Essays and bookmarks on technology, AI, and software engineering
Bookmarked on August 04, 2026 A benchmark for whether LLM applications honor instruction priority under realistic conflicts.
Bookmarked on August 04, 2026 A careful look at a dangerous gap between what an MCP approval surface shows and what reaches the model.
Bookmarked on August 04, 2026 A call for controlled memory benchmarks that do not confuse method gains with pipeline changes.
Bookmarked on August 04, 2026 A focused treatment of stale facts: retrieval memory needs a model of time, not just similarity.
Bookmarked on August 04, 2026 A timely comparison of the cost and grounding trade-offs between retrieval and long-context systems.
Bookmarked on August 04, 2026 A domain-specific example of assembling evidence across text, knowledge bases, and structured data.
Bookmarked on August 04, 2026 A benchmark for the subtle but costly problem of choosing the wrong skill among near-equivalent options.
Bookmarked on August 04, 2026 An argument for retrieving the possible interactions with a corpus, not only its documents.
Bookmarked on August 04, 2026 A sharper way to separate real retrieval gains from entity overlap and benchmark leakage.
Bookmarked on August 04, 2026 A provocative replay-first approach to making repetitive agent work cheaper and more predictable.
Bookmarked on August 04, 2026 A study of how an agent harness changes the value of search and retrieval choices.
Bookmarked on August 04, 2026 A reminder that memory evaluation should measure precision, not reward dumping the whole store.
Bookmarked on August 04, 2026 A concrete case for moving recurring LLM workflow work into deterministic compiled artifacts.
Bookmarked on August 04, 2026 A useful wide-angle comparison of retrieval choices for mixed text-and-table RAG.
Bookmarked on May 07, 2026 Search is not a side operation in the agent loop; it is one of the main things the agent does.
I think Codex’s `/goal` is underrated. Easy one-task ralph loops for everyone.
Bookmarked on May 05, 2026 "The fact that it was cheap for you to produce does not make it cheap for someone else to receive."
Bookmarked on May 05, 2026 "A farmer in Loudoun County sells a few acres of mediocre hay field to a hyperscaler for ten times its agricultural value."
I asked Alan to look at Simcluster and think about how he would try to win it.
Bookmarked on April 22, 2026 OpenAI's first claw-like thing since hiring the OpenClaw founder, with UX tight enough for non-nerds