Do you trust code written by LLMs? Linus shares his take. [Credit: mastery_learning on YT]
@GithubProjects There is @coderabbitai that does an incredible job at helping code review already! As a maintainer, it has saved me more time than any other AI tools.
@GithubProjects Trusting generated code blindly is risky. Verification and testing remain crucial for reliable software.
@GithubProjects @grok give me the most important things from the video
@GithubProjects I have special respect for this man
@GithubProjects Half the folks running LLM code in India don't even worry about trust, they just need something that saves time. The rest check the output, then rewrite entire functions anyway. The middle path just vanishes.
@GithubProjects Trust usually depends on what the code is running and how visible the failure would be. Early mistakes from these tools still spook some folks, but plenty of people use the outputs without ever checking twice.
@GithubProjects Trust usually depends on what the code is running and how visible the failure would be. Early mistakes from these tools still spook some folks, but plenty of people use the outputs without ever checking twice.
@GithubProjects I would have never thought that he has a positive feeling about LLMs generating code. 👌
