many people would be interested in programming contests. Guess this means I should drop in here more often👀 — Psyho (@FakePsyho) July 17, 2025 After winning the competition on Wednesday, Dębiak celebrated his win online, writing, "Humanity has prevailed (for now!)", adding that he was "completely exhausted" after just 10 hours of sleep in three days. He told BI he flew from Warsaw to Tokyo for the competition. The programmer wrote on X that he did not use AI tools, just "regular" Visual Studio Code, a software development platform, with "some basic autocomplete to speed up repetitive stuff." Dębiak, 41, told BI he only found out about OpenAI's entry a week before the contest. "No one knew that this would be, to some extent, an exhibition match between humans and AI," he said. The stars aligned for the human AI will win in cases that require straightforward engineering — implementing algorithms, solving the problem, and optimizing code — because it's simply faster, Dębiak told BI. But in longer contests where programmers "do everything from scratch," he said, AI has a harder time keeping up. "The longer the contest is, the more chances for humans and less chances for AI," he said, adding that he overtook OpenAI's tool near the end of the 10-hour competition. Dębiak said he competed in "algorithmic optimization," where the goal was to write the most efficient solution to a complex problem. In this competition, he had to code a program that guided robots across a 30x30 grid using the fewest moves possible. Reflecting on the competition, Dębiak said the variables lined up in his favor. "It's easy to imagine a different problem where AI would win and all the humans would be like far, far away," he added. AI has famously beaten humans in other high-profile competitions. In 1997, IBM's Deep Blue AI machine beat chess grandmaster Gary Kasparov, and in 2016, Google DeepMind's AlphaGo beat Go world champion Lee Sedol. In February, Altman said that by the end of the year, OpenAI could surpass humans in coding competitions. AI is already writing big chunks of code at Microsoft, Google, and Meta, the companies' CEOs have all said in recent months. Altman has said that demand for software engineers could eventually dip. "My basic assumption is that each software engineer will just do much, much more for a while. And then at some point, yeah, maybe we do need less software engineers," he said in March, referring to OpenAI's hiring strategy.