alt Hacker NewsNot offended at all, but just ranting about how someone is a shill instead of responding to the substance of their argument is simply not the kind of discussion we have on HN. Read the guidelines.
> "no, model performance is not degraded because I say so" serve as correcting misinformation?
Because zero evidence has been provided other than feelings. That is not evidence of degradation, and we know they don't serve quants.
You are an Anthropic shill, and this is an explicit marker that needs to be added to all of your comments, so that all information you provide can be adjusted for that bias. But I do understand why you ignore this point since it devalues all your comments (as it should), and instead cling to "ranting how someone is a shill bla-bla-bla".
Those people, unlike you, are actually using AI in development. And it is not a singular person who reports their frustration with the model being degraded after a certain period of time, so the anecdata does gradually become data. Your attempts at gaslighting are weak, you should really ask your bosses for a new guidebook on how to deal with reports of models performing at worse levels than before. Just writing "because I say so" is not cutting it.
> "we know they don't serve quants"
How do you know that unless you are working at Antrhopic? Yet another evidence of you being an Anthropic shill.