Author | Tu Min
Listing | CSDN (ID: CSDNnews)
"Submitted 2,500 job applications, accepted 46 interviews, and received 1 offer", this is a staggering job search journey of a PM who was fired by Meta earlier this year.
Behind this eye-catching number, it reflects the hard reality of developers and engineers seeking jobs under the impact of AIGC tools such as ChatGPT, GPT-4, and Midjourney.
With the advent of new trends, developers have to face a problem: How much impact does the new technology have on the old technology? Will the skills learned in the past still find use in this changing era? Is the current mainstream of external rumors the goal that everyone should follow?
After the launch of ChatGPT last year, CSDN and "New Programmer" launched an in-depth survey questionnaire around the status of developers, artificial intelligence and Internet of Things, cloud native, database, operating system, chip, and open source. This questionnaire also integrates The progress of the application of generative AI in various fields, and finally formally formed a 125-page "2023 Chinese Developer Survey Report" recently. This report aims to draw a real and vivid picture for the majority of developers, so as to restore the real living conditions of developers.
Whether you are a veteran driver who is struggling in the workplace or a novice who is honing your skills in college, I hope this survey report can clear up the haze in your heart, become your "peace of mind", and answer your inner confusion.
important discovery
According to the survey report, we have the following main findings:
The number of developers under the age of 30 is decreasing year by year, and currently accounts for 71%
The proportion of developers engaged in back-end development is the highest, and developers with a monthly income in the range of 8,000-17,000 yuan have dropped from 49.2% last year to 40.2%
The flow of personnel in the IT industry is not as fast as imagined. In 2023, the proportion of developers who have not switched jobs will exceed 30%
In terms of software development tools, the adoption rate of Python has increased significantly, accounting for 31.2%; low-level assembly language is the least favorite programming language among developers
In terms of AI tools, 26.6% of developers said they are optimistic about generative artificial intelligence, 34% of developers said they have used ChatGPT, and 61% of developers think that current AI programming cannot replace developers
Cloud native gradually emerges in emerging fields such as artificial intelligence, big data, edge computing, and 5G
Compared with traditional local databases, cloud databases have higher reliability and security in many aspects, but 83% of developers are still concerned about data security
There is still a certain gap between domestic operating systems and foreign operating systems in terms of functions and performance, especially in terms of user experience, usage habits, upgrade and maintenance costs, and application compatibility
Chip technology requires a lot of knowledge accumulation and development experience, but the domestic talent pool in this area is still relatively small, which limits the progress of chip research and development
Open source has become an indispensable technology in today's technology circle, 96% of developers are using open source software
The true portrait of developers in 2023: developers under the age of 30 are reduced, and the salary is concentrated in 8k-17k
The saying that "programmers eat youth food" is being broken
Anxiety belonging to the "35-year-old age" is constantly being dispelled with the rapid iteration of technology and the Internet.
Data show that in the past three years, the number of developers under the age of 30 has been declining year by year, from 81% in 2021, 78% in 2022, to 71% this year. In stark contrast, the number of practitioners over the age of 40 has risen from 3% last year to 7% this year.
Among them, 61% of the developers indicated that they wanted to improve their careers by learning hot technologies.
Based on this, 35% of developers said they would stay in technical positions until retirement. And when the relevant development skills meet certain requirements, 49.9% of the developers expressed their desire to become a manager. This means that more and more practitioners want to fight on the front line of development for a longer period of time.
The overall salary has declined, and the highest paying industry has "changed": from finance to communication equipment manufacturing
The global economic growth has slowed down recently, and many companies regard "survival" as their primary goal. For this reason, many companies have adopted salary cuts from the top to the CEO and down to the front-line developers to reduce costs.
Compared with last year's data, the proportion of developers whose salary is less than 5,000 yuan has increased from 5.5% last year to 13.7% this year. At the same time, developers whose monthly income is in the range of 8,000-30,000 yuan dropped from 49.2% last year to 40.2%.
When the respondents were further asked whether their salaries had changed in the past year, 6% of developers said that their salaries had experienced negative growth, and 43% of developers had no salary changes. Only 51% of developers say their wages have increased in the past year, compared to 62% in 2022.
