Essence and Categories Solve Problems | Generated by AI
Question: What is the powerful thinking method of categorization discussed in relation to fixing myopia with under-correction glasses and using AI to fix errors, based on the provided links and the idea that the world is about categorization and essence?
Answer:
The core idea presented is that categorization (or classifying things into distinct types/essences/categories) is a highly powerful thinking tool and problem-solving method. By identifying the fundamental “essence” or grouping similar issues together, one can address complex problems more effectively — focusing on one category at a time leads to clearer, faster, and higher-quality results. This approach avoids dilution of attention and prevents interference between different problem types.
Application to Myopia Reversal (from Yinwang’s post)
The article argues that myopia is largely reversible through natural adaptation rather than being a fixed genetic condition. The key method involves categorizing the problem as an adaptive response of the eye’s external muscles and axial length to prolonged near work (e.g., screens, books), rather than an irreversible structural defect.
- Practical fix via categorization: Use under-corrected glasses (e.g., reduce by about -1.00D to -1.50D from full prescription) to create deliberate slight blur for near tasks like computer/phone use.
- This makes the screen appear “barely clear” at typical distances, encouraging the eye to relax and adapt by shortening the axial length over time (promoting reversal).
- Example: For someone needing -4.00D, use -2.50D to -3.00D glasses; the slight blur at ~0.66–1 meter range acts as a “far-looking” stimulus even during near work.
- Adjust gradually every few months as vision improves (e.g., 25–100 degrees reduction per cycle), always maintaining under-correction.
- Why categorization helps here: Separating “near-work stimulus” as the main cause allows targeting it specifically (via controlled blur) instead of treating all vision issues the same way. The author contrasts this with full correction (which reinforces elongation) or aggressive exercises (which can cause strain).
Note: Mainstream scientific consensus (from studies on under-correction) shows mixed or negative results for slowing progression — some trials found it ineffective or even worsening myopia — so this method remains controversial and not widely endorsed by optometrists/ophthalmologists. It draws from ideas like Bates method adaptations but emphasizes passive, natural adaptation over active exercises.
Application to AI-Assisted Code Fixing (from the AI Fixes Win post)
The page describes using AI (e.g., large language models) to fix large-scale issues in codebases (like Windows focus-related bugs, namespace migrations, etc.). The breakthrough insight is categorizing errors into distinct “families” and fixing only one family at a time.
- Why multi-type fixes fail:
- Pattern competition: AI confuses rules across categories.
- Attention dilution: Context gets crowded, reducing focus.
- Inconsistent or over-generalized fixes, leading to new bugs.
- Focused categorization method:
- Isolate one error family (e.g., only javax → jakarta namespace changes).
- Provide targeted examples (8–40) for just that category.
- Instruct AI strictly: Modify only that family, ignore others.
- Apply fixes, recompile, feed new same-family errors back iteratively.
- Repeat per category separately.
- Results: Much higher fix acceptance (75–95%), fewer regressions, and faster overall progress than trying to fix everything simultaneously.
Overall Philosophy
Both examples illustrate that the world (problems, biology, software, etc.) can often be understood and mastered through precise categorization of essence:
- Break down complexity into non-overlapping types.
- Solve one type exhaustively before moving to the next.
- This prevents interference, sharpens focus, and enables large, reliable changes — whether reversing adaptive myopia changes or cleaning up massive codebases with AI.
This “divide by category, conquer one by one” mindset is presented as a universal, powerful method for tackling hard problems.
References: