@KoalaQuillHQ Well that's where the AI agent's intelligence comes in. If you set up access rules well, this won't happen.
So specifically for agents accessing servers or cloud environments, you create users and roles for those agents and limit their actions.
Every business has a “where.” So why do we still understand so much of our data through spreadsheets?
I wrote about why every business will eventually need a map - and made an example with https://t.co/1xxKHSfZ2I.
https://t.co/Ngby58qxTQ · https://t.co/0KgYSySBIM
@meijer_s Tailwind actually made frontend work way more fun for me - I get to focus on the product and engineering part of the project, instead of spending time naming classes, organizing CSS files, or jumping back and forth between HTML and stylesheets.
@robj3d3 Absolutely agree. Reviewing PRs made by engineers extensively using AI is painful. Lots of meaningless code, lack of greater picture of the software...
@windsurf_ai - this is something you could level up when generating tests:
• Simulate actual usage
• Validate failure paths
• Catch regressions
We don’t just want coverage — we want confidence.
When your AI pair programmer writes tests that confirm water is wet 💧
"Let's assert that the thing we just hardcoded... is the thing we just hardcoded."
Thanks, robot. Truly groundbreaking.
And then someone drops this into a PR like:
"Added unit test coverage 🚀"
Meanwhile the test:
✅ Confirms structs still hold values
✅ Assures hardcoded strings remain hardcoded
10/10 coverage, 0/10 value.
🧠 Please, don’t create interfaces prematurely in Go.
Large and bloated interfaces not only add complexity—they leak implementation details and lead to fragile tests.
Build interfaces when needed, not "just cuz".
#golang#SoftwareEngineering
@NoahAttwood nice work @NoahAttwood but having a compile time specified expression is the same as having a single function receiving an object and checking it's data members in an if condition, right? I guess there is no need to build the whole tree then