Google SWE points 48 AI agents at NASA Lidar, finds unmapped Maya pyramid in 72h
1024 points
• 1 day ago
• Article
Link
这次发现源于作者在前往秘鲁旅行途中，关于历史探索的一个简单问题。受到 Hiram Bingham 寻找失落城市的启发，作者思考现代技术能如何应用于考古发现。通过与包括 Gemini 、 ChatGPT 、 Claude 和 Grok 在内的多种 AI agents 合作，作者确定 lidar 是最合适的工具。由于卫星影像常被茂密的丛林树冠遮挡，lidar 利用激光脉冲绘制植被下方的地表，从而让人能够看到肉眼难以察觉的隐藏信息。
团队随后将研究重点从秘鲁转向玛雅地区，因为那里更容易获取公开的 lidar 数据。仅用了一个晚上，48 个 AI agents 就分析了约 1,200 平方公里的丛林影像。它们像高度勤勉、条理清晰的研究人员一样逐一手动查看图像，识别金字塔、广场和平台等人造结构，同时剔除自然地貌。到第二天早上，agents 已经标记出若干看起来像是未被发现的地点：这些地点呈现出与周围环境截然不同的锐利人工棱角与清晰的建筑布局。
其中有一处地点尤为引人注目：它包括一个已被整理得较为平整的广场、一座高约 8 到 12 米的金字塔、一个抬高的庭院，以及一个可能的蓄水设施。为确认该位置是否早已登记，agents 对既有文献和考古档案进行了细致核查。确认这一异常并非既知记录后，AI agents 便向作者提供了联系 Mexico's National Institute of Anthropology and History（INAH）的必要步骤。根据他们的指导，作者撰写了正式报告，并将其翻译成西班牙语以确保表达清楚。
提交结果后，墨西哥相关部门的回复在 72 小时内到达。 INAH 证实，在所提供的坐标处没有已登记的考古遗址记录。尽管这还未达到官方实地发现的标准，但现有数据表明该地点很可能是经典时期的玛雅政务中心，或许是用于举行仪式、进行地方治理以及日常生活的枢纽。作者表示，只要 INAH 决定派遣团队到现场核实，他愿意提供协助。
回顾整个过程，作者认为自己在考古领域的"外行"身份反而带来了优势：它让自己能够用一种全新、没有包袱的方式去面对一个长期存在的问题。通过把现成的计算资源和 AI agents 与公开数据结合起来，人们就能在极短的时间内完成重要的科学工作。这一成果凸显了 AI 作为各学科催化力量的潜力，也意味着，阻碍发现的主要剩余因素只有人类的好奇心与想象力。
The discovery began with a simple question about historical exploration while traveling through Peru. Inspired by Hiram Bingham's search for lost cities, the author wondered how modern technology could be applied to archaeological discovery. Collaborating with various AI agents, including Gemini, ChatGPT, Claude, and Grok, the author identified lidar as the ideal tool. Because satellite imagery is often obstructed by dense jungle canopy, lidar, which uses laser pulses to map the ground beneath the vegetation, offered a way to see what remained hidden to the naked eye.
The team shifted their focus from Peru to the Maya region, where public lidar data is more accessible. Overnight, forty-eight AI agents analyzed approximately 1,200 square kilometers of jungle imagery. They acted as highly diligent, methodical researchers, manually inspecting images to identify man-made structures like pyramids, plazas, and platforms while filtering out natural features. By the next morning, the agents had flagged several sites that appeared to be undiscovered, displaying sharp, artificial corners and structural layouts that clearly contrasted with the surrounding landscape.
One particular site caught their attention, featuring a leveled plaza, an 8 to 12-meter-tall pyramid, a raised courtyard, and a potential water reservoir. The agents performed a thorough review of existing literature and archaeological archives to confirm that the location was not already registered. Once confident in the anomaly, the AI agents provided the author with the necessary instructions to contact Mexico's National Institute of Anthropology and History (INAH). Following their guidance, the author prepared a formal report, which was translated into Spanish to ensure clarity.
After submitting the findings, the response from the Mexican authorities arrived within 72 hours. INAH confirmed that there was no record of an archaeological site at the provided coordinates. While not yet an official field discovery, the data suggests the site was a Maya civic center from the Classic period, likely serving as a hub for ceremonies, local governance, and daily life. The author has offered to support a potential field investigation should INAH choose to send a team to verify the findings on the ground.
Reflecting on the process, the author notes that being an outsider in the field of archaeology proved to be an advantage, as it allowed for a fresh, unburdened approach to a long-standing problem. By combining readily available compute resources and AI agents with existing public data, significant scientific work can be accomplished in a remarkably short time. This achievement highlights the potential for AI to act as a catalyst across all scientific disciplines, suggesting that the only remaining limits to discovery are human curiosity and imagination.
