Bash quoting: how one filename became two arguments
The file is there, but ls says it is missing. Follow what Bash does to spaces, patterns and variables before the command receives them.
Articles
Guides, news and explanations about AI and reliable systems.
Not sure where to begin? Choose a short reading path, or browse at your own pace.
A fast “still pending” reply does not make the export fast
Follow job-42’s four-minute queue wait separately from its eighty-millisecond status reply, and keep the missing download check in view.
Map the report’s journey beyond the quick API reply
A job number arrives quickly; the report does not. Follow each handoff, its owner and the check that distinguishes a stored file from a usable result.
Agents and tool calling: how an assistant actually looks something up
A question about service health needs a lookup, not a convincing guess. Follow the request into the application, through its checks, and back with a result.
AI history: from writing the rules to learning the task
Rules, examples, prompts and reasoning changed the work we ask computers to do. The older choices still have jobs worth doing.
Embeddings: finding the document when you’ve forgotten its words
You remember the job. Someone else chose the title. Embeddings can help a search bring the two together, even when the words don’t match.
LoRA fine-tuning: teaching the model to give the label you need
A support queue needs one label, but the model keeps adding a paragraph. Follow a small LoRA change through training, saving and the reports that decide whether to keep it.
Linux history: why the same kernel can feel like a different system
A missing command and a familiar kernel can coexist. Unix, GNU, Linux and distributions explain how the pieces came together, and where to look when one is missing.
OpenCode 2: follow a small repair from failing test to changed file
Keep the conversation beside the project, set explicit approval boundaries and follow one retry bug through explanation, edit and unchanged tests.
RAG: giving an AI assistant something useful to read
An assistant can read the relevant documents before it answers. Follow one small exception from search to the reply, and see where its meaning can get lost.
Tokens and context windows: what stays in the conversation
A chat can look intact while the next request contains only part of it. Tokens explain how much fits; context explains what the answer can use.
A practical Gemini 4 Argon trial: follow the migration through its callers
A green test suite can miss the caller that still breaks. Give Argon a bounded migration, preserve its requirements and compare the full cost of finishing.