LabHub

Blog

The Developer Learning Complete Guide: New Languages, New Domains, Papers, Lifelong Learning, and AI Learning Partners (2025)

2026-04-15 · 15 min read · #learning#skill-acquisition#programming-language#papers#ai-learning#zettelkasten#obsidian#career#lifelong-learning#deep-dive#english

한국어English日本語

Intro — "Ten years in and there is still so much to learn"

A senior in her late 30s:

"I tried Rust again and quit for the third time. Maybe my learning ability has slipped with age."

Answer: absolutely not. Most of what feels like "declining learning ability" is actually a problem with learning method.

Research:

This post covers:

  1. The science of learning — Carey, Oakley, Ericsson
  2. Conquering a new programming language in 30 days
  3. Entering a new domain (ML, Security, Finance, Embedded)
  4. Reading papers — the 3-pass approach
  5. Zettelkasten + Obsidian system
  6. How to actually finish a technical book
  7. Coursera, Udemy, bootcamp ROI
  8. Using AI as a learning partner
  9. Learning in your 40s–50s

Season 3 Episode 11. The last episode covered recovery from mental-health issues; this one is "how to grow from a recovered state".

Chapter 1: The Science of Learning

1.1 Cognitive Load Theory

John Sweller (1988):

Bad instruction: lots of extraneous load. Good instruction: focuses on intrinsic, induces germane.

1.2 Desirable Difficulty (Bjork)

If it is too easy, it does not stick. A little difficulty strengthens memory.

1.3 Focused vs Diffuse Mode (Barbara Oakley)

Focused: concentrated study, detailed understanding. Diffuse: the brain reorganizes during rest.

You need both. The shower epiphany is diffuse mode at work.

1.4 Deliberate Practice (Anders Ericsson)

Not mere repetition:

  1. Specific goal
  2. Focus
  3. Immediate feedback
  4. Attack weaknesses
  5. Uncomfortable zone

Developer example: random Leetcode grind is worse than working on weak DP problems with feedback.

1.5 The misunderstood "10,000 hour rule"

Popularized by Malcolm Gladwell. Ericsson's actual message:

"Quality, not quantity. Ten thousand hours of zoned-out practice will not make you an expert."

Chapter 2: Conquering a New Programming Language in 30 Days

2.1 30-day roadmap

Week 1: Syntax basics

Week 2: Data structures + error handling

Week 3: Concurrency + ecosystem

Week 4: Real project

Rust:

Go:

TypeScript:

Python:

2.3 Common strategies

1) Connect to what you already know:

2) Learn the language-specific idioms:

3) Read real OSS code:

2.4 "Beyond Hello World"

A common mistake: replaying tutorials forever, never actually coding.

Force the issue: on day 30, publish a blog post "What I learned building X in Rust".

Chapter 3: Entering a New Domain

3.1 ML/AI

Path:

  1. Linear algebra + probability basics (Khan Academy)
  2. Python + NumPy + Pandas
  3. Andrew Ng's Coursera (Machine Learning Specialization)
  4. Fast.ai (practical)
  5. Andrej Karpathy's "Zero to Hero" series (YouTube)
  6. Papers with Code (reimplement)

Timeline: 6–12 months.

3.2 Security

Path:

  1. Networking basics (TCP/IP)
  2. Web security (OWASP Top 10)
  3. HackTheBox, TryHackMe (hands-on)
  4. CTF competitions
  5. Bug bounty (HackerOne, BugCrowd)
  6. CISSP, OSCP certs (optional)

3.3 Embedded

Path:

  1. Relearn C fundamentals
  2. Arduino → Raspberry Pi → STM32
  3. RTOS (FreeRTOS, Zephyr)
  4. Rust for embedded (recent trend)
  5. Hardware (oscilloscope, logic analyzer)

3.4 Finance (Fintech, HFT)

Path:

