To prevent my custom AI chat assistant 'Kana' from repeating previous shopping inquiries in casual conversation, I added clickable research suggestion chips, backed by hourly Twilog status ingestion and 24-hour expiration limits.
I added small-talk updates and paired Codex image generation to my Pi and Qwen assistant, while debugging why Backlog tasks stalled at research and fixing container permission blocks.
Providing local Qwen3-Embedding-0.6B examples and regex hints boosted Qwen's chat intent classification from 57.5% to 90.0% on 40 synthetic test cases, and a secondary LLM verifier reached 95.0%. Here are the failure modes and gateway constraints.
Building a gateway to route tasks across multiple PCs to Codex, Claude Code, and Pi. How we manage persistent memory, task state, model verification, and external AI boundaries.
We evaluated MiniCPM-V 4.7 35B-A3B on a 16GB M4 Mac mini across coding, reasoning, and table OCR benchmarks, finding 100% transcription accuracy alongside arithmetic hallucination patterns.
Resolving api_key_model_discovery incompatibility errors in Codex. This article covers the version divergence mechanism between the CLI and daemon, compatibility check updates in 0.162.1, local logs, and daemon update commands.
Codex stopped making progress on external operations while continuing to respond. A session history audit showed permission policy shifts, MCP rejection under approval_policy never, and recovery via session handoff.
Tested on M4 Mac mini: Gemini 4 Argon availability in AGY, pricing comparison, and CLI benchmark runs between GPT-6.1 Sol, Claude Opus 5.5, and Gemini 3.8 Flash.
AWS Strands harness benchmarked against Claude Code and Codex on identical models using official data. Covers cost gaps, model swapping, and unprompted shell execution risks.
Tested on M5Stack CoreS3: routed Claude Code, Codex, and Antigravity completion hooks to StackChan via Tailscale. Fixed audio stutter by raising I2S speaker buffer to 1024.
Parsed Codex CLI and Claude Code session logs on an M4 Mac mini to track subagent state, time, and tokens. In a 6.5-hour session, the parent waited 68% of the time on children; across 30 days, Codex subagents spent only 5% of active time on tools and 95% waiting on model responses.
OpenAI's Navier–Stokes claim drew questions about credit and whether Buckmaster's Codex prompts influenced it; its September 10 update says prompts in the two months before the September 8 announcement could not affect it.