PinnedInLevel Up CodingbyJaroslaw Wasowski·Jun 13Stop Treating Agent Memory Like a Cache — It’s a Security LayerA production agent’s memory is an authorization layer, not a cache. Relevance is not permission — where to insert a deterministic gate…A response icon10A response icon10
PinnedInLevel Up CodingbyJaroslaw Wasowski·Jun 11Formal Verification in Spec-Driven Development — Enterprise LevelIn regulated enterprise, a specification becomes an executable contract. Why AI-generated code must be proven with formal methods, not…A response icon6A response icon6
PinnedInLevel Up CodingbyJaroslaw Wasowski·Jun 18The End of Coding Is Not the End of Developers — and That Distinction Decides 2030SDD and AI agents don’t end the developer profession — they shift its value from code to intent and verification. Here’s the competency map…A response icon15A response icon15
PinnedInLevel Up CodingbyJaroslaw Wasowski·May 11Stop Writing Specs. Start Writing Facts. The Entire SDD Movement Is Already Obsolete.One test survived Sonnet 3.5, 3.7, 4, and Opus 4.5+. The spec needed four reinterpretations.A response icon50A response icon50
InSystem WeaknessbyJaroslaw Wasowski·17h agoAI agent harness: abort, egress proxy, credential scoping, cache and trace-based monitoringThe lab incidents were not a rebellion but a harness with no legal exit, permissions in the prompt and no owner of weak signals
Jaroslaw Wasowski·21h agoAI Agents — AI agent self-assessment: deterministic gate, kill criteria and read-back in the…An agent reports its goal is at risk and the pipeline logs it. A gate with thresholds written before the run turns the report into action
InLevel Up CodingbyJaroslaw Wasowski·1d agoSpec-Driven Development — A pull request template for AI agents: an evidence pack and spec review…An evidence pack for agent changes: behavior, proof, risk, unknowns, rollback plan. Plus ten minutes on the spec before the codeA response icon1A response icon1
InCodeToDeploybyJaroslaw Wasowski·2d agoDeep Research at Work: Compare Four AI Tools’ Reports and Verify DisagreementsFour tools, one research prompt: the value comes from triangulation and verifying where the reports disagree.
InSystem WeaknessbyJaroslaw Wasowski·2d agoGen AI at work — Decision Brief Instead of a Summary: How to Brief AI on Documents Before a Meeting…A summary keeps the document’s proportions; a decision brief keeps the decision’s. Five-field template, prompt, 10-minute check.
InLevel Up CodingbyJaroslaw Wasowski·3d agoEnterprise AI Agents — AI Agent Reflection Loop: Holdout, Veto Over Average, and an LLM Judge…A shared reflection layer fixes agent silos, but only if the loop can’t write the tests, rubric or threshold that grade itA response icon2A response icon2