  1. Finance basics (stocks, bonds, options)
  2. Time series analysis
  3. C++ / Rust (low latency)
  4. Book: Options, Futures, and Other Derivatives (John Hull)
  5. Korea: financial certs (AICPA, CFA) — not required

3.5 DevOps/SRE

Path:

  1. Linux in depth (The Linux Command Line)
  2. Networking (TCP, DNS, HTTP)
  3. Containers (Docker → Kubernetes)
  4. Terraform / Pulumi
  5. Prometheus, Grafana
  6. Site Reliability Engineering (Google book)

3.6 Shared principles

Chapter 4: Reading Papers

4.1 S. Keshav's 3-Pass Method

Pass 1 (5–10 min):

→ Decide: "do I even need this paper?"

Pass 2 (1 hour):

→ Get: "what is the author claiming?"

Pass 3 (4–5 hours):

→ Complete understanding.

4.2 Which papers first

Classic CS papers:

Where to find them:

4.3 Paper reading groups

Ten times more sustainable than solo reading.

4.4 Reading papers with AI

Chapter 5: Zettelkasten and Obsidian

5.1 What is Zettelkasten

Niklas Luhmann (sociologist; 90,000 notes over 70 years):

5.2 Why Obsidian

5.3 Basic structure

Folders:

00_inbox/       # new notes
10_zettels/     # permanent notes (atomic)
20_literature/  # paper/book summaries
30_daily/       # daily entries
90_attachments/ # images

Note template:

---
id: 20250108-spaced-repetition
tags: [learning, memory, science]
source: [[book-make-it-stick]]
---

# The principle of spaced repetition

Long intervals between reviews are more effective
for long-term memory than short intervals.

## Why

- Re-exposure right before forgetting creates the strongest signal
- Spacing effect (Ebbinghaus, 1885)

## Links

- [[deliberate-practice]]
- [[anki-workflow]]

5.4 Daily notes

Daily Note: one file per day.

Promote to permanent notes during the weekly review.

5.5 Plugin recommendations

5.6 Compared with Notion

Obsidian strengths: personal knowledge, offline, ownership. Notion strengths: collaboration, databases, sharing.

Use both: Obsidian for personal notes, Notion for team work.

Chapter 6: How to Actually Finish a Technical Book

6.1 "Finishing" is not the goal

Most technical books are reference material. Reading cover to cover is wasteful.

6.2 Adler's four levels from How to Read a Book

  1. Elementary: reading the words
  2. Inspectional: skimming (TOC, preface, conclusion)
  3. Analytical: detailed analysis
  4. Syntopical: comparing multiple books

6.3 Strategy for technical books

Level 1: Inspectional (1 hour):

Level 2: Selective (5–20 hours):

Level 3: Deep (50–100 hours):

6.4 Classics guide

SICP (Structure and Interpretation of Computer Programs):

CLRS (Introduction to Algorithms):

DDIA (Designing Data-Intensive Applications):

Operating Systems: Three Easy Pieces:

The Pragmatic Programmer:

6.5 Skim vs deep read

Heavy with code/math = deep read. Essay/perspective books = skim is fine.

6.6 How to not forget what you read

  1. Annotation: underline and margin notes
  2. Summary notes: one-page summary per chapter
  3. Blog: public writeup
  4. Teach: study group
  5. Apply: use it in actual code

Chapter 7: Online Courses — Coursera, Udemy, Bootcamps

7.1 Coursera

7.2 edX

7.3 Udemy

7.4 Pluralsight / O'Reilly Learning

7.5 Bootcamps

Korea:

US:

Effectiveness:

7.6 ROI analysis

Free path: CS50 + The Odin Project + OSS contributions (unbeatable value). Paid courses: 10–20 Udemy courses total (200–500 USD). Bootcamp: Korea free–10M KRW, US 10K–20K USD.

