I remember that when ChatGPT was first launched, most of the early adopters and test results I saw were developers doing code-knocking tests. It can be said that a group of people with a strong sense of professional crisis.
In addition, there are actually some technical barriers to using ChatGPT, and most of the people who are willing to toss are programmers, so of course everyone will think that this group of people should use the most, but until I saw a global User research is still very interesting.
ChatGPT user survey results
The latest statistics from Measurable AI show that designers who use this AI are five times more than developers, and programmers only account for 4.4%.
It's still a bit of a shock, but it's not just me. Most of the research conducted by this organization on twitter thinks that developers will be the first.
However, real user usage shows that people who use Adobe account for 18.5% of the total number of ChatGPT users, and those who use Canva account for about 9.2%; these two software are the designer's drawing baby.
In addition to designers, there are also many text workers who use this AI tool. For example, Substack users account for 10% of the total, and Grammarly users account for 6.8%.
And perhaps somewhat surprisingly, GitHub users account for only 4.4%, much less than those who draw pictures and write articles.
Reasons why ChatGPT is not seen by developers
As for why ChatGPT is not welcomed by developers, in fact ChatGPT itself has some answers.
In any case, it is more or less reasonable, in addition to some analysis from my point of view.
1. Developers are the first to eat chicken and get tired of playing
ChatGPT came out in December last year. I remember that it was like a thunderbolt on the ground, and the number of registrations ushered in a peak; in January this year, the popularity dropped slightly.
It just so happens that the developers are basically the first batch of users, and they have almost played it. To be honest, I personally think that there is still a lot of room for improvement in its current maturity. Now it is enough to play a few times. , there is little motivation to play again.
Therefore, the user data fed back by this organization basically did not cover the first batch of developer user groups.
2. Currently, ChatGPT programming maturity is limited
I am not as good as Copilot in writing code. After all, Copilot is "designed for code generation". Many developers also reported that ChatGPT can write some code, but! The code contains some errors. For example, the following blogger's question:
But if you really have been a qualified programmer, you will know that the time spent writing code only accounts for 20%. In the other 80% of the time, more than 50% is debugging and modifying the code, and 30% is communication and coordination, including discussing demand synchronization information and so on.
But I also believe that in the future ChatGPT will definitely improve the ability to write code to a certain extent, with fewer and fewer errors, and even help developers write part of the code to reduce the developer's workload.
3. Application scenarios limit the value of developers
If it is just a simple early adopter, it is actually difficult to have a continuous use effect. Basically, after asking two questions to satisfy curiosity, it will not be used frequently, unless a specific scene is found to allow ChatGPT to truly play its value.
Generally speaking, the biggest difficulty in the development of technology-based products lies in finding a variety of application scenarios, allowing them to be continuously optimized and iterated, and then continuously expanding the scenarios to achieve commercialization.
Application scenarios of ChatGPT
1. Copywriting generation
ChatGPT can be used to generate various types of text, and it is also the scene with the widest application range and the largest number of users at present. It can be used in emails, video scripts, copywriting, translation, code, papers, etc., according to the keywords given by users or topics, complete the text according to the beginning or end given by the user, translate according to the source language and target language given by the user, etc.
I won’t say much about this point. Everyone is very slippery in writing poems and articles in ChatGPT. I still feel better after experiencing it, such as the outline of this article below.
2. Search engine optimization
Search engine optimization (SEO) is the process of improving a website's ranking and traffic in search engines by improving aspects such as website content and structure. By using ChatGPT, it can help websites to produce content that is more in line with search engine rules and user needs. For example, ChatGPT can generate web page titles, descriptions, tags, etc. based on keywords, and ensure that the content is highly relevant and valuable to the topic. In addition, ChatGPT can also generate long-tail keywords, internal links, external links and other elements to enhance the weight of web pages.
It is more helpful for products, operations, etc., which is equivalent to the role of a small assistant.
3. Customer Service Chatbots
ChatGPT can be used to create customer service chatbots that can provide information or entertainment, such as answering users' frequently asked questions, providing product or service introductions, handling user complaints or suggestions, etc.
It can be used in entertainment, consulting, customer service and other fields. I think this is the direction that should focus on mining value in the future.
Traditional chatbots are usually based on rules or preset dialogue processes, lack flexibility and intelligence, and cannot handle complex and diverse user needs. ChatGPT, as a chat robot model based on deep learning, can generate reasonable and coherent responses based on context and user intent, and has the ability to self-correct, challenge wrong premises, and reject inappropriate requests. ChatGPT can improve chatbot user satisfaction and retention while reducing development and maintenance costs.
Especially at present, developers have directly connected ChatGPT to their official accounts, web pages, and Apps, which will be very helpful for future customer service consulting work, and can also greatly reduce the labor burden.
4. Code writing
Although the ability in programming is not perfect, it can help developers to reduce a certain amount of workload to a certain extent, for example: programming language conversion, such as Python to Java. From my personal experience, currently supported languages include Python, Java, Go, Node.js, R, C++, etc.
In addition, it is also recommended to use it to generate program commands. This is what I have been using in my work, which can greatly speed up code development. There are also some SQL statement generation, generation of tables in unstructured data, etc. can be tried to use.
Finally, there is a 0-threshold experience path for friends who have not yet been able to experience OpenAi's native ChatGPT. There is a ChatGPT applet on the homepage of FinClip App, which can be used by clicking on it. It feels pretty good after the experience.