loc bengaluru, ist | local --:-- srijanshukla18@gmail.com →
[category]/writing/ai

ai

Filtered public notes from the archive.

[01]

entries

22
Oct 2026 AI builder notes - October 2026 Notes from my October liked tweets and GitHub stars on Anthropic's Scion source control, capping agent code-review loops, PostHog's 38 AI product lessons, and streaming MoE weights over NVMe. → Sep 2026 AI builder notes - September 2026 Notes from my September liked tweets and GitHub stars on how Anthropic made claude.ai 3x faster with instruction counters, Spotify's hooks vs skills rule, and ultra-fast System One decision models. → Sep 2026 writing a model packer for Inco Splash Inco's Splash engine ships a package reader and no writer, so it only serves the models they publish. I wrote the writer to run an uncensored Qwen3.8-27B, proved it byte-identical to theirs, and then spent hours on a layout bug that passed every check I had. → Sep 2026 running durable agent sessions with OpenAI SDK, Docker, and Temporal From a survey of the agent-platform field through strict sandbox requirements to a durable agent session: one OpenAI SDK run in Docker, hardened with a Temporal workflow. → Sep 2026 three ways to put an LLM inside a Temporal workflow Three production runtime designs for LLM workflows, compared by who owns context, iteration, execution, validation, and safety. → Sep 2026 checked work graphs for multi-agent projects A dependency graph with versioned contracts and independently checked outputs can coordinate large projects across many workers, as demonstrated by Prove2Me's formalization of Fermat's Last Theorem. → Aug 2026 AI builder notes - August 2026 Notes from my August liked tweets and GitHub stars on Uber and Stripe's internal agent platforms, why >150 skills breaks routing, KV prefix cache economics, and new agent harnesses. → Aug 2026 training an AI detector on my own blog (and beating it) Two generic detectors failed on my content. So I built my own on my own ground truth: pre-2026 posts are human, 2026 posts are machine-written. It scored 44 for 44. Then I attacked it for an evening, drove its verdict on this very post down to 0.313, and packaged the attack as a skill. → Jul 2026 AI builder notes - July 2026 Notes from my July liked tweets and GitHub stars on Claude 5 context engineering, turning .har traces into CLIs, issue-repro agents, and real-world agent receipts. → Jun 2026 AI builder notes - June 14, 2026 Notes on agent loops, deterministic workflows, Claude Fable, OpenRouter Fusion, and Google Open Knowledge Format. → Jun 2026 AI builder notes - June 8, 2026 AI-assisted notes from my liked-tweets feed, organized around agent loops, cloud agent infrastructure, skill security, memory, and runtime context. → Jun 2026 notes on agent engineering Notes on agent engineering: what counts as an agent, why long tasks fail, harness design, eval loops, observability, and when teams build vs buy. → Jun 2026 notes from my talk on coding agent architecture The written version of my talk on agentic design. The model is one layer; the interesting system is the harness, context, memory, control loop, and review process around it. → May 2026 AI builder notes - May 2026 AI-assisted notes from my liked-tweets feed, organized around agent workflows, browser traces, model loops, and guardrails. → May 2026 tracking my time and calories over WhatsApp with Hermes One WhatsApp thread with Hermes Agent captures what I do every 15 minutes and what I eat, writes two JSON files, and this site renders two public readouts from them. Why it works, and why I only track two things. → Apr 2026 AI builder notes - April 2026 AI-assisted notes from my liked-tweets feed, organized around harnesses, managed agents, memory, workflow packaging, and enterprise AI work. → Apr 2026 high thinking budgets make agents harder to control High thinking budgets can reduce control, especially when stale context and autonomous tool use mix together. → Apr 2026 syncing custom skills across laptops (and stopping the agent from overengineering it) One skills repo, two laptops, five agent runtimes. I let an agent design the sync, and every unstated decision gave it room to optimize for control instead of maintainability. → Feb 2026 stopping parallel sub-agents from stepping on each other (TAO) A prompt rule for running parallel agents without conflicts: the coordinator owns the shared work, sub-agents only get isolated slices. → Jan 2021 vigil - a live file-diff watcher for coding agents Once I had LLMs editing code autonomously, I wanted to watch a directory breathe in real time. Existing tools gave me file events or diffs or a TUI, never all three, so I built vigil. → Apr 2019 splitting a multi-step cleanup prompt into gated turns in Pi I wanted one cleanup workflow after each feature and tried building a Pi extension for it. Packing all three steps into a single prompt failed quietly, so the extension now runs them as sequential gated steps. → May 2018 syncing two AI agents across machines with Git instead of a shared mount A shared mount looked like the cleanest way to make two agent instances talk across machines. A 30 minute directory cache made it unreliable, so Git became the bus instead. →