Blog
Notes from building Stoicsoft. Tools, processes, and the occasional rant about software that wastes your time.
Showing 1–10 of 59 · Page 1 of 6
ai-workflow· Jun 26, 2026· 8 min readSafety Trails: What AI Coding Agents Need Before They Touch Real Files
Before an AI coding agent touches real files, the workflow needs rollback checkpoints, gated execution, durable evidence, cached context, and a faster loop that still leaves a trail.
ai-workflow· Jun 26, 2026· 7 min readWrite-Only Memory: Why Saved Context Needs Retrieval and a Review Loop
Saving is not remembering. A retrieval layer and a weekly review loop, not a bigger pile of saved links, is what turns a write-only vault into usable AI memory.
developer-tools· Jun 26, 2026· 8 min readThe File Jobs You Don't Want in the Cloud: Confidential OCR, Signing, and Legacy Rescue
Merging a PDF, OCRing a scan, converting media, resizing a banner, rescuing an old archive — the work is local and the files are sensitive. The only thing that ever needed an upload was convenience.
ai-workflow· Jun 26, 2026· 6 min readRunning Four Coding Agents at Once: The Control Plane for Parallel AI Development
Run three or four coding agents at once and the bottleneck stops being the model and becomes the control plane: agent state, approvals, API truth checks, session persistence, and cost-aware routing.
- ai-workflow· Jun 26, 2026· 8 min read
Agents Write Faster Than They Verify: Put Real App Testing in the Loop
AI agents generate faster than they verify. Across six developer threads, the missing layer is the same: proof the code actually ran, against the right context, with a record of what was tested.
ai-workflow· Jun 26, 2026· 7 min readWhen Code Outruns the Spec: AI Coding's Operational Debt
AI can write code faster than you can read it. The new bottleneck is keeping sessions, spend, and specs in sync with what the agents actually shipped — the operational debt of spec drift.
ai-workflow· Jun 26, 2026· 8 min readPortable AI Memory: Exporting Chats, Switching Accounts, and Preferences That Follow You
The AI memory problem is really a portability problem. Five real situations show why durable context must be exportable, account-independent, and continuous across tools.
ai-workflow· Jun 26, 2026· 7 min readReusable Project Memory Beats Chat History
The value in AI work is shifting from one-off chat transcripts to private, reusable project context. Five real scenarios show why durable project memory beats chat history.
ai-workflow· Jun 26, 2026· 7 min readYour Open Tabs Are a Memory System: Capturing Why Context Mattered
The tabs you keep open for days are placeholders for intent your tools forget. A look at why knowledge workers hand-build AI memory systems, and what a real one should capture.
- developer-tools· Jun 26, 2026· 7 min read
Run the Tool Where the File Lives: No-Upload, Composable File Actions
Single-purpose web tools wrap simple file tasks in accounts, uploads, and watermarks. The case for composable, local-first file utilities that run where the file already lives.