Self Improvement

Self-Refine: Iterative LLM Output Improvement via Self-Feedback

Self-Refine is a framework where LLMs generate feedback on their own outputs and iteratively refine them — no human feedback, no gradient updates. Shows that a single LLM can be its own critic and improve outputs across diverse tasks.

STaR: Bootstrapping Reasoning With Reasoning

Large Language Models Can Self-Improve

Huang et al. (2022) show that LLMs can bootstrap their own reasoning ability by generating chain-of-thought rationales, filtering with self-consistency majority vote, and finetuning on the high-confidence outputs — no human labels needed. A clean demonstration that LLMs can improve themselves without supervised signal.

Language Models Can Teach Themselves to Program Better

Haluptzok, Bowers, and Kalai show that language models can generate their own programming problems and solutions, verify correctness with a Python interpreter, then fine-tune on the verified examples — more than doubling test accuracy. A clean demonstration of self-improvement via external verification.

What Got You Here Won't Get You There (Book Summary)

James Clear's summary of Marshall Goldsmith's book on the behavioral habits that prevent successful people from reaching the next level — the 20 bad habits executives have that are actually making them less effective. Surprising because the habits are things that worked earlier in their careers.

Smart Guy Productivity Pitfalls

Tom Forsyth's (Book of Hook) essay on the specific productivity failure modes that affect intelligent people — analysis paralysis, over-engineering, and the tendency to confuse thinking with doing. A direct counterpoint to the 'just work harder' productivity genre.

Raw Nerve

Aaron Swartz's Raw Nerve is a collection of essays about psychology, self-improvement, and confronting uncomfortable truths about how we think. Written in late 2012, these are among his most personal and intellectually honest posts.

Two Important and Often Overlooked Aspects of Creating a Lasting Morning Routine

Joel Gascoigne (Buffer co-founder) on two underrated aspects of building a lasting morning routine: making it enjoyable and starting small. Part of his transparent self-improvement blogging that built Buffer's early audience.

How to Code a Life

A BuzzFeed essay framing life decisions using programming metaphors — an early example of using software concepts to explain life choices to a general audience. Part of the 2012 wave of coding-as-metaphor content.

Why Most People Dream and Only Some Do: The Go-Getter Theory

A personal development post arguing that the gap between dreamers and doers isn't talent or luck — it's the willingness to take the first uncomfortable action when no one is watching. The 'go-getter' distinction is about tolerance for initiation costs, not capability.