cgarvis

I've seen some newer testing frameworks where you prompt for the outcome you want under a specific scenario. The first run is a LLM agent that figures out the steps but those steps are then used for recurring runs. If something fails, it goes back to the agent loop to see if it's an actual error.

This works well in e2e type tests where the test implementation might change but the customer experience hasn't.

So maybe there is a hybrid approach in our future.

This is the bet I'm making with Croft. LLMs to figure out the messy bits but once it's figured it, software is the better tool. Let the LLM write the software.

This is what software engineers do but just more expensive and time consuming.

MCP is for when your end user (the one driving an LLM Agent) is non technical. Most non technical people are not going to install a CLI on their computer. Most are not going to be driving the LLM agent via a terminal.

It also very helpful when you need auth. MCP OAuth with CIMD makes it easy instead of cumbersome process of generating API Keys.

If you are technical and already using CLIs, then MCP doesn't give you much.

Fly.io brings back some of that easy of deploy.

never knew you could do css selectors on custom elements! does seem like syntactical sugar over classes. I do like the current trend with tailwind: style with markup using smaller components. But maybe there is something here with so many components being just styles...

Are you doing got worktrees in the backend?

Dish outsources sales calls to a third party. This third party doesn’t have access to customer accounts so they have to send you to an actual Dish call center. Not as nefarious as you think but still frustrating.