I asked an AI about selling dead roaches—thought it was a terrible idea. It still found niche uses like gag gifts or science. I don’t recall the model, but it intrigued me. Now I’m trying to stump it with something so bad it finally says, “Whoa, that’s an awful idea!” 😂

I just went to their website to check and got this strange message? What the heck is going on?

Depending on the use case, I think it would vary the results that someone would experience. What do you think?

I'm building an AI history app. It's 60–75% accurate, but the inaccuracies make it unreliable. A new process boosts accuracy but raises costs—making it impractical. Maybe usefulness depends on having the right processes for the specific use case, and cost makes that difficult to implement.

I am a little familiar with Playwright. Never thought to use it to help with designing. I’ll definitely give that a shot. Thank you!

Yeah. Claude Code definitely seems to do better with coding the functionality, but not so great when it comes to designing. I'm thinking there must be a better way with all the ai advancements we've had. 🤔

Appreciate the advice. I'll give it a shot!

Hey! Thanks for this! I appreciate it!

Just checked out zen MCP. Very interesting! I’m not sure, but I want to say there’s a model on OpenRouter that does something similar and automatically chooses the best models for the tasks. I wasn’t really sure about it. Have you noticed a significant difference rather than just using one model?

Oh yeah. I had briefly come across WebStorm since it seems most of what I do is using React and JavaScript. I hear they make great products.

Just checked out Zed. Love the video where they're switching tabs at over 120fps. So unexpected. I've actually be looking for something more simplistic in design yet powerful. Haven't heard of Zed but it looks cool. I've currently been looking at Warp.