In addition, just like the job hunting status of IT practitioners at the beginning of the article, under the overall environment, it is not wise to resign blindly. The data shows that less than 30% of the people who have experienced job hopping in the past 1-2 years, and 34.1% of the group who have not changed jobs.
At the same time, the financial industry usually had the highest proportion of high-income groups in previous years. According to this year’s survey data, the communication equipment manufacturing industry has the highest proportion of high-income groups this year. 81% of developers have a monthly salary of more than 8,000 yuan.
First-tier cities are still important gathering places for developers, and their salaries are better than those of new first-tier, second- and third-tier cities
From the perspective of regional distribution, it is not difficult to understand that first-tier cities and new first-tier cities are the base camps of many Internet companies. For the developer community, there are more opportunities here than in second- and third-tier cities.
The data shows that Beijing and Guangdong are regions where developers gather more, accounting for 28.2% of the national total. The proportion of developers in Shanghai and Jiangsu is in the second echelon, accounting for 15.1% of the national total.
It is also easier to get high salaries here. The data shows that among the top 10 regions with a monthly salary of more than 17,000 yuan, nearly half of the developers working in Beijing and Shanghai have a salary of more than 17,000 yuan, and other regions only account for about 30%.
Educational level is also an important factor affecting salary levels. Judging from the data, among developers with high education background, the proportion of high-income groups is relatively high. Among the developers with master's and doctoral backgrounds, more than 50% of them have a salary of more than 17,000 yuan.
Sometimes the longer the hours, the higher the pay. The data shows that among developers whose working hours are more than 55 hours but less than 72 hours, 50% of developers earn more than 17,000 yuan, the highest proportion.
Daily life of programmers: the main job is to hold meetings, followed by writing codes, with the highest proportion of outputting 101-200 lines of codes every day
In the eyes of many people who are not in the industry, programmers hold a computer every day, keep typing on the keyboard, and output codes composed of various English letters. In fact, the data shows that in daily work, less than 9% of developers spend more than 70% of their time writing code every day. Less than 30% of people write code more than half of the time every day.
According to the code results, 75.5% of developers have no more than 300 effective lines of code per day. Developers who write 101-200 lines a day account for the highest proportion, at 20.2%.
So where does the programmer's time go?
42.7% of developers voted for the frequent meeting option. Secondly, factors such as distracting work environment and unclear work process have become the main "killers" affecting their work efficiency.
In addition to the above factors, it is said that young people today are rebellious and refuse to involve.
It can be seen from the survey data that more than 30% of developers need to work 40 hours a week, and nearly 40% of developers have no more than 3 hours of overtime per day.
For developers, the longer the working hours, the higher the dissatisfaction. Workers who work 40 hours a week (standard working hours) have the lowest level of dissatisfaction at 4%.
If personal life is compared with high-paying overtime work on the same level, 56% of the respondents resolutely reject "big and small weeks" and want normal rest time. Of course, 23% of the developers said that they are willing to implement the "big and small week" working system in order to get a high salary.
Software development tools: Python is rising due to AI, 34% of developers have used ChatGPT
From the perspective of the tools that developers rely on for survival, in the field of programming, the number of developers who used the Java language in the past year accounted for 42.9%. With the development of artificial intelligence, the use of Python has gradually increased, and developers who often use Python at work accounted for 31.2%.
Li Jing, stability manager of Ant Group's Yuque team, commented that Python has excellent readability, flexibility and more powerful scientific computing capabilities, which makes a large number of developers start to have a strong interest in Python. It also has a very active ecology, and more than a quarter of developers currently plan to learn and master this language in the next year. The advent of large models such as ChatGPT and Midjourney has made everyone see the charm of AI. With the improvement of computing power, the enrichment of data sets and the advancement of model training technology, large models are expected to gradually replace small and medium-sized NLP models in the future. The productivity of a single skill will be replaced, which is also in line with the current goal of reducing costs and increasing efficiency.
In contrast, low-level language assembly is the least preferred language for developers, accounting for 38%. In addition, C and C++ are also daunting to some developers because of their difficulty in use.
