Does anyone else feel like AI is slowing them down when working with "real" systems? It's just so hard to trust.
Does anyone else feel like AI is slowing them down when working with "real" systems? It's just so hard to trust.
`VACUUM FULL` locks all read/writes to the table while it cleans up dead tuples. Which can take hours on the 200GiB table we have. On a production database, we would have seen an outage immediately, which we probably would have missed in Staging.
Copilot generated a rails migration to add the index (OK not so bad), but in the migration, it added a SQL statement `VACUUM FULL` to reduce the TOAST bloat
Do you know what this does?
Using copilot to understand why a database table was growing faster than expected. It told us that not having a GIN index on the table was the "real root cause". because our objects were stored as JSONB and searching was causing TOAST to bloat.
WRONG - but it gets worse:
MailPace.com inbound + webhooks
Failing that, fly.io with something like github.com/nodemailer/s...
Registering a new domain this morning, @porkbun.com gave me a chuckle
Thought you all should know that MailPace has pivoted and is now an AI startup mailpace.com/blog/changel...
Itβs supposed to be 4 hours a week π
It was great until someone asked me why I ran a toilet paper company π€¦
π