Very well reasoned summary, Hemant. Really appreciate the details involved. In most forums, I keep getting the same question from the audience and I have found this write-up (fireside chat) in the TCS Annual Report 2023 very useful. This was written by Mr Anantha Krishnan, CTO TCS and an Industry Veteran and it gives a detailed insight into why GenAI can be value accretive to the IT Industry. He also brings in the economics of why this phenomenon cannot be compared with farm mechanization, which led to effort deflation. I hope you find this useful too.
What is the evidence for this thesis (IT Industry wont become redundant with the advent of GenAI)?
KAK’s response - The evidence is empirical. Every new generation of technology has led to reduction in programming effort per function point. But while that has steadily fallen, aggregate spending on IT services has only risen year after year, over decades. Take for example, the switch from assembly language to C. Its compilers came with large, extensible libraries of reusable pre-defined procedures. A developer could invoke a procedure with one line of code in C and embed its entire logic in the codebase, without actually coding all of it from scratch. Three lines of C accomplished what took 30 lines in assembly language. The 10x effort deflation didn’t result in mass layoffs of programmers. Instead, there was an explosion in software development because the same IT team could now build ten times as many function points.
Similarly, enterprises adopted offshore outsourcing, it led to a big cost deflation, but nobody’s IT budgets deflated. Instead, those savings went into building new systems and volumes rose to fill budgets and spending on IT services has only expanded. Likewise with low-code, no-code platforms.
**Why is that so? Farm mechanization caused effort deflation and rendered the agricultural workforce
redundant in the West. **
KAK: With most goods and services, when the price falls, any increase in volume is limited by how much of that good or service the market can consume in a defined period. When farm mechanization reduced the cost of tilling, the increased demand for men in tractors was not large enough to compensate for the effort deflation because there was only so much land available
to till.
Demand for IT services behaves differently. In every enterprise, there is significant unmet demand. Every CIO has limited capacity for new system development, resulting in a requirement
backlog that never gets fulfilled. Technologies like generative AI or low code-no code can help a CIO expand capacity and accomplish much more with the same budget. But even then,
the backlog never goes away because there is no limit on business users’ ingenuity or competitive drive. Demand just rises to fill the incremental capacity created by new technologies. The emergence of new technologies triggers more ideas, experimentation and more demand for our services. To that
extent, business application of generative AI, along with other technologies, will itself drive the incremental demand that fills up the capacity it frees up through higher productivity."
Subscribe To Our Free Newsletter |