385 comments • Comments Link
- 对大规模 LiDAR 数据集的自动化分析，标志着考古发现潜力的一次重大飞跃：它把原本依赖人工、费时费力的检查，转向了快速、机器辅助的模式识别。
- 大规模 agentic workflows 往往会轻易宣称有新发现，但实际上这些所谓新发现并不新鲜；它们常常也无法拿现有（有时还相当晦涩的）学术文献全集来核验自己的结论。
- 真正的科学发现离不开深厚的领域专业知识来解读研究结果。仅凭表面层面的结构识别是不够的，还需要把结果与人口密度、社会关系、农田分布等方面的人类学理解做对照。
- 未经核实、快速发布的疑似遗址公告，带来了助长"vibe archaeology"的风险：在这种情境下，AI 生成的报告即便没有经过严格的实地考察或同行评审，也会被当作权威。
- "发现工具"的民主化带来了一种两难局面：它确实加速了研究，但也可能把盗掘与恶意破坏行为推到工业化规模。如今，相关人员可以更轻松地锁定高价值目标。
- 在技术业余项目中使用"Google SWE"的品牌塑造引发了不必要的怀疑与争论，尽管这也凸显出企业支持的研究能力与个人学术研究者有限接入之间的差距越来越大。
- 科学机构在公众眼中常常是知识的把关者。但它们在核验主张时的谨慎，能够遏制错误信息的扩散，也能避免因鲁莽探索而造成考古背景不可挽回的损失。
- 用 48 个 agents 完成一项任务（而不是单一、优化后的处理流水线）之所以有效，说明人们正在转向更依赖"蛮力破解"、高计算量的解决思路。这种做法或许会成为公民科学家的一种标准方法论（尽管成本不低）。
- 核心争论仍在于：AI agents 到底是真正的"intelligent"，还是只是高效的检索与模式匹配工具——尤其当结果高度依赖人类精心策划的数据集时。
- 现代考古学的进步正在被一种面向数据的"everything, everywhere"式路径重新定义。其主要障碍不再是发现异常的能力，而在于人类能否在识别出异常之后对其进行保护，并做出妥善、可信的记录。
这场讨论折射出一个更广泛的张力：AI agents 在利基研究领域正被迅速、以高热度方式采纳；而要维持科学完整性，则必须经历缓慢得多、谨慎得多的验证过程。虽然大范围扫描并从中找出模式，确实是一项具有变革性的能力，但它也带来巨大风险，包括大规模盗掘的可能性，以及缺乏坚实基础的"vibe-coded"发现被迅速传播。当前已形成较为明确的共识：机器可以作为平凡识别任务的倍增器发挥作用，但它们在背景智慧上仍不足以替代人类专业知识——尤其是在敏感遗址的道德保护方面。最终，这一事件也成为计算领域"Bitter Lesson"的一个案例：尽管规模化、以"蛮力"驱动的 agentic workflows 能产出令人瞩目的结果，但它们在透明度与怀疑态度方面所要求的标准，往往高于技术文化实践通常能提供的水平。 • Automated analysis of large-scale LiDAR datasets represents a significant leap in potential for archaeological discovery, moving the task from manual, time-intensive examination to rapid, machine-assisted pattern recognition.
• Large-scale agentic workflows are prone to claiming novelty where none exists, often failing to verify findings against the full scope of existing, sometimes obscure, academic literature.
• True scientific discovery requires deep domain expertise to interpret findings, as evidenced by the contrast between a superficial identification of structures and an anthropological understanding of population density, social relationships, and farm distribution.
• The rapid, unverified announcement of potential sites poses a risk of encouraging "vibe archaeology," where AI-generated reports are treated as authoritative without rigorous ground-truthing or peer review.
• Democratizing discovery tools creates a dilemma: while it accelerates research, it simultaneously increases the threat of industrial-scale looting and site destruction by bad actors who can now identify high-value targets with minimal effort.
• The "Google SWE" branding in technical hobby projects creates unnecessary skepticism and debate, though it highlights the growing divide between corporate-backed research capabilities and the limited access of individual academic researchers.
• Scientific institutions are often viewed by the public as gatekeeping knowledge, yet their cautious approach to verifying claims protects against the spread of misinformation and the irreparable loss of archaeological context caused by reckless exploration.
• The effectiveness of using 48 agents for a task—rather than a single, optimized processing pipeline—suggests a shift toward brute-force, high-compute problem solving that may become a standard, albeit costly, methodology for citizen scientists.
• Determining whether AI agents are genuinely "intelligent" or simply efficient search-and-pattern-matching tools remains a central point of contention, particularly when outcomes rely heavily on human-curated datasets.
• Modern archaeological progress is being redefined by an "everything, everywhere" approach to data, where the primary barrier is no longer the ability to find anomalies, but rather the human capacity to protect and properly document them once identified.
The discussion reflects a broader tension between the rapid, high-hype adoption of AI agents in niche research fields and the slow, deliberate verification processes essential to scientific integrity. While the ability to scan vast geographical areas for patterns is clearly a transformative capability for archaeology, it carries significant risks, including the potential for mass looting and the dissemination of "vibe-coded" findings that lack rigorous foundation. A clear consensus emerges that while machines can effectively act as a force multiplier for mundane identification tasks, they currently lack the contextual wisdom to replace human expertise, particularly regarding the ethical protection of sensitive sites. Ultimately, the episode serves as a case study for the "Bitter Lesson" in computing, illustrating that while scale and brute-force agentic workflows can yield impressive results, they also demand a higher standard of transparency and skepticism than current tech-culture practices typically provide.