"Computer World" magazine once wrote an article saying that "programming will disappear by 1960" because IBM developed a new language FORTRAN, which allows engineers to write the mathematical formulas they need, and then Submit it to the computer to run, so the programming ends.
A few years later, we heard a new saying: Any business person can use business terms to describe their problems and tell the computer what to do. Using this programming language called COBOL, companies no longer need programmers. .
Later, people said that IBM had developed a new language, RPG, that allowed any employee to fill in forms and generate reports, so most of the company's programming needs could be handled by it.
In the 1980s and 1990s, we had components and graphical development, and business personnel could develop programs by dragging and dropping.
Entering the 21st century, Low Code, No Code appeared again. Since there is no code anymore, there is no need for programmers.
What is the actual situation?
Not only has programming not been ended, the threshold has been continuously lowered, and there are more and more programmers.
Now the ultimate big boss - the big model - is here. It is different from the previous new technologies and new languages. It does not require people to write the code, but directly generates the code.
Can it end programming and make programmers obsolete?
First of all, we have to admit that the industry's top AIGC tools, such as GPT-4, now have very powerful code generation capabilities. Some people on the Internet have used GPT-4, Midjourney, and DALL·E 3 to generate a code similar to Angry Birds. Game: "Angry Pumpkins", the effect is amazing!
However, the code generated by AIGC cannot be comprehensive or correct. Programmers have to check, debug, and guide it to make modifications after discovering problems. This is a very laborious task.
Mastering tools like GPT-4 can make programmers become super individuals.
Secondly, what many people don’t realize is that writing code is only one part of software development. Before writing code, there is also requirements analysis and design (architectural design and detailed design) that need to be done.
AIGC has made great progress in detailed design, but the architectural design is not yet good. With the rapid iteration of AIGC, within 5 to 10 years, AI with strong design capabilities and code generation capabilities may be available, but it will be difficult to surpass AIGC. The last big mountain: needs analysis.
It is very difficult to figure out what customers need. You have to understand their field, their business, and their processes, and then through constant conversations and confirmations, you can roughly figure out what they need.
Many times, customers only suddenly realize when they see the finished software: Oh, the function I want is not like this.
All programming-related AI tools on the market, such as GitHub Copilot and Amazon CodeWhisperer, cannot generate code directly from customer needs. They all require programmers to tell it exactly what to do in jargon.
Simply put, there is no way to go from customer needs directly to code generation.
However, FuncGPT (FuncGPT) launched by the full-stack fully automatic software development tool SoFlu software robot can improve this situation to a certain extent.
As an important part of Feisuan SoFlu software robot, FuncGPT (Hui Function) supports the creation of all types of functions. Use natural language to describe Java function requirements and generate high-quality, highly readable Java function code in real time. The generated code can be directly copied to IDEA, or imported into the Java fully automatic development tool function library with one click.
Because FuncGPT (FuncGPT) has undergone large-scale machine joint training, it has already had in-depth learning and understanding of Java syntax. Therefore, FuncGPT (FuncGPT) can generate high-quality function code that conforms to Java syntax rules.
Specifically, FuncGPT (FuncGPT) will follow Java syntax specifications when generating function code to ensure that the generated code is syntactically correct and complies with Java programming standards. This allows engineers to focus on describing function requirements and implementing business logic without worrying too much about grammatical errors.
When development engineers are faced with situations where they are not familiar with certain Java syntax rules, the existence of FuncGPT (Functions) provides them with a powerful auxiliary tool. Development engineers can input function requirements into FuncGPT (Hui Function) in the form of natural language, and it will generate function code that conforms to Java syntax for them, avoiding coding difficulties caused by syntax errors.
FuncGPT (FuncGPT) five major capabilities:
● Natural language: Functions can be generated through natural language, lowering the threshold for software development.
● Second-level function development: Fully automatic second-level function development, completed in seconds, improving efficiency by a thousand times.
● One-stop function development and application: functions can be directly applied after generation, with guaranteed efficiency and safety.
● Reliable quality: The generated function code complies with industry standards, has good readability, and is consistent with the optimal solution.
● Flexible generation: It can be generated according to the specific needs of developers and supports adjustment and modification.
It can be said that the arrival of FuncGPT (FuncGPT) brings auxiliary programming functions that can greatly improve developers' programming efficiency and free them from the daily simple and repetitive "writing code", which originally costs 80%. The coding time is left to AI tools to complete. The SoFlu software robot is also equipped with special functions such as real-time code visibility, human-computer interaction-intelligent brain, and other special functions. On the basis of helping developers improve both efficiency and quality, they can free themselves to focus more on 20% of the work involves thinking deeply about problems from a higher perspective, changing from "programmer" to "architect", from "writing code" to "designing programs", and completely unleashing the innovative potential of every developer.
And its surprises don’t stop there. There are more exciting functions. You can follow the [SoFlu Software Robot] official account to apply for a trial.
Spring Boot 3.2.0 is officially released. The most serious service failure in Didi’s history. Is the culprit the underlying software or “reducing costs and increasing laughter”? Programmers tampered with ETC balances and embezzled more than 2.6 million yuan a year. Google employees criticized the big boss after leaving their jobs. They were deeply involved in the Flutter project and formulated HTML-related standards. Microsoft Copilot Web AI will be officially launched on December 1, supporting Chinese PHP 8.3 GA Firefox in 2023 Rust Web framework Rocket has become faster and released v0.5: supports asynchronous, SSE, WebSockets, etc. Loongson 3A6000 desktop processor is officially released, the light of domestic production! Broadcom announces successful acquisition of VMware