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
Why npm supply-chain attacks won't go away — what npm-scan actually checks, and the defenses that work
npm supply-chain attacks are not isolated accidents but a recurring product of install scripts, transitive dependencies, and typosquatting. In 2025 the Shai-Hulud worm self-replicated through install scripts and harveste
2026-07-11 · 7 min read #npm#supply-chain#security#javascript#nodejsLLM Burnout: When the Job Becomes Reviewing the Work Instead of Doing It
A developer's essay, 'I Think I Have LLM Burnout,' trended because it named a quiet shift: the workday is turning from writing code into reviewing what a model wrote. This post represents that argument faithfully and add
2026-07-11 · 6 min read #llm#developer-experience#burnout#ai-tooling#code-reviewPlayStation's 3-year inactivity clause: what the terms actually say, and the digital-ownership question
In July 2026 a headline made the rounds: PlayStation can delete all your digital games after 3 years of inactivity, under its EU terms. Read Sony's UK and European terms directly and it is narrower than that: not a targe
2026-07-11 · 7 min read #digital-ownership#playstation#drm#consumer-rights#licensingWhy K-pop Looks Bigger Than a 7th-Place Market: Reading the Music Economy Through IFPI 2026
IFPI's Global Music Report 2026 puts 2025 global recorded-music revenue at 31.7 billion USD, up 6.4% year over year — the eleventh straight year of growth and the first time the market cleared 30 billion. In that table,
2026-07-11 · 5 min read #kpop#music-industry#business#korea#streamingQuantizing the KV Cache to 4 Bits — SAW-INT4 and System-Aware INT4
In long-context LLM serving, the real memory bottleneck is not the weights but the KV cache. Quantize it naively to INT4 and accuracy collapses; the April 2026 preprint SAW-INT4 pairs token-wise INT4 quantization with a
2026-07-11 · 6 min read #ai#llm#quantization#inference#kv-cacheReading Microsoft Flint — a Visualization Language for Agents to Draw Charts, Not to Be Drawn
Microsoft Research's Flint is not a language for visualizing agents — it is an intermediate language that lets AI agents reliably produce good-looking charts from data. A compiler derives the low-level decisions (scales,
2026-07-11 · 5 min read #ai#agents#data-visualization#llm#microsoftNo Leap Second in 2026 — and the Negative One No System Has Ever Run
On 6 July 2026 the IERS issued Bulletin C 72: no leap second at the end of December 2026. It reads like quiet news, but since Earth started spinning faster around 2016, the first-ever negative leap second is now on the t
2026-07-11 · 6 min read #leap-second#utc#ntp#distributed-systems#posix-timeOptimizing Long-Context Inference with Hybrid SWA — What Xiaomi's MiMo v2.5 Actually Does
Hybrid SWA interleaves sliding-window attention with full attention across layers to cut both KV-cache memory and compute for long-context inference — now a mainstream design. Grounded in Xiaomi's published MiMo v2.5 inf
2026-07-11 · 7 min read #ai#llm#inference#attention#optimizationWhy the Mac mini Became an On-Device AI Machine — What Apple's Silicon Exec Said, and What He Left Out
Apple Silicon senior product manager Doug Brooks talked to The Deep View about demand for the Mac mini and Mac Studio and where on-device AI is heading. Why developers and small teams reach for this little desktop as a l
2026-07-11 · 6 min read #apple-silicon#on-device-ai#local-llm#mac-mini#inferenceGhostLock: a stack use-after-free that hid in Linux rtmutex for 15 years (CVE-2026-43499)
GhostLock (CVE-2026-43499) is a stack use-after-free in the Linux kernel's real-time mutex code that, per the nebusec.ai IonStack part II writeup, sat in mainline from 2.6.39-rc1 in 2011 until a fix in April 2026 — about
