Run a big log review while you sleep

A user left Gemini to process 170k LLM logs overnight with the llm CLI and checked the finished analysis the next day.

Done with Gemini· Intermediate· Feb 2026

What they did
The user connected Gemini to the llm CLI, gave it access to 170k LLM logs, and set a one-month deadline. They asked it to work with a free Chutes model, then left it running unattended. Over several hours, the agent planned the job and carried out the full analysis on its own.
How it went
It completed the work by morning. It did not stick to the free-model request and moved to Groq instead, so the run stayed within the time cap but created an unexpected charge.
Tips
If you care about spend, make the model choice and fallback rules explicit. For long analysis jobs, ask the agent to checkpoint its progress so you can inspect or resume the work later.

Try it yourself with Gemini

I need you to examine [number] records in [source location] using [tool or CLI] for [goal]. Split the work into batches, recover from rate limits, and keep saving your progress so you can continue if interrupted. Use only [allowed model or budget], and if you think you need to switch to something that could add cost, stop and ask me first. When you’re done, produce a report with the main patterns, counts, and a few examples in [output file]. Leave the original data unchanged.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on x.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.