Will AI Replace Programmers ?

‍

The first reaction of a developer when they hear about Dualite has always been will AI replace programmers and whether they’re gonna lose their jobs ? 

‍

And the answer is NO. Just as calculators didn’t make mathematicians lose their jobs, but only made their work faster, Dualite is a tool for developers to catalyze their stressful development work by automating the repetitive task of manually coding UI designs from a Figma file to code through the power of AI. 

‍

We believe developers won't be replaced by AI, but those who don't use AI-driven tools will fall behind significantly. With Dualite our mission is to empower developers to build faster and better, and truly focus on the things that really matter, for there’s no substitute in any LLM model or GPT for Human intelligence.

‍

That being said, there are still larger consequences to be aware off and plan cautiously, something which we’ll discuss 

‍

Moving to a higher ground

‍

If you were building toys as a laborer in a factory 50 years back, you probably could’ve made 100-150 toys per day. But now, a machine can probably produce 5000 in the same timeframe. 

‍

On an immediate note you might think that your core job has been automated. The task of 5000/100 = 50 workers to be precise has now been replaced by a single machine. But if you look at it from a different perspective, a person who now sits at the operation panel, controlling the same machine, earns significantly more per day in comparison to if he was a working laborer. 

‍

So just like the emergence of a machine resulted in the demand of more skilled operators and technicians who are much more prosperous, the advancement in AI-based technologies will only allow far greater productivity and therefore better compensations.

‍

The idea of job replacement is very much true. But the definition of work itself has changed. We’re moving to an age wherein even with very small efforts, we can generate a much higher compensation.

‍

Furthermore, this shift from traditional linear work systems, i.e, you’re compensated based on the hours you put in, is slowly shifting to a nonlinear one wherein a skilled worker with minimum input you can get significantly higher output, and therefore higher compensation.

‍

Today’s designers and artists leverage AI products like Midjourney and tools to quickly get from idea to product with just a few prompts. The prompt writing itself now has become a skill which few people have leveraged to get high-quality outputs. And the final touches or finishing still need manual intervention and refinement. 

‍

Similarly, software development cycles are going to get transformed where people use AI coding tools and prompts to fix bugs, automate repetitive tasks like UI development and build a better and cleaner documentation. 

‍

You can read more about these nonlinear work systems here

‍

But purely looking at it through a macro-economic lens, the other 49 people who lost their job due to a lack of education in operating a toy machine face challenges and difficulty. Thus, we have to aggressively integrate these new-age technologies in our learning systems and institutions so that this lack of awareness does not lead to inequality. 

‍

People like Musk also suggested the introduction of a universal basic income due to the potential of technology to create inequality by concentrating huge wealth in the hands of a few. He further argues that a harder social challenge would be to give meaning to people’s lives since a lot of people derive purpose and a sense meaning from their work.

‍

When’s the Terminator coming for us

‍

The next question probably comes of whether AI has achieved human, or even beyond human intelligence.

‍

ChatGPT is a great natural language search engine, and an incredible translator from computer languages to human languages. It can extrapolate and scale up a paragraph from a single sentence, but distilling down to core things and explaining something to the core, which is a sign of true intelligence, it falls apart.

‍

Secondly if you give it anything outside its training set, it’ll fall apart and not actually reason through it. A great example was on Twitter wherein a user asked “There are two knights - they’re both telling the truth. How do you find out which one is telling the truth?” A simple answer was just asking any one of them. But because ChatGPT was trained on a dataset of zillions of examples of the famous two knights problem, where one is lying and the other is telling the truth, it couldn’t simply ‘think’ beyond that and went through a convoluted process and couldn’t do simple rational reasoning that even a 5 year old could have.

‍

Thus, the right question is whether AGI - Artificial General intelligence as it is known - has well and truly arrived. An actual conscious machine which can critically think, reason, and rational. That is actually concerning. 

‍

Conclusion

‍

But practically speaking, the complexity and the layered nature of the world we live in, the current technologies can expand more and more and become more knowledgeable, but will never be able to sufficiently compensate for its lack of understanding. Sure, within specified, well-defined use-cases and domains, it can certainly exceed human abilities in the way that a calculator exceeds an average human’s math capabilities. 

‍

Understanding and grasping something is fundamentally standing on top of knowledge and digging deep into it. AI might spit out a definition of “love”, but it truly can never understand the depth and gravity of it that we all humans do.

‍

Author

‍

Rohan Singhvi

CEO of Dualite

LinkedIN

‍

‍