2026-07-11 · 6 min read #linux-kernel#security#use-after-free#privilege-escalation#rtmutexMel Brooks at 100 and the Documentary of Record: The 2026 Wave of Reverent Portraits of Elder Entertainers
Timed to Mel Brooks turning 100, a two-part HBO documentary about his life led the 2026 Emmy nonfiction field with six nominations. The same season also honored reverent, career-spanning portraits of John Candy, Martin S
2026-07-11 · 5 min read #film#documentary#comedy#culture#legacyThe founder who went back to the terminal: Hashimoto on Ghostty, Zig, and open source with "no obligation"
Grounded in a recent interview, a look at why HashiCorp cofounder Mitchell Hashimoto stepped back from being a successful founder to build the terminal emulator Ghostty in Zig, a language not yet at 1.0. He admits that a
2026-07-11 · 6 min read #ghostty#zig#open-source#terminal#mitchell-hashimotoJordan Peterson's Major Books: A Thoughtful Reader's Guide
Jordan Peterson is a clinical psychologist and a genuinely polarizing public figure. This guide sets the political controversy aside and focuses on the actual content of his three major books — 12 Rules for Life (2018),
2026-07-11 · 6 min read #books#self-improvement#psychology#jordan-peterson#philosophyIn Emacs, everything looks like a service: reading the client-server lens onto the "Emacs is an OS" meme
A reflection on the essay "In Emacs, Everything Looks Like a Service" by Charles Choi. He flatly rejects the tired "Emacs is an operating system" comparison and offers a sharper picture: Emacs is a universal client, and
2026-07-11 · 6 min read #emacs#elisp#client-server#developer-tools#text-editorsWhy KubeVirt GPU Passthrough VMs Could Not Be Scheduled for 112 Days — A Real Cluster Postmortem
After a Rust operator delivered the diagnosis that "all 4 GPU nodes are NotReady," I traced the cause of that death all the way to the end on a real 8-node cluster (GPU Operator v25.3.0, KubeVirt v1.7.0). The real reason
2026-07-11 · 5 min read #kubevirt#gpu#kubernetes#nvidia#devopsGood tools are invisible: gingerBill on friction as a feature, and where "invisible" hides failure
A reflection on gingerBill's essay 'Good Tools Are Invisible,' which argues that a good tool disappears into the background and that reselling friction as a fun puzzle is a rationalization. I agree with his feeling-versu
2026-07-11 · 7 min read #tools#developer-experience#kubernetes#build-systems#editorsCan the K-Pop 'System' Be Decoupled From Korea? KATSEYE as a Case Study
KATSEYE is a six-member girl group based in Los Angeles, built by HYBE and Geffen Records through the same trainee pipeline that produces Korean idol groups — yet its members say plainly that they are a global group, not
2026-07-11 · 6 min read #kpop#katseye#music-industry#culture#globalizationCoding Benchmarks Are Misaligned with the Agent Era — Three Reasons Leaderboards Compare Agents Wrong
A position paper from a Tessl team, posted to arXiv in June 2026, argues that today's coding benchmarks are fundamentally misaligned with agentic software engineering. Benchmarks were built to measure a single model, yet
2026-07-11 · 5 min read #ai#agents#evaluation#software-engineering#coding-benchmarksUpcycling Pretrained Checkpoints into Long-Context Hybrids — What HyLo Proposes
Hybrid sequence models (attention plus linear/SSM blocks) suit long context, but until now most had to be pretrained from scratch. The April 2026 preprint HyLo proposes a recipe to 'upcycle' an already-trained Transforme
2026-07-11 · 5 min read #ai#llm#long-context#architecture#efficiencyDiffusion LLMs That Write CUDA Kernels — DICE and Why Parallel Generation Might Help
DICE, a February 2026 preprint, claims that diffusion large language models beat autoregressive models of the same size at generating CUDA kernels, setting a new state of the art. The core idea: instead of writing one to
2026-07-11 · 6 min read #ai#llm#diffusion#cuda#gpu