the fact that claude is widely reputed as one of the best LLMs for coding indicates that they at least are focusing on the quality of their product for what their target markket wants, as they are producing a reasonably good product.

i also like to emphasise and mention Apple in the group of companies that are highly visible in this field, whose historical marketing gimmick is about conditioning their users to feel superior for their choice.

i'm just an engineer. what matters to me more is the tool does the job, reasonably well. i can't say the same about any of the others. i've used the previous latest model from google, and it was such a confused, retarded thing that i literally never saw it actually achieve anything. about an hour or two trying it and i was done. same with GPT-4o and GPT-5. also extremely unimpressive reasoning for my task.

perhaps my choice of language colors this. according to anthropic's announcement, Go is the least well supported language. but the circus that i watched with other models working with go tells me theh kind of reasoning that it is "weakest" at, is the kind that the others can't even do at all.