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·24d ago·5 min read
Stop Telling Me to Ask an LLM
TL;DR
Relying solely on large language models (LLMs) like Claude for complex questions can lead to frustration when nuanced, experiential insights are needed. Personal interactions often yield richer, context-driven answers that LLMs cannot replicate.
✦ Why It Matters
Prioritize direct communication with experienced colleagues for complex problem-solving instead of relying solely on LLMs.
Key Takeaways
- Engage with experienced colleagues for nuanced insights.
- Avoid defaulting to AI for complex, unresolved questions.
- Recognize the limitations of LLMs in providing personalized advice.
- Value human opinions over generic AI responses.
Full Summary
The author reflects on the limitations of using large language models (LLMs) like Claude for complex, nuanced questions that require personal insight. Despite spending time querying these models, the author finds that they often still have unresolved questions that only experienced individuals can answer.
The article highlights a growing trend where people default to suggesting LLMs as a response, which can feel dismissive and unhelpful. The author emphasizes the importance of human experience and opinion, especially when faced with conflicting information or a lack of industry consensus.
While LLMs can provide information, they cannot replace the value of thoughtful, personalized advice from someone with years of experience. This distinction is crucial for fostering meaningful conversations and problem-solving in professional settings.