Menu Close

The age of Multi-discipline

the AI age of multiple disciple and emotional intelligence

In the future, developers don’t write code. We debug code

Following the insights from Github Universe, I am now more convinced than ever about a pivotal shift in the future of software development. The role of developers is evolving, steering towards becoming professional debuggers, akin to doctors in their approach. This transition sees developers integrating various elements, stretching their expertise across multiple domains with a remarkable depth of knowledge.

Imagine blockchain engineers who not only excel in their field but are also adept at full-stack development and possess a comprehensive understanding of cybersecurity. Envision mobile developers with a robust grasp of web mechanics, bridging the gap between platforms seamlessly.

This future heralds a move away from narrow specialization towards a broader, more integrated approach to learning and application.

My expectations are that training and education will pivot to focus on first principles, laying a solid foundation from which to navigate the complexities of technology. It’s a future where adaptability and foundational knowledge eclipse niche specialization, shaping a more versatile and dynamic breed of developers.

Apart from developers, I also think Muscle Memory(MM) and Emotional Intelligence (EQ) would be the greatest assets any human wields


The Devin is in the details

Since I last wrote about this topic in early November, there has been some recent advancements in AI assistants that can write code, one of which is Cognition Labs recently released AI programmer called Devin. Which has been the most remarkable so far.

From their website the mention that with the recent advances in long-term reasoning and planning, Devin can plan and execute complex engineering tasks requiring thousands of decisions. Devin can recall relevant context at every step, learn over time, and fix mistakes.

Devin is equipped with common developer tools including the shell, code editor, and browser within a sandboxed compute environment β€” everything a human would need to do their work.

I think one of their most remarkable implementations is the ability to report on progress in real time, where Devin can accept feedback, work together with a human through decision choices and collaborate on work. Machine learning folks have a term for this called H.I.T.L; and if we’ve learnt anything from Self Driving Cars. HITL is the most reliable implementation of Autonomous AI β€” So it’s safe to say that Devin is driving in the right direction.

Till then.


Skip to content

Share This

Copy Link to Clipboard