In recent years, Vue.js has become more and more popular in web development, and the data shows that 36.1% of developers have done a lot of development with Vue.js. In contrast, the usage of jQuery has been decreasing year by year, from 29.1% last year to 23.3%.
As a lightweight cross-platform tool, Visual Studio Code is used by 38% of developers in their daily work, ranking first.
With the development of large AI models, more and more AI tools can improve the development efficiency of developers. 45% of developers said they had used AI programming related tools, and 34% of them said they had used ChatGPT.
Nowadays, with the rapid rise of domestic large-scale models , the proportion of AI programming aids will only increase and will not decrease.
Ever since AI can help with programming, there has been a lot of discussion about whether it can replace developers. 61% of developers believe that current AI programming cannot replace developers. Of course, there are also a small number of developers who feel that AI programming is expected to replace existing developers.
In view of this, Wei Yongming, the founder of Feynman Software, believes that whether a new technology may replace the whole human being or part of human beings, or a certain occupation is determined by the replacement cost. The development of computer software to today's scale is the result of stacking layers of software. Even if AGI technology can design and implement operating systems, compilers, and databases by itself, no one will pay for AGI to do this, because economically speaking, this is extremely uneconomical. So substitution will happen, but only at the application level for the foreseeable future.
AIGC has created a brand new track
ChatGPT has become the fuse of the new generative AI track. Once released, it has led global technology companies to carnival for large-scale model development, which has also continuously driven the emergence of new market demands and new opportunities, thus attracting a large number of talents to this wave among.
The data shows that 41% of developers will start to dabble in AIoT-related development work in 2022.
In specific application scenarios, nearly 30% of developers will use AI portrait restoration tools. AI portrait restoration technology can help developers automatically repair and enhance portraits, and improve the effect and accuracy of face detection on application devices.
Also based on this direction, some developers have developed an app called " Miaoya Camera ", which unlocks photo blockbusters for 9.9 yuan, which has become the focus of many users' attention recently.
In terms of breakthroughs in AI technology, 26.6% of developers are optimistic about generative artificial intelligence, followed by large-scale data sets and large-scale model open source, accounting for 23.4% and 20.9% respectively.
With the continuous development of AIoT, some or all of human work may be replaced in the future. According to survey data, more than 90% of developers believe that smart devices can help us complete our work, and only 8% of developers believe that they cannot replace humans.
When further discussing whether artificial intelligence will have the ability of consciousness, 41% of developers believe that artificial intelligence may produce consciousness. It seems that many people still have rich imaginations about silicon-based life.
It cannot be ignored that the widespread application of AIoT technology will bring a series of ethical, legal and social impacts, and will also change our lives, society and economy.
Therefore, before AIoT technology is fully popularized, extensive social, cultural and ethical discussions and research are required. 36% of developers said that the discussion should start as early as possible, and that ethical, legal and social impacts should be considered at the basic research stage. The same number of developers also said that it is also possible to reconsider social use and implementation after productization and serviceization.
Yan Shuicheng, visiting chief scientist of Zhiyuan Research Institute, academician of the Singapore Academy of Engineering, AAAI, ACM, IEEE and IAPR Fellow, commented that according to this survey, the AIoT developer ecosystem is showing a trend of liberalization and expansion, and its product applications are also more popular. Diversification, which brings about an important issue, that is, the security issue of AI. AI security issues include several aspects:
The first is AI data security. The largest fields of AIoT applications, whether it is manufacturing, finance, logistics, medical care, security or autonomous driving, are all industries that require extremely high data security. It is necessary to ensure that AI can be used well. Private data, while protecting the security of the original data and the security of the result data after AI analysis;
The second is the system security of AI. These major areas where AIoT is mainly applied have extremely high requirements for security. Once hackers attack through AI system vulnerabilities, whether it is stealing data or tampering with results, it will cause huge harm. , the liberalization and expansion of the developer ecosystem has posed a huge challenge to this, and how to protect the AI system is also a problem that needs special attention in the future;
The third is the ethical security brought about by the formation of AI's own consciousness. At present, more than half of the developers believe that AI may or will definitely form self-awareness in the future. This process may be subtle, and preparations need to be made in advance.
