Maybe n=1, but I disagree? I notice that Sonnet 4.6 follows instructions much better than 4.5 and it generates code much closer to our already in-place production code.
It's just a point release and it isn't a significant upgrade in terms of features or capabilities, but it works... better for me.
Are you using a tool like Claude Code or Codex or windsurf? I ask because I've found their ability to pull in relevant context improves tasks in exactly the way you're describing.
My own experience is that some things get better and some things get worse in perceived quality at the micro-level on each point release. i.e. 4.5->4.6
Haha yeah I've had this happen to me too (inside copilot on GitHub). I ask it to make a field nullable, and give it some pointers on how to implement that change.
It just decided halfway that, nah, removing the field altogether means you don't have to fix the fallout from making that thing nullable.
In my limited experience, that's mostly since the 4.6 release. I noticed that with the same prompt, it answers much more briefly. A bit jarring indeed, but I prefer it. Less bs and filler, and less burning off electricity for nothing.
Even (uncommon) country TLD's too. I own a .vg domain which is a perfect match with the initials of my last name. My mails end up in spam quite often too, despite having set up SPF, DKIM, DMARC and all that stuff correctly. It's just not common so some security systems block it.
I agree. On top of that, in true Google style, basic things just don't work.
Any time I upload an attachment, it just fails with something vague like "couldn't process file". Whether that's a simple .MD or .txt with less than 100 lines or a PDF. I tried making a gem today. It just wouldn't let me save it, with some vague error too.
I also tried having it read and write stuff to "my stuff" and Google drive. But it would consistently write but not be able to read from it again. Or would read one file from Google drive and ignore everything else.
Their models are seriously impressive. But as usual Google sucks at making them work well in real products.
I don't find that at all. At work, we've no access to the API, so we have to force feed a dozen (or more) documents, code and instruction prompts through the web interface upload interface. The only failures I've ever had in well over 300 sessions were due to connectivity issues, not interface failures.
Context window blowouts? All the time, but never document upload failures.
I'm talking about Gemini in the app and on the web. As well as AI studio. At work we go through Copilot, but there the agentic mode with Gemini isn't the best either.
What I love about Gemini mobile is that, if you look at the app wrong, it completely loses the response. It still generates it (and uses up your quota), but it never displays it!
This is the company that made Android, and it can't make an Android app that fetches a response from a server. Astonishing.
For me it honestly matches pretty well. I give it an instruction and go reply to an email, and when I'm back in my IDE I have work (that was done while I was doing something else) to review.
Going back from writing an email to working, versus going back from email to reviewing someone else's work feels harder.
Nah that's not how it works. Streaming video is usually cut up into small segments. By having a couple of variants per segment, they can serve you a unique and identifiable sequence of segments without having to decompress (and encrypt) them for each user.
reply