Thomas Broadley

This is probably still necessary, even given #55. It still takes 14 seconds to clone Bazel to one commit. Maybe just print the output from `git clone`?

Hey, thanks for filing an issue! First, `-e` doesn't take a comma-separated list of extensions, but a space-separated list (e.g. `-e ipynb txt md rst py`). However there is a...

Thank you for pointing this out, Daniel! Sorry for the annoyance. I do think it'd be a good idea for the tool to detect if the `fix-typos` branch already exists...

Sure, I would accept such a PR! Thank you for the suggestion.

Also, thank you for writing about this tool! That's cool to see.

Sigh. I think it's just slow on my computer because Hunspell is slow. Still something to improve.

Sami's talking about these generated fields on `trace_entries_t`: https://github.com/METR/vivaria/blob/196f603c2da77fa68574e65496f6dc69d61e8246/server/src/migrations/schema.sql#L159-L171 And the `runs_mv` materialized view:https://github.com/METR/vivaria/blob/196f603c2da77fa68574e65496f6dc69d61e8246/server/src/migrations/schema.sql#L660 Runs using the passthrough are runs with at least one generation trace entry that has the...

Yeah I think this would be OK. One way we could do this: - Move Driver and DriverImpl out of the `task-standard` subdirectory of the vivaria repo - Update links...

I even think clean branching wouldn't be too hard to support. As long as the uploaded agent's code still exists on the server (which it might not, because it gets...

I do wonder if it'd be better to upload the agents to some kind of more durable object storage service like S3. That might be overkill.