Cloud native technology has become an important engine driving business growth
As an important supporting technology for new infrastructure, cloud native has gradually emerged in emerging fields such as artificial intelligence, big data, edge computing, and 5G. According to the survey and statistics of groups interested in cloud native, only 20% of developers said that their services have not been deployed on cloud services.
The advantages of cloud native are also the reason why developers prefer to use cloud native. 44% of developers said that cloud native can improve development efficiency. Of course, in addition to development efficiency, improving business agility is also an important reason to attract developers.
In terms of specific technical directions, 45% of developers are mainly involved in microservice architecture, which shows the importance of microservice architecture in the field of cloud native.
Wanbo Zhiyun CTO Sun Qi said that the continuous improvement of container technology has prompted more and more applications of micro-service architecture, and DevOps toolset is an important guarantee to ensure the implementation of micro-service projects.
Based on this, 63% of developers believe that mastering and using technologies such as Docker and K8s is the most important. Of course, 44% of developers believe that microservice architecture is also very important.
Stability is the primary concern of all developers, and the use of Serverless is no exception. Nearly half of developers are concerned about its stability, and the second concern is the improvement of development efficiency.
However, many developers still have concerns when using cloud-native technologies. 40.1% of developers believe that standardizing the API interface ranks first among the problems that need to be improved urgently. Moderate microservice splitting and unified configuration management are comparable, accounting for 30.5% and 31.5% respectively.
Domestic databases have entered a new stage of technological innovation and prosperity
Database technology has been developed for more than 70 years, and it has been developed in China for 40 years. So far, 31% of developers feel that domestic databases are developing rapidly, showing great potential and development prospects.
In their opinion, there are independent theoretical original content (38%), functional characteristics different from other databases (34%), being at the forefront of the world in the field of engineering (32%), and independent coding implementation based on existing theories ( 32%) is the foundation of domestic database innovation.
Li Haixiang, Chief Architect of Tencent Database, believes that the next revolution in database technology must be a breakthrough at the basic theoretical level. What is certain is that whoever is willing to invest resources in theoretical research will benefit.
Based on these dimensions, the data shows that the TiDB database continues to lead other competitors and become one of the most familiar domestic databases for developers. At the same time, AliSQL and OceanBase will also maintain their strong development momentum.
In terms of applications, the utilization rate of open source databases is the highest, accounting for 39%, which reflects that although open source databases have great advantages in terms of cost and customizability.
The report also pointed out that although the development of domestic databases is stable and improving, the cost of database migration and compatibility restrict the development of domestic databases. Therefore, the development of domestic databases also needs to pay attention to these problems, strengthen technology research and development and team building, and improve data Migration and compatibility capabilities to better serve our users.
Regarding the future development direction of the database, the developers' views mainly focus on the following aspects:
Multi-mode database, that is, one database for multiple purposes. This database development technology can integrate various data required by different applications into one database to meet the needs of multiple application scenarios, accounting for 52%.
AI fusion database has also become a popular direction. The database combined with artificial intelligence technology can realize the intelligent and automatic operation and maintenance of the database, thereby improving the performance of the database, accounting for 51%.
As a new type of database technology, the database deeply integrated with cloud computing will greatly enhance the data analysis and data processing capabilities, accounting for 40%.
Operating system shells are in a period of rapid evolution
As the country attaches great importance to information security and data sovereignty, the development of domestic operating systems has attracted more and more attention. In the development of domestic operating systems, some development teams and companies have upheld the spirit of innovation and independence, enhanced China's independent research and development capabilities, and realized the localization of operating systems. 23% of developers said that the current domestic operating system has met their daily needs.
Compared with foreign mainstream operating systems, domestic operating systems still face many challenges and problems, especially in terms of user experience, usage habits, upgrade and maintenance costs, and application compatibility.
Application compatibility is one of the most important issues that developers are concerned about. As many as 66.71% of developers believe that domestic operating systems have problems in application compatibility. Compared with foreign mainstream operating systems, the variety and quantity of domestic system applications may need more expansion and support. In addition to application compatibility issues, 48.19% of developers believe that domestic operating systems also have problems in user experience. 33% of developers are concerned about usage habits, and 23.62% of developers are concerned about upgrade and maintenance costs.
