AI and the top technologies 2023 - A summary of the stack overflow report 2023

The survey

As in our Stack Overflow 2022 article, we start with the setup of the survey: 89,184 software developers from 185 countries participated for 2023. However, more than a fifth of them come from the USA (21.21%). Together with Germany, which comes in second place with 8.34%, that's already a third of the total survey. Germany has overtaken India with 6.4% and thus the two countries swap positions from last year. Malicious tongues explain this phenomenon by the fact that the Indian developers were finally granted their EU Blue Card and could move to Germany. Spain was kicked out of the top 10 by Australia's move into 9th place with 2.36%.

‍

‍

Developer Profile

Similar to last year, more than half of the people (55%) who start programming are between 18-24 years old. Considering Stability AI CEO Emad Mostaque's statement "There Will Be No (Human) Programmers in Five Years", we are curious to see what this development will look like next year - in general we think it will change the way software developers work, but they will still be around.

The willingness to learn online has increased from 70% to 80% compared to 2022, i.e. through YouTube and videos in general, blogs, forums, etc. Still, 55% learn in your "traditional" way like educational institutions like school or university.

Programming, scripting, and markup languages

In terms of programming, Javascript (63.61%) takes the crown - and has done so for 11 years now! HTML/CSS follows with 52.9% and in 3rd place Python takes over from SQL with 49.3%. This is partly due to the fact that Python is especially popular with people who are just starting out in programming. Professional developers, on the other hand, rely on the tried and tested SQL from 1974.

The following languages also occupy the top 11 (see Fig. 2): TypeScript, Bash/Shell (all shells), Java, C#, C++, C and PHP. At 1.33%, Solidity, the language characteristic of Web3, is slightly below its value of 1.45% from the previous year.

‍

‍

Databases

In contrast to 2022, PostgreSQL could prevail over MySQL - almost every 2nd professional uses PostgreSQL (50%), while the newbies prefer MySQL (46%). MongoDB remains a popular database in both camps. One reason for this is the large number of platforms and languages MongoDB supports.

Cloud

There is no change in the cloud either: AWS tops the list as the most used platform with 49% overall as well as for professionals (53%) and newbies (19.4%). Alongside this, Microsoft Azure (26%) was able to defend its second place against Google Cloud (23.9%). One reason for this is the strong access and focus on companies and professional developers.

‍

‍

‍

Web frameworks and technologies

As in the previous year, Node.js and React.js dominate in the area of frameworks. This is hardly surprising given the dominance of Javascript as the top programming language. Of the first six places, five are Javascript technologies, only Angular as a Typescript framework falls out here.

‍

‍

Integrated development environment

For all developers, Visual Studio Code is still by far the preferred IDE, as it was the year before. The same applies to the second and third places in the ranking with Visual Studio and IntelliJ, which also occupied this position in 2022.

‍

‍

AI Search Tools

This part is now new in the survey. Unsurprisingly, ChatGPT leads the way with an outstanding 83%. In 2nd place comes Bing AI with 21% and in 4th position Google Bard AI with 9.9%, - this follows analogously to the times of their respective releases.

WolframAlpha in third place is new to us - and as we now know, a semantic search engine that can solve numerous tasks with the help of AI.

Perplexity AI, the product of the aforementioned Emad Mostaque, follows in 8th place.

‍

‍

‍

KI Developer Tools

We found the question about AI developer tools particularly exciting. When it comes to the top 3, professionals and newbies are unanimous: GitHub Copilot with a wide margin (55%) to second position Tabnine with 13% and in third place AWS CodeWhisperer. In the ranking of professionals, GitHub is used five times more than Tabnine (56% vs 11.7%). So it's even more exciting to find out which tools can reduce GitHub's omnipresence in the future - you'll find out in the next section!

‍

Most loved, dreaded, and wanted

Next up is the popular "Most loved, feared and wanted" section of the annual Stackoverflow survey. A certain look into the future of programming in the coming years. We focus on the "most admired" part in each area and only look at the languages, databases and of course the AI Search and Developer Tools. 

Programming, scripting, and markup languages

Rust is at the top with just under 85%, followed by Elixir, TypeScript and Zig. Zig is an up-and-coming modern C alternative that promises quite a bit of potential. Zig is also the best-paid language in 2023.

‍

‍

‍

Databases

For databases, PostgreSQL (71.3%) is just ahead of Datomic (70.5%) and Redis (69.9%).

Short excursion about Datomic, for those who don't know it: Datomic is a distributed database designed for applications that rely on scalability, performance and consistency. So it's commonly used in applications that require high availability, integrity and security, such as financial services, healthcare and e-commerce. Among others, Netflix and Facebook use the database.

‍

AI Search Tools

Here, too, ChatGPT narrowly won the "most admired" category with 77.7%, followed closely by Phind with 77.5% and WolframAlpha in third place with 68.2%. Phind (pronounced "find") is particularly exciting for developers, as it is specifically geared towards programming assistance and advertises with the slogan: "Personal search-enabled assistant for programmers." - Have you tried it out yet? Let us know in the comments.

‍

‍

AI Developer Tools

There is also no deviation in the top spot for AI tools for software developers, with GitHub Copilot scoring 72.3 %. But who moves up after that? Codeium is the closest with 65.1%, well ahead of Whispr AI (54.7%), AWS CodeWhisperer (52.7%) and Mintlify (52.1%, for documentation). Based on this list, it is already foreseeable that different AI applications will specialise in different areas.

Lastly, let's take a final look at the field of artificial intelligence.

‍

Artificial intelligence

Talking about specialisation. In the following illustration and the question about AI tools in the next year, a clear picture emerges. First and foremost, the step-motherly treatment of documentation (especially the, let's say "older" generations, whoop it up here), debugging, which has already cost us all our last nerve, and the evaluation of a code base. All great use cases in programming - who knows, maybe I'll start programming too - it's never been easier than it is today. 🙂

‍

‍