Stack Overflow Developer’s Survey is one of the most important resources available to understand the new trends on software development. I have dedicated some time to analyze the full report and share some insights on the impacts of the AI on this industry.

Before digging into the data, it is important to note that there might be some bias on this report, since the audience is composed of users who have an account on Stack Overflow and most of the answers came from users from the United States. I assume that every programmer nowadays use Stack Overflow, but not everyone has an account and contribute to it. Also, the market in the United States can be completely different from the market in your country, so this has to be considered when taking insights from this report. It can be a base for an overview of the software development market, but we must not take this data as a single source of truth for the general panorama of this industry.

Can AI help people learn how to code?

The first section where AI is mentioned on the report is Online resources to learn how to code. Although most of the respondents still rely on the technical documentation, it is very interesting to see that almost 40% of them stated that they use AI to help them learn. And I have to say that I am a developer who is constantly using AI to learn how libraries, frameworks and snippets of code work. I truly believe that AI will become an indispensable resource for learning programming in the next years.

It is also interesting to note the AI mentions on the section Technical documentation preferences to learn how to code. Both free and paid AI-powered search/dev tools are ranked lowest when respondents where asked what is the source of the technical documentation you use most often to learn to code. I believe people are interested in using AI to help them learn how to program, but when they need to get answers from a technical documentation they prefer to rely on the traditional methods.

The supremacy of ChatGPT

In the section AI Search and Developer Tools we can see that ChatGPT is still the favorite tool developers use the most and the second tool, Github Copilot, also relies on OpenAI technology to deliver it’s core functionality to the users.

I am a heavy user of both ChatGPT and Github Copilot and I have to say that I am with the majority on this one. I also think that is worth mentioning Codeium which is supposed to be Github Copilot’s strongest competitor, offering a free ( at least for now ) plan. The only reason I haven’t tried Codeium yet is because I have already paid for the annual plan of Github copilot, but I really hope that a free alternative can be available to help programmers do their job better without the need to leave they IDE.

AI Tools on path to mainstream in the development process

In the section AI tools in the development process we can see that less than 25% of the respondents don’t plan to use AI on their development process. Taking into account that some of them can be motivated by having fear that AI can take their job or simply are just used to their work this way and don’t wanna change, it is quite impressing that something that didn’t even existed two years ago is now either used or considered by 3 in each 4 of the developers who answered the survey.

The section AI tool sentiment also confirms this trend by showing that the vast majority of the respondents have a positive sentiment towards the use of AI tools for development.

Learning and Productivity are the main benefits expected from AI Tools

To be fair, I couldn’t expect anything far from this on the section Benefits of AI tools. As previous stated, I am a heavy user of AI tools and my expectation also fall heavily on these two benefits.

Developers still don’t trust 100% on AI Tools

One thing is very interesting to note on the section Accuracy of AI tools. Although the majority ( 40.3% ) of the respondents stated that somewhat they trust AI Tools just a minority of them ( 2.7 % ) stated that they highly trust them.

This means that we still have a lot to improve on these tools. People want AI Tools, but they are not yet completely satisfied with what they have today! This is a huge opportunity if you are an entrepreneur thinking of targeting this industry.

This is also highlighted again on the section AI tools’ ability to handle complex tasks where we can see that, again, just a minority ( 3.3% ) of the respondents consider the AI Tools they use very good at handling complex tasks!

Writing Code driving the demand of AI Tools

In the section AI in the development workflow we can see that the vast majority of the respondents are using AI Tools to help them with writing the code. To be fair this is quite expected since the output of the work of a developer are lines of code. Not saying that there is no brain work in there, but in the end, you get things done only the moment your code is shipped to production. Search for answers is also a strong use case among the developers who answered the survey. I use AI Tools in development a lot to help me find the best libraries, frameworks, architecture or approach to a specific problem. I pretty much started to skip Google when I need answers about a topic for my job as a software engineer. This also highlights how AI can be a threat to traditional search engines, who are now running to incorporate AI in their products.

New opportunities for AI Tools for development

In the section AI tools next year we have something very interesting. Although most of the respondents stated their are using AI Tools to help them write code and search for answers, when asked where they think AI will be used more in the next year, the majority answered Documenting and Testing code. I relate 100% with that because I HATE documenting and testing stuff ( honest take: I almost never do this ). It can be a waste of time Documenting and Testing code because people who like building products want to create features and see their users using them to solve problems. Documenting and Testing is a necessary step but one that most of the developers would gladly skip to focus on the tasks that will create value for the users. I recognise that maintaining a huge code base and quality control of an app is extremely important, but this is a job I would gladly delegate to an AI Tool.

Final considerations

In the section Are AI tools a threat to your job we can se that just a minority of the respondents are worried that AI can be a threat to their job while the majority believes that it won’t.

I also believe that AI won’t be able to completely replace a good developer or software engineer, but I also believe that AI Tools will be required for developing and maintaining complex code bases in the next years.

In the section Most important ethical issues for AI we can see that misinformation and disinformation are the main ethical concerns for developers while in the section Challenges with AI at work we can see again that the accuracy is the biggest problem and it is also mentioned that most of the current tools cannot comprehend the whole context of a complex codebase because of technical limitations.

I am very confident that the context limitation we have today will be soon fixed and we will have AI Tools capable of understanding a whole complex codebase and thus, we will have way less problems with accuracy. Again this is another huge opportunity of entrepreneurs. I myself think a lot about this problem of the context limitation on LLMs and how to solve it.

It was a good amount of data to analise and I have tried to be more objective as possible and also to contribute with my professional experience working with AI on B2B powered SaaS. I hope you have enjoyed the reading and if you think the insights on this post can be valuable for others, feel free to share it or mention it on your social networks, blog or newsletter.