@jerryjliu0: I’ve gotten reasonable success by creating a skill that reflects my writing style, and hillclimbing the skill based on …
Summary
Jerry Liu shares that creating a skill to reflect his writing style and hillclimbing based on real samples solves 70% of the style problem, but notes that articulating clear insights remains a post-training challenge.
View Cached Full Text
Cached at: 07/04/26, 08:44 AM
I’ve gotten reasonable success by creating a skill that reflects my writing style, and hillclimbing the skill based on my real writing samples.
I think that solves 70% of the style problem, though doesn’t solve the inherent ability to articulate clear, differentiated insights, which is probably an inherent posttraining problem
Lee Robinson (@leerob): Are current LLMs incompatible with great creative writing?
I can’t tell if it’s cope or not, but it seems like even with the best models, I still can’t get them to write like humans would.
For coding, there is a verifiable reward like it compiling or tests passing. But for
Similar Articles
@spicycandy00: Three human-touch writing skills recommended — AI writing Skills worth collecting. There's no universal solution to removing the AI flavor; these three happen to cover three directions. @Khazix0918 Kazik, a hands-on practitioner in AI writing, only open-sources what he uses every day. human-writin…
Recommends three open-source AI writing Skills (human-writing, humanizer-zh, ljg-plain), each targeting a different direction for removing the AI flavor, and introduces a technique for combining multiple Skills through a bridging document.
@akshay_pachaar: Best way to create an agent skill: 1. Do the thing 2. Do the thing better 3. Even better 4. Now say "okay, make that a …
A quick tip on creating an AI agent skill by iteratively improving the task and then converting it into a skill.
@AlphaSignalAI: https://x.com/AlphaSignalAI/status/2069064122218717387
This article explores how AI agents can automatically write and optimize their skill files using techniques like SkillOpt from Microsoft Research, which treats skill documents as trainable state and delivers significant performance improvements. It addresses the challenge of manual skill tuning and presents frameworks like GEPA and EvoSkill as evolutionary approaches.
SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use
SkillCoach introduces a self-evolving rubric framework that evaluates and enhances LLM agent skill-use by analyzing skill selection, following, composition, and reflection, providing process-level supervision beyond outcome-only metrics.
@wlzh: On the same day, I uncovered two gems — Liu Run's Business Reference Distillation Skills and Luo Zhenyu's Quote Distillation Skills. What does that mean? It's not about learning how Liu Run and Luo Zhenyu write — it's about turning their writing systems into templates that AI can use directly. That is, you feed in a piece of material, and AI will use Liu Run's methods to deconstruct the structure, create analogies, and build SCQA logic chains; then switch to Luo Zhenyu's mode to generate quotes, insights, and personal engagement.
Introduces two open-source AI writing skill templates that mimic Liu Run's business analysis style and Luo Zhenyu's quote style, enabling AI to produce professional and shareable articles, lowering the barrier to writing.