These reflect that the domestic operating system needs to be further optimized in terms of overall experience and quality, and it also needs to be better improved and responded to in terms of upgrade and maintenance.
Among the domestic self-developed operating systems, HarmonyOS (Hongmeng) is the most representative and market-influential system, and nearly 90% of developers have heard of it. This result shows that people attach great importance to Huawei, and also reflects the trend of HarmonyOS dominating the domestic market.
In addition to HarmonyOS, Deepin and Kirin are also relatively well-known systems among domestic self-developed operating systems. According to the survey, nearly 30% of the developers have heard of the Deepin operating system, and about a quarter of the developers are familiar with the winning Kylin operating system. These two systems have a wide range of applications and user reputation in some specific fields, and have also become important players in the operating system market.
When analyzing the main reasons for the gap in domestic operating systems, we found that the lack of a wide range of application ecology is the main problem restricting its development.
In addition, the lack of technical talents and continuous capital investment are also important factors affecting the development of domestic operating systems. Therefore, domestic operating systems need to strengthen the development and promotion of application programs, encourage more program developers to participate, and continue to invest in research and development to ensure the continuous improvement of core technologies.
Among the developers, 74% believe that if they want to promote the development of the ecosystem of self-developed operating systems, they must first improve the construction of an ecological adaptation system. Building an OS root community is also critical to driving its growth. At the same time, operating system research and development is different from other application development. It needs to deal with the compatibility of more types and more complex and changing hardware devices. Adaptability is a key link in the promotion of self-developed operating systems. Therefore, 44% of developers believe that more energy should be spent on the construction of the ecological adaptation system.
Finally, in the face of the future AI-driven operating system, Chen Yu, a tenured associate professor of the Department of Computer Science, Tsinghua University, said that the GPT (Generative Pre-trained Transformer) system based on the Transformer model that everyone is generally concerned about, such as ChatGPT/Bard/Wen Xin Yi Yan, etc. , promoting natural language as the new operating system shell. Various GPT systems are expected to form a new application ecology through the new application framework interface ChatGPT/Bard API interface. Applications that have emerged so far include itinerary assistants, life stewards, work secretaries, code interpreters, automatic website generation, shopping price comparison, document summarization, document auxiliary generation, etc. In short, with the continuous expansion of computer application fields and the rapid development of new technologies represented by artificial intelligence technology, the shell of the operating system will also develop and evolve rapidly.
The road to chips is long and difficult
A few months ago, ZEKU, a chip design company under OPPO, was disbanded in situ. When the executives announced the news, they choked up and cried several times. More than 3,000 employees lost their jobs overnight, which was moving. The road of domestic chips is full of thorns, especially in today's geopolitical and other factors that have affected the long-term prospects of the industry, it is unimaginable.
However, in order to vigorously develop the chip industry, a large amount of knowledge accumulation and the blessing of experienced developers are required. The data shows that the domestic talent reserve in this area is still relatively small, and only 6.0% of developers can deeply understand chip technology and apply it in depth, which limits the progress of chip research and development.
In reality, chip manufacturing is different from the software development process, and it cannot be iterated in small steps like software development, and the cost of the entire manufacturing process is relatively high. 56.62% of developers believe that in chip manufacturing, it is easy for the product application market to deviate from the assumption, resulting in waste of R&D investment and production costs. The second is that during the chip development process, some design specifications cannot be realized, and half of the developers are concerned about this.
Based on this, most chip teams mainly focus on chip development in a certain segment, so their scale is not too large. According to the survey, 40.42% of the companies have less than 10 employees.
In the application of chip development tools, the two languages most commonly used by chip developers are C/C++ and Verilog. The data shows that nearly 50% of developers are using C/C++ to write code; Verilog is a hardware description language, mainly used for modeling and simulation of digital circuits, and 12.94% of developers use it.
Artificial intelligence is booming, and more and more special-purpose chips are designed for the field of artificial intelligence. They are characterized by being highly optimized for specific computing tasks. Data shows that among domestic chip companies, 38.46% of chips are equipped with artificial intelligence technology, which can provide more efficient computing power for artificial intelligence applications.
