The rise of Artificial Intelligence (AI) has been a game-changer in many industries, including software development. With the development of AI tools, developers can now automate several aspects of the software development lifecycle, from coding to testing and deployment. This has led to significant improvements in the speed, efficiency, and accuracy of software development, enabling developers to build better products in less time.
As an example of said set of skills applied to multiple industries, ChatGPT wrote the first paragraph of this article. Although technically accurate and perfectly well written, it lacks those slight nuances that (for now at least) are needed to really deliver what we want in the way we want it.
Intro on AI that writes code
Before AI’s dramatic entry into mainstream media with ChatGPT’s skills, there had already been other breakthroughs into other forms of AI that helped many of us on different tasks, such as coding.
The reason ChatGPT became such a visible bomb is because if tapped into something anyone could understand and interact with: “Chatting”. Unlike coding, which is already pretty far fetched to most people, chatting is a daily skill mostly reserved for analog intelligent beings. However, AI has been around for longer than this.
Github’s Copilot was developed by OpenAI back in 2021 and uses a combo of natural language processing and machine learning to help developers write code. This was a huge breakthrough in the use of AI in general, and software development in particular. Large language models (LLMs) such as GPT-4 can suggest code snippets, answer technical questions and even write simple applications from scratch.
For the development community, this was probably a lot more baffling than looking at ChatGPT interacting with people. After all, most ppl can have a conversation, but not everyone can write code.
Given this new AI skillset, and knowing that software development is one of the most notoriously sought after skills these days, how will this impact the world? Are developers becoming obsolete? Not quite just yet.
Limitations and risks
For the time being and probably in the near future software development will retain its place in the sun, potentially more now than ever. Here are just a few reasons:
Machines need to be shown the way
ChatGPT doesn’t technically think about what is saying, instead, it analyses and recognizes patterns and reproduces the most fitting response based on a given context. Without any immense pool of data to base its opinion on, it will never have any.
This is a very clear limitation when facing new problems and specially when pioneering on any given field. Unlike AI, when we lack a frame of reference, humans will strive to create one.
AI isn’t always right
Although very opinionated and savvy, AI is not always right, nor can it be for that matter. There are many different ways to approach a specific problem and many times there isn’t one solution to fit all.
To make a decision on something one needs enough information, which AI can have plenty of, but also vision, creativity, experience and ultimately a dash of arrogance to make that ultimate decision on top of all the other possibilities.
Quality of code
Most of the time things are better done than perfect, but some things, especially in the long run, need to be perfect. Security is the classic example of this. If there is one thing where we can’t cut any corners, it is around security. Cyrex, our mothership expores these dangers in depth in https://www.linkedin.com/pulse/how-ai-technology-changing-your-games-security-cyrextechnet.
Here at Cloudoki, we take a lot of pride in the code that we built and how we build it. Each project is their own and for now, it’s still impossible for AI to jump into a tech lead’s shoes, understand a project’s past, future, business model, audience and quirks to decide on how to implement a specific feature. For now, we still need to be very opinionated and steer AI into this to make sure our products have the security, scalability and code quality our clients look for.
Opportunities & the future
Resting assured that, for now, developer’s status quo isn’t changing, we should instead embrace all the positive things that AI is bringing into the table, as developers.
Since there have been people writing code, there have been people fed up with writing the same basic pieces of code for a website, a function, a feature. Especially when going back & forth through multiple different languages, this can become very time consuming. With AI tools we can start to delegate these very opinionated but harmless pieces of code while we rush to focus on the parts that really matter.
Improving code quality
Although it’s still a controversial topic around AI, the truth is that we have been listening to software’s opinion on our own code for many years now. Every time your build fails due to some (community sanctioned) lint warnings, that’s basically software giving you the “you shall not pass” gesture.
With Copilot, you can crank it up to 11 and get amazing insights and tips on how to improve your code for readability and performance.
New set of possibilities
The state of AI as it is today, is already bringing a number of new business and technological possibilities that we haven’t even begun to fully explore. Moreover, it lights up a path of continuous improvements that can take us to insane new breakthroughs in virtually any field.
From a software house’s perspective, there are many new products and clients to steer through this. From a software developer’s perspective, this is a never ending playground of cool stuff to play with.
Cloudoki’s comfort zone is around new technology, this is why we can only embrace all of the new tech and challenges that come with it.
Just like us, society will also adopt, adapt and learn to delegate what it can and should to AI so it can focus more on places where only a human brain can make a difference.
Cut the BullSaaS on AI: Insights from Cloudoki’s Founder
Explore the transformative power of AI integration in the SaaS industry through an engagin...Read more