TEDx · Building Better Software
My talk at TEDxSOAUniversity — what 'better software' actually means once you've shipped past the demo, and why most of it isn't about code.
also published at: https://www.ted.com/talks/karthikeyan_ng_building_better_software
In November 2017 I gave a talk at the inaugural edition of TEDxSOAUniversity in Bhubaneswar, Odisha. The theme was “CREATE with Butterfly Effect” — small choices that compound into outsized outcomes. My talk was on building better software.
Watch on TED.com.
The argument, in one paragraph
Most “better software” advice is about code — patterns, languages, frameworks. That’s not where the leverage is. The leverage is in the questions you ask before you write the first line: who are you really building for, what do you refuse to ship, what compounds, what decays. The teams I’ve watched ship the best software weren’t the ones with the cleanest architecture. They were the ones that had the most honest relationship with the problem.
What I’d add eight years later
(notes for future-me to flesh out before publishing)
- The intelligence-explosion era changes the input/output ratio of an engineer by 10–50×. “Better software” now means: better taste in what to delegate to models, better instincts for what humans must still own (judgment, accountability, relationships).
- Security is no longer a department. After running compliance for a regulated crypto exchange, my whole frame on “better” shifted from craftsmanship to reliability under hostile conditions.
- The boring systems still win.
Backstory
The talk was given at SOA University, Bhubaneswar, on the chapter’s maiden edition. Theme curators were big on the butterfly-effect framing — small actions, compounding outcomes — which lined up well with what I wanted to say about software being the result of thousands of small decisions, most of them invisible.
If you watched it and want to argue with anything, drop me a line. Best counterargument gets featured in a follow-up post.