Blog
Writing on GPUs, LLMs, MLOps, Kubernetes — and mindset · 3517 posts
#2026-03 765#english 592#culture 264#deep-dive 254#kubernetes 247#career 229#ai 216#llm 208#devops 193#2026-04 146#security 141#database 114#observability 113#communication 109#history 107#architecture 100#productivity 96#finance 88#economy 84#mindset 81#psychology 80#ai-papers 79#food 78#it 78#travel 78#deep-learning 77#japanese 77#networking 77#performance 72#business-travel 70#linux 70#gpu 69#ai-agent 66#cs-fundamentals 63#postgresql 60#rag 58#self-improvement 55#learning 53#mlops 53#ai-platform 51
What Studying Well and Playing Sport Well Share Is Narrower Than It Looks
Is there really an ability that carries over between studying and sport, between coding and playing an instrument? The conclusion of the cognitive-training transfer literature runs against popular expectation. Near trans
2026-08-16 · 13 min read #mindset#learning#expertise#transfer#evidenceLunch, Hoesik, and Hallway Conversation — What Flows Through the Informal Channel
An organization runs two information networks. The official channel carries what has been decided; the informal one carries what has not been decided yet. This post admits honestly that not participating has a cost, and
2026-08-16 · 14 min read #career#workplace#korea#communication#boundariesManaging Up Is Not Flattery — It Is Designing the Flow of Information
Most trouble with a boss is not a personality clash but information asymmetry. Your manager sees wide and shallow while you see narrow and deep, so the kind of thing each side misses is the same every time. This post cov
2026-08-16 · 13 min read #career#workplace#managing-up#communication#managerThe Order of Operations After a Mistake — Contain, Report, Record, Prevent
The cost of a mistake is set not by the mistake itself but by the four stages that follow it, and people invert that order in almost exactly the same way every time: hunting for the cause instead of stopping the bleeding
2026-08-16 · 13 min read #career#workplace#mistakes#incident#accountabilityBad News Is Judged by Timing More Than by Content
A good response delivered late is judged worse than a bad response delivered early, because bad news is evaluated on two axes: the size of the problem, and the number of options the listener has left. The size of the pro
2026-08-16 · 13 min read #career#workplace#communication#escalation#reportingThe Structure of Saying No — Work Piles Up by Routing, Not by Rating
When work piles up on you, that is not praise. It means the cheapest path for someone to ask ran through you. This post first explains the structure that turns the reward for being capable into a penalty, then shows that
2026-08-16 · 14 min read #career#workplace#boundaries#prioritization#communicationThe First 90 Days After a Move — What Gets Decided Before Performance Does
What actually gets decided in your first quarter at a new organization is not performance but classification. People sort a newcomer into one of a few boxes quickly, and the box is hard to change afterward, because the i
2026-08-16 · 15 min read #career#workplace#onboarding#transition#internal-moveProcesses and Signals, the Complete Guide: Tracing Why a Process Will Not Exit
Signal numbers and default actions, the difference between SIGTERM and SIGKILL, zombies and orphans, process groups and sessions, nohup and setsid, and how to design a graceful shutdown. Traces the cause of a process tha
2026-08-15 · 18 min read #리눅스#프로세스#시그널#운영#커널A Complete Guide to Linux Performance Tools: How to Read the Numbers in the Output
Interprets the numbers printed by top, vmstat, iostat, pidstat, sar and perf one column at a time. Covers which tool answers which question, which values create optical illusions, and the order in which to choose a tool
2026-08-15 · 18 min read #리눅스#성능#모니터링#운영#sysstatA Complete Guide to File Descriptors and Inodes: What Happens When df and du Disagree
Starts from the relationship between inodes and directory entries and works through hard links, deleted-but-still-open files, inode exhaustion, and file descriptor limits. Covers why df and du diverge and how to recover,
2026-08-15 · 16 min read #리눅스#파일시스템#inode#파일디스크립터#운영The complete guide to SSH operations: from key management to hardening a server without locking yourself out
sshdconfig hardening, choosing and distributing key types, ProxyJump and port forwarding, and the layered order to debug a connection that will not come up, organised from an operator viewpoint. Includes the procedure th
2026-08-15 · 17 min read #리눅스#ssh#보안#운영#서버관리The complete Linux troubleshooting command guide: from the first 60 seconds to root cause
The diagnostic commands to run, in order, from the moment someone tells you a server is misbehaving. uptime, top, free, df, iostat, ss, journalctl — which number in each output to read, and what that number rules out ver
2026-08-15 · 19 min read #리눅스#트러블슈팅#운영#장애대응#명령어A complete guide to Linux firewalls and access control: handling nftables, firewalld, and ufw without locking yourself out
Starts from the structure of nftables and lays out what firewalld and ufw do on top of it. Covers the procedure that keeps you from locking yourself out while changing rules remotely, and the diagnostic order for when ru
2026-08-15 · 18 min read #리눅스#방화벽#nftables#firewalld#보안The complete backup and restore guide: designing backwards from the recovery scenario
Starting from RPO and RTO, the criteria for choosing between rsync, tar, LVM snapshots, and snapshot-based backup tools. Covers the recovery rehearsal procedure and the five ways a backup quietly rots.
2026-08-15 · 16 min read #리눅스#백업#복구#재해복구#운영The Complete Guide to TLS Certificates: Handling Them End to End with openssl
A purpose-by-purpose reference of the openssl commands for reading certificates, creating CSRs, verifying chains, converting formats, and diagnosing server connections. Also covers how to identify expiry, a missing chain
2026-08-15 · 18 min read #리눅스#tls#인증서#openssl#보안The complete guide to Linux log operations: tying journald, rsyslog, and logrotate into one system
Designing the log plumbing of a Linux host, from journal retention policy to forwarding into rsyslog to the rotation modes of logrotate. Sets out the criteria for keeping the records an incident investigation needs witho
2026-08-15 · 16 min read #리눅스#로그#journald#rsyslog#logrotateUndoing Things in Git, the Complete Guide: What Comes Back and What Is Gone Forever
Maps the boundary between reset, revert, restore, reflog, and fsck by what can actually be recovered. Covers reflog expiry policy, when gc really deletes objects, force-push accidents, and worktree and submodule situatio
2026-08-15 · 17 min read #git#버전관리#복구#운영#협업The Complete Guide to DNS: Following a Name All the Way to an Address
Splits name resolution into five separate segments — stub resolver through authoritative server — and works through record types, TTL and caching, delegation, reading dig output, propagation delay and stale caches, DNSSE
2026-08-15 · 25 min read #리눅스#dns#네트워크#운영#트러블슈팅The Complete Guide to Transaction Isolation Levels: The Part the Application Owns, Not the Database
Lays out the four isolation levels of PostgreSQL 18 exactly as documented, then covers what the application must implement on top of that: the 40001 serialization-failure retry layer, the row-lock ladder from FOR UPDATE
2026-08-15 · 19 min read #database#postgresql#transaction#isolation-level#concurrencyThe Complete Guide to Partitioning and Sharding: The Order for Moving Past a Single Node
A look at what declarative partitioning in PostgreSQL 18 actually solves and does not solve, then the criteria for moving from partitioning to sharding once partitioning hits its limit. Covers partition pruning, the cons
2026-08-15 · 19 min read #database#postgresql#partitioning#sharding#scalability