The OP of the thread is pointing out that Google's "gemini-2.5-pro-preview-03-25" model was "suddenly and silently" redirecting to "gemini-2.5-pro-preview-05-06", confusing and frustrating "developers, benchmark maintainers, and researchers" who thought they were using one model but were actually being served another.
The current HN link is to page 3 of that discussion, of which I assume the highlighted comment is from this user who provides a possible reason for that redirection:
> It literally didn`t handhold and gave critical honest feedback, even under wrong assumptions. Where all other models were just parroting what they thought the user wants to hear.
karencarits · 3h ago
Sorry, I wasn't aware the page parameter when pasting the link so don't read anything into that, I have not personally posted in that thread. But I strongly support H_Express in his case that dated models should not suddenly be redirected to another endpoint
globie · 3h ago
It's a strange thing Discus does loading only a portion of discussion, as if it's prohibitively expensive to pull out a few extra kb to load the entire text of the thread at once.
I do enjoy seeing a Google page load about 30 scripts from AWS Cloudfront, though :)
The current HN link is to page 3 of that discussion, of which I assume the highlighted comment is from this user who provides a possible reason for that redirection:
> It literally didn`t handhold and gave critical honest feedback, even under wrong assumptions. Where all other models were just parroting what they thought the user wants to hear.
I do enjoy seeing a Google page load about 30 scripts from AWS Cloudfront, though :)