In recent years, the geographic information industry has been experiencing a "cognitive revolution" driven by AI.
In early 2025, OpenAI co-founder Andrej Karpathy proposed the concept of "Vibe Coding": "Describing requirements in natural language and AI completing the entire process of coding".
This concept quickly resonated with the global tech community, and several tech giants in Silicon Valley have set up related tracks. According to a report by First Financial, overseas AI programming tool cursor has been valued at nearly billions of dollars in just three years thanks to its ambient programming model; There are at least 7 AI programming unicorns with valuations exceeding 1 billion US dollars, and their total financing has exceeded 24 billion RMB.
In China, "atmosphere programming" has also sparked heated discussions on major social media platforms. This mode is called "the artifact of liberal arts students" by many netizens because it can significantly reduce the programming threshold.
Data shows that when Baidu's codeless programming product, Dida, was launched for 8 months, it generated over 500000 commercial applications, with a daily increase of over 150% in new applications. Among them, 81% of users were non programmers.
When the concept of "atmosphere programming" enters the field of geographic information, this transformation not only impacts traditional development models, but also triggers deep thinking about talent structure in the industry - the weight of technical ability to write code is decreasing, and humanities students who are better at understanding and refining needs are rewriting the technological discourse power of the geographic information industry?
1. Implementation of "Atmosphere Programming": From "Code is King" to "Requirement Definition"
The essence of atmosphere programming is to shift the core of software development from "technical implementation" to "requirement analysis". The core logic is that users describe their goals through natural language, AI automatically completes code generation, debugging, and optimization, and the developer role shifts from a code creator to a requirement provider and result checker.
Google CEO Sandal Pichai said frankly in Google Podcast that just like the Internet promotes the professionalization of content creation, atmosphere programming can enable more non-technical people to try new careers, and can easily build applications and websites without programming foundation.
In January 2026, a local information technology practitioner shared a practical case in the technical community: "Just input 'develop a QGIS based plugin that can automatically recognize building contours in satellite images and generate vector layers', and AI can directly generate executable project products. ”
This scenario is not an isolated case, but a microcosm of the technological transformation in the geospatial industry - AI is "de professionalizing" the code logic.
In the traditional development model, programmers need to break down complex steps such as cleaning, modeling, and spatial analysis of geographic spatial data into precise code logic; Nowadays, AI empowerment tools such as ArcGIS Pro and SuperMap iPortal can directly parse natural language instructions, automatically generate professional code for map service interface calls, spatial overlay analysis, and more.
The CSDN blog once published an article disclosing a set of data cases: a city imported the latest satellite images through ArcGIS Pro during its annual map update, and used a deep learning based object detection model to complete the annotation work of buildings, roads, vegetation and other elements that used to require manual teams to spend several months in just one week. The classification accuracy improved from 82% of traditional methods to 95%. Users can complete professional terrain recognition tasks without writing complex code.
2. Opportunities for Humanities Students: Defining Needs and Abilities as Core Competitiveness
The core reason why "atmosphere programming" is considered to potentially reconstruct the talent structure of the industry is that it has turned "demand generation and precise description" into the core of the development process, which is the natural advantage of humanities students.
When talking about the substitutability of AI for jobs, Liang Jianzhang, co-founder of Ctrip and a population economist, pointed out that artificial intelligence has a significant impact on primary information related industries in the short term. Previously, computer science was a popular major, and the programmer group formed a large employment group. Among them, some programmers engaged in simple mental labor have been significantly impacted by artificial intelligence in the short term.
The uniqueness of the real estate industry highlights the value of "atmosphere programming".
As a computer system specialized in processing spatial data, GIS has application scenarios covering multiple fields such as resource investigation, environmental monitoring, urban planning, etc. The demand side is mostly composite talents with professional backgrounds in geography, planning, public management, etc., including a large number of humanities students.
In the previous development model, requirements had to go through multiple stages of "professional description product manager conversion programmer development", which easily led to information loss; And 'atmosphere programming' allows the demand side to directly interface with AI, greatly compressing the communication chain and error space.
Li Weisen, President of the China Geographic Information Industry Association, once predicted that "artificial intelligence technology is driving the transformation of spatial intelligence software towards digitization, refinement, and intelligence, which means that the industry's demand for composite talents who understand business and can express themselves will significantly increase
In the real estate industry, demand highly relies on policy interpretation and spatial semantic understanding. Humanities students, with their insight into industry logic, are becoming a key "bridge" for the implementation of AI programming, and their value is mainly reflected in three dimensions:
Firstly, translating policy semantics into technical rules. Geographic information applications need to strictly comply with policy requirements such as national land planning and emergency management, and humanities students have a natural advantage in policy interpretation.
