3 ms·
Personally, I’M terrified of people using ChatGPT for serious coding work. They will be seeding the future with vulnerable/buggy software that no one understand
by glau 3y ago
Personally, I’M terrified of people using ChatGPT for serious coding work. They will be seeding the future with vulnerable/buggy software that no one understands.
Granted, us humans are doing that now. Just many orders of magnitude slower. Probably slow enough that we can find/fix the important stuff.
The other aspect that terrifies me is the potential for nation state entities with deep pockets to inject vulnerabilities. What would it be worth to the NSA to seed the future with programs they could exploit?
- throwaway4aday 3y agoLLMs are also capable of analyzing, describing, and debugging software. There have been several papers published on this. So I wouldn't worry too much about any new buggy software being produced. In a few years, LLMs will probably be issuing PRs on your repos to fix things you haven't gotten around to or noticed yet.
- glau 3y agoThis actually seems like a great use case. Read and flag potential bugs for review. Although if it wrote the code in the first place how would it identify the bug? What it produced was something close to the most likely result given what it has seen so far
- throwaway4aday 3y agoThis paper explains it better than I could in a comment https://arxiv.org/abs/2304.05128 https://arxiv.org/abs/2304.05128
- mym1990 3y agoAre you able to back up the claim that ChatGPT code, when refined and deployed to production, is more buggy than code that is patchworked together by 10-20 stack overflow posts until something sticks?
- glau 3y agoI didn’t make that assertion. My fears are based on the speed that code of unknown quality, which is also poorly understood can be produced. I specifically said: Granted, us humans are doing that now. Just many orders of magnitude slower. Probably slow enough that we can find/fix the important stuff.