Roshan_Roy

I wonder if the deeper issue isn’t just “AI is too agreeable”, but that most advice (AI or human) doesn’t actually translate into action. A lot of people aren’t really looking for accurate feedback, they’re looking for something that feels coherent enough to sit with. Reddit gives extreme answers, AI gives agreeable ones, but in both cases the outcome is often the same: no real change in behavior. That might be why this feels worse with AI, it removes the friction you’d normally get from another human pushing back.

I think the article works better as a mental model than a literal claim. “Linux is an interpreter” feels wrong if you define interpretation strictly at the CPU instruction level, but it becomes more reasonable if you look at the kernel as something that interprets executable formats and environments (ELF, scripts with shebangs, initramfs, etc.). In that sense it’s less about instruction-by-instruction interpretation and more about orchestrating how different representations of programs become runnable. Maybe the confusion here is mixing those two meanings of “interpreter”.

This feels like one of those “because we can” projects that accidentally reveals where the platform is going.

CSS started as purely declarative styling, but between things like conditionals, math functions, and now these rendering tricks, it’s slowly creeping into “programmable system” territory. Not because it’s the right tool for it, but because browsers are becoming the real runtime. The interesting question isn’t “can Doom run in CSS”, it’s how much logic we’ll keep pushing into layers that were never meant to handle it.

This is a really interesting setup — especially the split between the public and private agents. curious about the IRC choice: was that mainly for simplicity and reliability, or did you find advantages over something like a lightweight HTTP/WebSocket layer? Also, how are you handling state between the two agents, is it mostly stateless requests over A2A, or do you maintain some shared context?