May I suggest that you try to acquaint yourself with the concept of the desktop, which is what we are talking about. Windows accounts for around 96% of desktop operating systems. So in that respect Linux is nowhere. May I suggest that you try to acquaint yourself with the concept of the desktop, which is what we are talking about. Windows accounts for around 96% of desktop operating systems. So in that respect Linux is nowhere. Latest 2026 beachmark Windows: ~60%–73% globally macOS: ~15%–16%. Linux: ~3%–4.7% Chrome OS: ~1%–4% So you are correct Do you really not understand that Mac is a different platform? We are talking about PCs, which is what Windows, Linux and Haiku are designed to run on. I think you are just arguing for the sake of it, which is what you were doing in the discussion about Vista. Give it a rest. Please. What are you on about? Mac OS is a desktop OS for PCs. Always has been. Also, how is this relevant to the topic at hand? If you’re just going to stay off-topic the whole time just give this thread a rest. this thread is so off topic at this point Truely, the issue is not the LLM-generated code. There was a time when spaghetti code was an issue. You should ask: The issue for you may not be with LLM-generated code, but this statement does not reflect the broad range of objections that people have posted here; many of them have more to do with the LLM part than the code part. A (probably incomplete) list would have to include: (Not everyone who objects to LLM code being used in Haiku or publicized on haiku’s forums views all of these issues as significant or even real, but I tried to list all the ones I’ve seen here.) I’m sure I’ve missed at least one there, but that’s a much broader set of objections than just spaghetti code. The issue for you may not be with LLM-generated code, but this statement does not reflect the broad range of objections that people have posted here; many of them have more to do with the LLM part than the code part. Agree on the legalese, and I object to certain use cases of AI/LLM (i.e. not all). Kinda like disliking business practices of laying off 1,000 people for one computer. Makes sense on the spreadsheets, but the results can tank an entire town or city (or maybe the business).. I think the gatekeepers mentioned the kernel code submission limitations versus application development… I thought maybe your point was, spaghetti code was a tangible attribute that could be gauged in the object itself, where LLM origin is like “code written on a Monday” Which depends on me to volunteer the information that I made some changes to the code on a Monday. So, sure, it isn’t that bad - in April 2026, there are some tangible attributes that could be spotted in code that was substantially generated that way, so that’s part of the issue today: whether to allow code that bears evidence of assistance from April 2026 AIs, along with voluntarily reported assistance. I’m not saying that’s a silly policy direction, it’s just something that seems likely to evolve pretty fast.. Keep in mind that most of that number is the corporate and institutional world. If you exclude corporations, businesses, schools, hospitals etc.. And only focus on home users (our market) The playing field shifts towards more of a balance between M$ and Apple Anyway, your ability to go completely off-topic is absolutely fantastic
TB – super‑fast text‑only browser using Jina Reader API