Secondly, spatial narrative drives functional innovation. The user experience of geospatial products highly relies on spatial scene design, and the scene construction and narrative ability of humanities students play a key role in this.
Thirdly, cross-border collaboration triggers an efficiency revolution. Liberal arts students transform business knowledge such as satellite image interpretation standards and illegal spot detection rules into structured instructions, while AI focuses on completing complex calculations such as data preprocessing and model training. This division of labor model not only significantly reduces project labor costs, but also reduces cross role communication losses.
3. Substitution or complementarity? Cross disciplinary talents reshape the local information ecosystem
Although the argument of "liberal arts students replacing programmers" has emerged, the industry consensus still believes that AI programming reconstructs the division of labor among talents rather than simply replacing human labor. Liberal arts students and programmers are not in opposition, but form a new pattern of complementary coexistence.
Firstly, the technical barriers for programmers in complex scenarios remain insurmountable. Andrej Karpathy once admitted in a speech shared on the CSDN blog that ambient programming is more suitable for rapid development of basic applications or prototype products, while serious and complex large-scale production applications still require professional teams to collaborate and tackle.
Wu Jun, former Vice President of Tencent and an expert in artificial intelligence, also holds a similar view. He pointed out that if a computer wants to meet specific programming needs, the user must have programming skills themselves. Otherwise, once the program written by AI has vulnerabilities, it is often difficult for humans to troubleshoot and fix them. So you must understand very well that you cannot completely let the calculator fix it, because if it feels wrong, it will not write such code
In the field of geology and information technology, core projects such as urban level flood simulation and global land spatial planning systems require complex technologies such as hydrological models, spatiotemporal big data analysis, and multi system cross platform integration. Currently, AI can only complete basic code generation, and key links such as core logic design, performance optimization, and multi-source heterogeneous data fusion still need to be led by senior programmers. Practical data shows that for complex geospatial systems involving spatiotemporal big data analysis, manual coding still accounts for over 70% of the workload.
More importantly, there are potential risks associated with AI generated code. According to statistics, enterprise level geospatial systems require "test driven governance" to ensure reliability, and tasks such as test case design and logic vulnerability investigation heavily rely on the engineering experience of programmers.
Secondly, humanities students also face significant capacity bottlenecks. The cognitive shortcomings of humanities students in GIS basic technology concepts such as spatial benchmarks and topological relationships. In addition, mainstream geospatial development still relies on the deep application of professional tools such as ArcGIS Pro and SuperMap. Liberal arts students need to learn additional basic technical knowledge such as plugin calling and API interface adaptation in order to collaborate more efficiently with AI.
Fourth, Future Vision: "New Geospatial Ecology" under Human Machine Collaboration
The ultimate goal of the transformation of the real estate industry is not to replace programmers with humanities students, but to achieve an efficient closed-loop of "demand code" through "atmosphere programming", releasing technological productivity and knowledge value. This transformation is driving the formation of a new ecosystem in the industry:
Job restructuring has given rise to new professions. AI trainers and spatial product managers have become core new positions: the former is responsible for transforming industry knowledge such as farmland protection and ecological red lines into AI understandable rule libraries and semantic labels; The latter needs to coordinate demand analysis and functional design, with the ability to interpret policies, conduct spatial analysis, and master AI tools, becoming the core hub connecting business and technology.
Atmosphere programming "is not an exclusive carnival for humanities students, nor is it the end of the programmer's career. It is a key catalyst for promoting the transformation of the local information industry from" technology intensive "to" knowledge intensive ". When AI undertakes repetitive and procedural coding tasks, the core competitiveness of practitioners will no longer be limited to a single skill, but will shift towards industry insight, cross domain collaboration ability, and AI tool mastery ability.
AI will not replace programmers, but programmers who use AI will replace those who do not use AI. 360 Group founder Zhou Hongyi emphasized in the 2026 AI panoramic forecast that "silicon-based employees" will be officially included in the enterprise employment system, forming a mixed work team of "carbon based+silicon-based", and humans need to become planners and supervisors of AI. After AI enters the second half, the focus shifts from big models to intelligent agents. Intelligent agents will drive organizational restructuring, and "super individuals" skilled in managing intelligent agents will become the core of the workplace.
In the wave of transformation in the local information industry, the boundary between technology and manpower is being redefined. Liberal arts students need to strengthen their ability to adapt technical tools, while programmers need to enhance their understanding of the industry and AI collaboration skills. Only practitioners who actively embrace change and achieve capability upgrades can firmly establish themselves in the new ecosystem of human-machine collaboration and jointly paint the future picture of the land and information industry.
(Note: Images and some text content are generated by AI)
