matt swanson 😈

🎯 Building Karl Pilkington is my spirit animal
🫎 #1 ranked Canadian Football DFS

[matt swanson 😈 @_swanson]

@DanielLockyer you hate to see it

[matt swanson 😈 @_swanson]

@michiels its almost entirely from background API syncing and webhooks, on the order of 10M jobs/day

[matt swanson 😈 @_swanson]

Fix your egress with this one simple trick: dont load full ActiveRecord models that have huge json columns just to get two id columns out and then run code a few million times a day

[matt swanson 😈 @_swanson]

@dusangran planetscale insights did all the work: found the queries running the most with the biggest payload, then agent traces back to the job

[matt swanson 😈 @_swanson]

@mscccc can we get an MCP tool to dismiss recommendations with a reason please? agent is trying to do stuff like

[matt swanson 😈 @_swanson]

@hollylawly when redis

[matt swanson 😈 @_swanson]

I havent been raving about PlanetScale lately because honestly I havent had to think about our database in a while. Great product, great MCP, highly recommend.

[Mike Coutermarsh @mscccc]

Bring us your Rails apps. I love rails apps.

[matt swanson 😈 @_swanson]

It feels like there are actually three distinct levels of "run AI 24/7 to look for security issues"

1. Loop to check for published security patches in languages, libraries, platforms you need to upgrade

2. Loop to check your source code for app-specific security issues (pen-test, OWASP type stuff)

3. Loop to red-team your running app (combines 1 and 2) to cover chained issues or production config

[matt swanson 😈 @_swanson]

@IanLandsman how is your survivor pool doing

[matt swanson 😈 @_swanson]

@mscccc the weirdest papercut I run into all the time is the list_databases tool being in "Organization tools" and not "Database tools" on the auth screen -- I was skipping all org tools and then agent is like "I dont know the database name"

[matt swanson 😈 @_swanson]

@sqs it's very good, but finding the notifications annoying when i am also at my computer...i know its a hard problem (insert that Slack notification flow chart) but will probably be forced to turn off the android notifs

[matt swanson 😈 @_swanson]

We recently rewrote all the READMEs for our repos to:

- remove stuff about setting up environments (moved to .agent/setup and targeting a machine reader not human)
-add info about production: where it runs, what MCPs can do, project ids in external system, Slack channels to use

[matt swanson 😈 @_swanson]

This week I added a little feature to delete old export files that we never cleaned up (as they had reached 100GB total)

Built the feature then told Amp to report in on the backlog progress in Slack: "post to #talk-code with a summary of what was purged in a table"

[matt swanson 😈 @_swanson]

@danlovesproofs switched to this, plus having dependabot do a weekly PR to update them

[matt swanson 😈 @_swanson]

@butcher__51 what are you even talking about lol

>It could have been an simple explain select in production

Yeah, I know explain select and the agent knows it too, it tried but didnt have access so it made a local scratch db to do the next best thing?

[matt swanson 😈 @_swanson]

"but the agents dont understand the characteristics of our production system"

"oh"

[matt swanson 😈 @_swanson]

@butcher__51

[matt swanson 😈 @_swanson]

@searls Gotta be April Fools day for me

[matt swanson 😈 @_swanson]

@IanLandsman Kick push kick push...

[matt swanson 😈 @_swanson]

@dwestendorf @nateberkopec @danlovesproofs Yeah I guess if you have 1000 openAI keys and 1 more in encrypted creds you might just skip it