Why Do LLMs Rely on Raw Intelligence Over Effective Communication Post-Training?
matt_slotnick · x · 2026-08-13
@mattslotnick raises an interesting question regarding LLM capabilities: why aren't models post-trained on "effective writing" guides, or if they are, why does it seem ineffective?
He notes that for humans, extreme intelligence doesn't necessarily equate to being an effective communicator, and an inverse correlation often exists. He wonders why AI models seem to rely on raw intelligence in this domain rather than more functional post-training, suggesting that improving this area could be of high value.
More from Models
- Web Dev AI Leaderboard: Claude Takes First, Kimi Follows Closely — arena · 2026-08-14
- Hands-on: Gemini 3.7 Flash is Now Accessible via API — Angaisb_ · 2026-08-14
- Grok Leads AI Models in Document Understanding and Data Reasoning, Per Databricks — xiaosun86 · 2026-08-14
- Custom Agent Harness Test: GLM-5.2 Best Overall, DeepSeek Flash Top Value — codes_astro · 2026-08-14
- DeepSeek V4 Flash Makes AI Agents Affordable for Everyone — Teknium · 2026-08-14
- DeepSeek Ships V4 Pro, Open-Sources Agent Software, Raises API Prices — The Decoder · 2026-08-14