We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
Hacker News·19d ago·5 min read
The LLM Critics Are Right. I Use LLMs Anyway
TL;DR
Despite widespread criticism of large language models (LLMs), many engineers, including those developing LLM tools, continue to use them. At a recent conference, a developer revealed that they auto-close most pull requests (PRs) generated by LLMs to maintain quality.
✦ Why It Matters
Engineers should consider implementing strict PR review processes to filter LLM-generated contributions effectively.
Key Takeaways
- Acknowledge the valid criticisms of LLMs while using them.
- Implement the '/grill-me' technique to enhance understanding.
- Focus on producing high-quality outputs rather than quantity.
- Consider the implications of LLMs on junior engineers' roles.
How It Works
The '/grill-me' technique involves asking the LLM detailed questions to ensure a shared understanding of the problem before proceeding, which helps maintain quality and relevance in the output.