Yeah. I should have used these definitions for clarity:

empty? → truthy/falsey check
sort! → danger ahead (mutates, raises, etc.)

I guess truthy works given the intent is a boolean-like conditional check. I prefer coercing to boolean for predicate methods, but it's not a hill I'd die on.

Then you need to keep following up to ensure that the analysis is correct. I use follow up prompts like, "Are you sure? Perform more research and verify that your last response was accurate and comprehensive." It takes a few iterations to dial it in.

Yeah. I've had to fork the conversations several times due to weird AI behavior like that. It's also good to check against multiple LLMs and with different web search tools.

It takes iteration. Start with a product idea & have AI perform competitor analysis to establish an initial pricing model. Then pick ad platforms that you intend to use, enter budgets, & calculate metrics like CPM, CPC, CR, CAC, LTV, etc. You'll learn quick whether or not it's feasible.

How about defining and removing constants and changing behavior of other classes/modules via patching? Still technically message passing but the side effects are bigger. Another possibility is throw/catch.

Have you looked into Tebako? www.tebako.org

I'm curious. Did you get packaging with traveling ruby working? @bradgessler.com

Terminalwire looks really slick. Looking forward to giving it a spin.

Hoping to fold that agent stuff into a gem. Only need more time in the day. 😜

github.com/hopsoft/loca...