Choice criteria:

Chapter 8: AI as a Learning Partner

8.1 Wrong uses

8.2 Good uses

1) Socratic questioning:

Prompt: "I am learning Rust ownership. Quiz me.
If I answer correctly, ask a harder question next.
If I get it wrong, just explain."

2) Explanation requests:

"Explain what each line of this CSS is doing."
"Explain why this SQL query is slow given this EXPLAIN output."

3) Understanding checks:

"I will describe what I think I understand.
Point out what is wrong: [my explanation]"

4) Pattern comparison:

"Differences between Rust's Result and Go's error returns?"

5) Brainstorming:

"Give me 3 weaknesses of this architecture and 2 alternatives."

8.3 Effective prompt templates

Learning by teaching:

"I want to explain [X] to [Y]. Pick the 5 concepts I need to know,
and explain each with an analogy suited to [age/level]."

Deliberate Practice:

"In [topic], my weak areas are [A, B, C]. Give me 10 problems that
attack these weaknesses, in increasing difficulty. After each problem,
provide hint first, then answer."

8.4 Knowing AI's limits

Verify: cross-check against official docs, authoritative books, and the community.

8.5 AI in the IDE

Learning tip: turn copilot off, try on your own, then turn it on and compare. Avoid pure copy-paste.

Chapter 9: Learning in Your 40s and 50s

9.1 Age and learning

Research:

9.2 Shift in strategy

Quantity to quality:

Deeper T-shape:

Learn by teaching:

9.3 Health is the foundation

9.4 Real examples

9.5 Traps

Chapter 10: Learning Communities

10.1 Korean developer communities

10.2 Study groups

Online:

Offline:

Formats:

10.3 Global communities

10.4 Conferences

One to two per year is a good target.

Chapter 11: Lifelong Learning Routine

11.1 Daily

11.2 Weekly

11.3 Monthly

11.4 Quarterly

11.5 Yearly

Chapter 12: 12-Item Learning Checklist

Chapter 13: Ten Learning Anti-patterns

1) Tutorial hell

Endlessly replaying beginner tutorials, never starting a real project. Real projects are real learning.

2) Book hoarding

Buying dozens of tech books and finishing zero. Finishing one beats owning ten.

3) Tool-setup addiction

Two weeks polishing Obsidian templates. Zero actual learning. Tools 10 percent, learning 90 percent.

4) "Next year I will learn it"

Updating the list every year, never executing. Start 30 minutes today.

5) Multi-start

Starting Rust, Go, and Elixir at once. Dropping all three. Finish one, then the next.

6) YouTube-only

Passive lecture watching, no coding. Passive is worse than active.

7) "I will code once I fully understand"

Waiting for 100 percent comprehension. Never starts. Get to 50 percent, then practice.

8) Not sharing

Keeping learnings to yourself. No feedback. Share through blog/conversation.

9) Avoiding hard things

Sticking to easy material. No growth. Desirable Difficulty.

10) "Too old for that"

Self-limiting. The brain can still learn in its 60s. Age is an excuse.

Closing — Lifelong Learning Is an Engineer's Fate

Principle 1: Small, daily

30 minutes a day for 5 years beats 10 hours every weekend for a month.

Principle 2: Teach while learning

Blog, mentor, speak — the fastest growth path.

Principle 3: Community beats solo

A group beats going alone. Motivation, feedback, fun.

Principle 4: Health is the foundation

Sleep, exercise, nutrition. Without them the brain does not function.

Principle 5: Tools are means

Do not over-engineer your system. Simplicity is powerful.

Principle 6: Read the originals

Next up — "Developer Networking and Branding: A Practical Guide Even for Introverted Developers"

Season 3 Ep 12 will cover:

See you next time.

Read the original: https://www.youngju.dev/blog/culture/2026-04-15-developer-learning-complete-guide-new-language-domain-papers-lifelong-learning-ai-partner-deep-dive-guide-2025.en

Comments

No comments yet.

Sign in to leave a comment