mattsan
Joined 3 years ago
Joined 3 years ago
additionally, similar to how large PRs are more likely to just be skimmed and replied with a "LGTM!", an LLM missing some bad stuff but still producing a seemingly thorough review would increase the chance of the bad stuff making its way in.
allowing LLMs to write code would be fine if its truly verified by a human, but let another LLM hallucinate and cloud a persons judgement and you've got a problem
The main difference is that it has a pretty good base autopilot, such that you can give it a direction vector to point to.
The downside is that you have to learn a specific language for it, but you can also make your own language that transpiles as a fun project :)
Example of what's possible: https://www.youtube.com/watch?v=szsIL_0m2Qo