« Machine Learning Has an AI Problem | Main | Data-Centric AI: the Systematic Engineering of Data to Build AI Systems »

July 06, 2023

Comments

Chris Ward

I am a software engineer for IBM, and my personal feeling about generative AI for software is that is hard and time consuming enough already to understand,revise for bug removal and feature implementation, and build appropriate test cases, and this will only get more time consuming if an AI comes along and effortlessly multiplies the lines of code count by a factor of 10 or so. And if the software is intended for use by a global business audience... as IBM software is ... all the more so.

My sort-of.nightmare has been that I would be presented with a million lines of code written at Chinese universities and given to the world as
open source. And my manager would give me the task of fixing a bug.
All the comments and variable names would be in Chinese, and I neither read nor write Chinese.

To this you can add a new nightmare that an AI will mak it 10 million lines ... should only take it a few seconds to do. That would be job security until retirement if my manager really wanted me to diagnose and fix a bug for the sorr of customers that IBM has.

The comments to this entry are closed.

My Photo

April 2024

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30        
Blog powered by Typepad