In a recent podcast, OpenAI’s chief product officer Kevin Weil had an in-depth conversation with Avalon CEO Varun Mayya and YouTuber Tanmay Bhat. Kevin mentioned that AI's progress in coding is impressive, and AI is expected to surpass human programmers in competitive coding benchmarks by the end of this year. This prediction is based on significant advances in AI coding models in recent years, especially advanced models launched by companies such as OpenAI and Anthropic.
Valen quoted Anthropic's research data in his podcast, pointing out that by 2027, the automation rate of functional code is expected to reach 99%. He specifically mentioned Anthropic's Claude3.5Sonnet and OpenAI's GPT-4o, o1 models, which performed well in coding tasks. In addition, Anthropic has also launched Claude Code, a tool designed to automate the software development process. Developers can directly entrust coding tasks through the terminal, further improving development efficiency.
Anthropic CEO Dario Amodei is also optimistic, who believes AI is making very rapid progress in programming. He predicts that AI will be able to write 90% of the code in the next three to six months, while that percentage may be close to 100% over the next 12 months. This means that AI will gradually take on most of the work in software development, greatly changing the current situation of the programming industry.
Asked about the timeline for full automation of functional code, Kevin said 2027 could be a conservative estimate at the moment of progress. He believes that with the continuous improvement of AI's inference capabilities, coding automation may be implemented earlier. Kevin explains that writing quality code requires complex inference capabilities, and this capability is rapidly increasing as AI models continue to optimize.
Kevin also mentioned that there are currently about 40 million programmers worldwide, and OpenAI's o1 model ranking in competitive coding benchmarks has increased from the initial 1,000 globally to one million. He revealed that OpenAI's upcoming o3 model has ranked 175th in the world in terms of ranking, showing the continuous progress of AI in the field of programming. Kevin believes that with the continuous training and optimization of the model, AI's performance in competitive coding will further improve, and is expected to surpass human programmers within this year.
In addition to progress in coding, Kevin also shared his views on the rapid development of AI. He mentioned that GPT-4.5 has shown human-like characteristics and the way in which it works in the future will undergo tremendous changes. He also revealed the expected timetable for GPT-5 and discussed how AI revolutionizes the field of education and how it can change the industry landscape through democratizing software creation. These views of Kevin provide a new perspective on the future development of AI, demonstrating its potential in multiple fields.
Overall, AI is making exciting progress in programming. As companies such as OpenAI and Anthropic continue to launch advanced coding models, AI is expected to surpass human programmers in the near future and become the dominant force in software development. This trend will not only change the programming industry, but will also have a profound impact on the entire technology field.