|
|
|
|
|
|
by obstacle3596
1283 days ago
|
|
|
In my opinion some of the most nefarious effects of this ChatGPT craze will be on open source software. Quality will diminish a lot, everyone will want to have a shinny GitHub portfolio without actually putting the effort, so they will use regurgitated code from GPT without disclosing the source. People will clone and use this projects, and by virtue of their open source nature, that code will be in the next iteration of the training data set, so it will be a spiral... And that is best case scenario, because my biggest fear is that this type of language model will scare good programmers from releasing novel/useful code to the open source community. Why would you want to provide the training data for free to OpenAI, and reap none of the benefits of your work? Currently people at least have to clone your code/examples, make an effort to understand it, and train so that they can market themselves as experts in <Spark, React, Kubernetes, etc> to advance their careers. |
|
|