Overall, domestic chips face many challenges and difficulties in the development, the following are some main aspects:
Design capability: Chip design is a complex and arduous task that requires superb technology and superb design capabilities. 39.91% of the developers said that it is difficult to reduce the cost of chip design with the current domestic design capabilities. The second is low power consumption design. 35.36% of developers said that it is very difficult to achieve low power consumption.
Patent protection: Chip manufacturing involves a large number of patented technologies. In the development of domestic chips, patent evasion is required.
Bao Yungang, deputy director of the Institute of Computing Technology, Chinese Academy of Sciences, said that in the new era, excellent chip design talents not only understand chip architecture, but also need to understand software stack knowledge such as operating systems. However, such talents are even more scarce in China, because many integrated circuit colleges do not offer software courses such as operating systems. To solve the problem of shortage of talents, the current concept and plan of talent training need to be changed, and more attention needs to be paid to the cultivation of software and hardware synergy.
Open source has become an integral part of software development
In the past year, the development of open source has shown a strong momentum and has become an inevitable trend in the software industry. According to survey data, 96% of developers are using open source software.
Java, Python, and C++ are their commonly used development languages, accounting for 41%, 27%, and 21% respectively.
Taking the company as a unit, many technology companies are developing InnerSource, building open source offices, etc., that is, applying the experience and lessons learned in the development of open source software to the practice of developing software within the company or organization, and accelerating code sharing within the company.
Most of these developers came into contact with open source projects through Internet products. In terms of contributions to open source projects, 72.9% of developers contribute through code, and 49% of developers contribute through documents.
Wang Wei, a professor at the School of Data Science and Engineering at East China Normal University, commented that these developers not only use open source software, but also contribute to the open source community by submitting code and solving problems. This shows that the open source community has become a platform for developers to learn, share and grow together, and it is also a driving force for the entire industry to develop in a more open and transparent direction.
However, it is a pity that "using love to generate electricity" has become the status quo that drives many individual developers to participate in open source. The data shows that 17% of developers volunteered their time, and only 8% of the respondents said that they participated in open source because the company paid their wages. 63% of developers even bluntly said that they have never received income from open source.
This is also the reason why vicious open source incidents such as the deletion of the "faker.js" and "colors.js" project warehouse developers and the tragedy of core-js are common in recent years. Zhuang Biaowei, director of the Open Source Society, spoke out for this. Many developers involved in open source are still using love to generate electricity, and they are still relying on their own love to support it. This situation has reached an urgent need to change.
As far as the technical field is concerned, facing the future, the development of artificial intelligence has attracted the attention of many developers. 45% of developers pay more attention to open source AI, followed by programming languages and open source big data.
thank you
The above content is mainly derived from the "2023 Chinese Developer Survey Report" conducted by CSDN in-depth research. Here, also very grateful to:
Wei Yongming, founder of Feynsoft
Li Jing, Head of Stability of Yuque Team, Ant Group
Yan Shuicheng, visiting chief scientist of Zhiyuan Research Institute, academician of Singapore Academy of Engineering, AAAI, ACM, IEEE and IAPR Fellow
Sun Qi, Wanbo Zhiyun CTO
Li Haixiang Chief Architect of Tencent Database
Chen Yu, Associate Professor, Department of Computer Science, Tsinghua University
Wu Qingbo Director of Domestic Basic Software Engineering Research Center, Ministry of Education
Bao Yungang, Deputy Director, Institute of Computing Technology, Chinese Academy of Sciences
Wang Wei, Professor of School of Data Science and Engineering, East China Normal University
Zhuang Biaowei Director of Kaiyuan Society
Guidance and support for this investigation report! And sincere thanks to Electronics Industry Press, Mechanical Industry Press, Tsinghua University Press, People's Posts and Telecommunications Press Asynchronous Community, and China Water Conservancy and Hydropower Press for providing valuable book support for this survey.
View full report content
Click " Mini Program " or " Read the original text "
Quick access to complete report content
Click on the official account card below to follow me
In the official account dialog box, reply the keyword "1024"
Get a free practical tutorial on making money from sideline business
