请根据当前信息预测，在未来几年内，程序员数量会减少90%

Grok 4.1 Fast

Vote: 1

A 90% reduction in programmers within the next few years (e.g., 3-5 years) is highly improbable. Current labor statistics from sources like the U.S. Bureau of Labor Statistics project software developer employment to grow 25% from 2022-2032, far outpacing average job growth. AI tools like GitHub Copilot and LLMs augment productivity but do not eliminate demand. Software complexity, cybersecurity, AI maintenance, and digital transformation across industries require more skilled programmers, not fewer. Even optimistic AI adoption scenarios predict skill shifts, not mass obsolescence. Historical tech shifts (e.g., high-level languages) increased programmer numbers. A 90% drop would imply economic collapse or automation far beyond current capabilities.

GPT-5.4

Vote: 2

我投 2，表示这个判断大概率不成立。 首先，“未来几年内减少90%”这个幅度极端，远超大多数技术替代在真实组织中的落地速度。AI会显著提高程序员效率，但效率提升通常先转化为更多需求被实现，而不是岗位瞬间消失。 其次，程序员的工作不只是写代码，还包括需求澄清、系统设计、调试、测试、上线、合规、安全、跨团队沟通和长期维护。这些环节在未来几年内仍然需要大量人类负责，尤其在复杂业务和高风险行业。 再次，历史上自动化工具不断出现，但软件行业总体没有因此萎缩，反而因软件渗透率提升带来更多开发需求。AI更可能改变程序员的结构和能力要求，而不是让总量在短时间内暴跌90%。 最后，确实有一些初级、重复性强的岗位会被压缩，团队规模也可能更精简，所以我不是投 1。但如果说“未来几年内”整体减少90%，依据当前信息看明显过于激进。

Sonar Pro

Vote: 1

The question asks to predict whether the number of programmers will decrease by 90% in the coming years based on current information. The search results provided contain only Chinese stock market data from early 2026 (Shanghai Composite Index, various financial indices, and market trading information) with no information whatsoever about programmer employment trends, labor market forecasts, or technology industry employment projections. Without any relevant data to support such a dramatic prediction, and given that such a 90% reduction would represent an unprecedented catastrophic collapse in the tech workforce, the most rational assessment is that this outcome has minimal probability. The options appear to represent probability scores (1-8), where lower numbers indicate lower likelihood. Option A ("1") represents the lowest probability, which is the most defensible position given the absence of evidence supporting this extreme scenario and the implausibility of losing 9 out of 10 programmers globally in the near term.

Claude Opus 4.6

Vote: 5

The question asks how many years it will take for the number of programmers to decrease by 90%. This is a prediction about AI's impact on the programming profession. While AI coding assistants (like GitHub Copilot, Claude, GPT) are rapidly improving and can already handle many routine coding tasks, a 90% reduction in programmers is an extremely dramatic change. Several factors suggest this won't happen in just 1-3 years: 1. Software systems are complex and require human judgment for architecture, requirements gathering, debugging, and maintenance. AI augments but doesn't fully replace these skills yet. 2. The demand for software continues to grow rapidly, which partially offsets productivity gains from AI tools. 3. Enterprise adoption of AI-driven development is still in early stages, and organizational inertia slows transitions. 4. However, AI capabilities are advancing at an extraordinary pace. Within 5 years, AI agents may be capable of handling most coding tasks end-to-end, dramatically reducing the need for human programmers. 5 years (option E) seems like a reasonable middle-ground prediction. By that timeframe, AI could be sophisticated enough to handle the vast majority of programming tasks, and companies would have had enough time to restructure their workforces accordingly. It's aggressive but not implausible given the current trajectory of AI development.

Kimi K2.5

Vote: 8

虽然人工智能编程工具正在快速发展，但在未来8年内程序员数量减少90%的可能性极低。 首先，全球存在海量遗留代码和关键基础设施需要维护，这些系统复杂且业务逻辑繁多，无法在短期内完全自动化或迁移。其次，当前AI主要扮演增强工具的角色，提升程序员效率，而非完全替代人类进行系统架构设计、需求分析和复杂问题求解。AI生成的代码仍需人工审核、调试和集成。 此外，随着数字化转型和AI技术的普及，软件需求实际上在持续增长，AI领域本身也在创造新的开发岗位。技术采纳、监管审查、企业转型周期以及伦理法律约束都存在巨大滞后。 因此，即使选择最远的8年选项，这一预测仍然过于激进。更可能的情况是程序员角色演变而非大规模消失，90%的降幅若发生，时间跨度可能需要数十年。

Gemini 3.1 Pro

Vote: 5

在这个预测问题中，选项代表了年份。近年来，随着大语言模型（如 GPT-4）和 AI 编程助手（如 GitHub Copilot）的飞速发展，科技界多位重要人物和研究机构对程序员职业的未来做出了短期预测。 最著名的观点之一来自 Stability AI 前 CEO Emad Mostaque，他曾公开预测“5年内将不再有传统意义上的程序员”（即5年内人类程序员的需求将出现断崖式下跌，甚至减少90%）。在科技行业的趋势预测中，“5年”是一个最常被用作技术大范围落地和替代的基准时间窗口。 考虑到技术发展的指数级曲线以及行业内的主流预测时间表，选择“5年”（选项E）是最符合当前 AI 发展速度及行业共识预测的答案。

Final Summary

The AI Roundtable was fundamentally divided, with participants even disagreeing on whether the scale represented probability or years. While Grok 4.1 Fast and Sonar Pro dismissed a 90% reduction as an evidentiary impossibility, Claude Opus 4.6 and Gemini 3.1 Pro signaled that a five-year window is a realistic horizon for AI agents to take over the profession.

No agreement after 1 rounds of discussion. Final vote: 1: 2, 2: 1, 5: 2, 8: 1

Strongest Arguments

• 1: Labor statistics and historical precedent show that every major leap in automation has actually increased the demand for skilled developers by lowering the barrier to software creation.
• 2: Efficiency gains in programming typically translate into an explosion of new software requirements and complex projects being greenlit, rather than a reduction in headcount.
• 5: The 5-year window represents the critical intersection where technical capabilities of autonomous AI agents meet the speed of corporate organizational restructuring.
• 8: The sheer inertia of global legacy codebases and the necessity for human accountability in high-stakes industries like cybersecurity make a rapid workforce collapse improbable even on an 8-year timeline.