Everything should be Tailwind because all code is written by LLMs these days.
If you need a converter for a normal HTML/CSS component, my free extension SnipCSS is the best Tailwind converter. I haven't seen anything else come close.
For example making a computer use agent… Made the plan, implementation was good, now I want to add a new tool for the agent, but I want to discuss best way to implement this tool first.
Clearing context means Claude forgets everything about what was just built.
Asking to discuss this new tool in plan mode makes Claude rewrite entire spec for some reason.
As workaround, I tell Claude “looks good, delete the plan” before doing anything. I liked the old way where once you exit plan mode the plan is done, and next plan mode is new plan with existing context.
I get where you're coming from. But you'll likely get better results by starting fresh and letting it read key files or only just a summary of the project goals/spec. And then implement the next feature building up on the previous one. It's unlikely you'll need all the underlying code of the foundation in context to implement something that builds up on it - especially if interfaces are clean. Models still get dumber the more context is loaded, and the usable window isn't all that big, so starting fresh gives best results usually. I try to avoid compaction in any way possible, and I rarely continue the session after compaction, for that reason.
I've tried to make AI work but a lot of times the overall productivity gains I do get are so negligible that I wouldn't say it's been transformative for me. I think the fact that so many of us here on HN have such different experiences with AI goes to show that it is indeed not as transformative as we think it is (for the field at least). I'm not trying to invalidate your experience.
If you're being honest, I bet your codebase is going to shit and your skills are in rapid decline. I bet you have a ton of redundant code, broken patterns, shit tests, and your coworkers are noticing the slop and getting tired of fixing it.
Eventually, your code be such shit that Claude Code will struggle to even do basic CRUD because there are four redundant functions and it keeps editing the wrong ones. Your colleagues will go to edit your code, only to realize that it's such utter garbage that they have to rewrite the whole thing because that's easier than trying to make sense of the slop you produced under your own name.
If you were feeling overwhelmed by management, and Claude Code is alleviating that, I fear you aren't cut out for the work.
Sir! Elon has responded to our pressure tactic. Your interview seems to have had an effect. "Well - what did he say?" It's better if you see for yourself.
GIF reply "why are you gae" (this was his actual response btw)
This is great, but I wouldn't consider a food to be "clean" just from this testing.
At a minimum it needs glyphosate testing. I suspect the avocado oil has no plastics but high glyphosate, it's one of the many reasons I only use high-quality olive oil and coconut oil in cooking.
The way it works is we prompt the LLM to generate code to call the tools directly. We then parse the AST to validate it and restrict it to a very small/safe subset (whitelisted function calls.)
HF smolagents does something similar and there are a few papers out there validating the approach.
That makes sense because Cline existed prior to tool use being part of the standard LLM APIs. I think OpenAI had it at that point, but Anthropic and Ollama hasn't yet standardized on an API.
But the vendors are training their models specifically on tool use with their expected internal formats so you might as well leverage that if you are starting fresh now.
If you need a converter for a normal HTML/CSS component, my free extension SnipCSS is the best Tailwind converter. I haven't seen anything else come close.
https://www.snipcss.com
reply