◉GEMINI LABJP
●3.8LIVE — Two audio-to-audio models reached GA for the Live API: gemini-3.8-live as the low-latency default, and -extended-thinking for background reasoning mid-conversation●09/30 — Thirteen days until gemini-omni-flash-preview shuts down. The replacement is gemini-omni-1.1-flash, so count your call sites before you swap●MCP — A timeout written as ten minutes on the extension side is reportedly cutting out at one. Anyone handing long work to an MCP server runs straight into it●NEW — When an AI function in Sheets refuses to generate, suspect where the file lives before you blame the 24-hour cap●GEM — When a Gem built for your team will not share, walk the admin settings and the Drive sharing settings in a set order●10/16 — Twenty-nine days until Gemini 2.5 Pro, Flash and Flash-Lite shut down together. The path forward is the 3.5 Flash line●3.8LIVE — Two audio-to-audio models reached GA for the Live API: gemini-3.8-live as the low-latency default, and -extended-thinking for background reasoning mid-conversation●09/30 — Thirteen days until gemini-omni-flash-preview shuts down. The replacement is gemini-omni-1.1-flash, so count your call sites before you swap●MCP — A timeout written as ten minutes on the extension side is reportedly cutting out at one. Anyone handing long work to an MCP server runs straight into it●NEW — When an AI function in Sheets refuses to generate, suspect where the file lives before you blame the 24-hour cap●GEM — When a Gem built for your team will not share, walk the admin settings and the Drive sharing settings in a set order●10/16 — Twenty-nine days until Gemini 2.5 Pro, Flash and Flash-Lite shut down together. The path forward is the 3.5 Flash line
TAG

agent design

3 articles
← Back to all tags
Related:
Gemini API3Function Calling1Timeouts1Idempotency1Memory Bank1Memory profiles1schema design1indie development1Managed Agents1environment hooks1sandbox1
◈ Gemini API/2026-08-08Advanced

A Timeout Was Never Evidence of Failure — Designing Around Blocking Function Calls

When a tool call cannot return until the real-world effect finishes, two habits reverse on you: parallel dispatch and generous timeouts. Measured numbers from a sandbox harness, plus the design that replaces retries with observation.

◈ Gemini API/2026-08-03Advanced

Measuring Update Policies for Memory Profiles: The Guard That Cost Me 16 Points of Accuracy

Memory profiles went GA in Memory Bank, making structured memory available to downstream code. I built three update policies and compared them under identical conditions. The one that looked obviously correct turned out to be the worst. Full harness code and the path to per-field TTLs.

◈ Gemini API/2026-08-02Advanced

Measuring a Guard in environment hooks: 46 Microseconds to Decide, 23 Milliseconds to Start

A record of building a destructive-command guard for Managed Agents environment hooks. A regex denylist let 9 of 20 dangerous commands through; argv parsing reached 100 percent detection. The startup cost turned out to be 500 times the decision cost.