Ask HN: Gemini Reliability Degrading?
I've been using Gemini (via the web interface) as my main 'daily driver' since 2.5 Pro came out. It's been super reliable until probably the past few days, where I am seeing a lot of slowness and 'errors', with chats not working or stopping after a few hundred tokens response.
I've also seen problems with some AI IDEs Gemini integrations recently too, which makes me suspect it is a capacity problem.
Curious if people are seeing this with bigger/more sophisticated workloads.
Wonder if Google are having serious scale problems, which probably would be the first time I've ever written that!
Could part of the issue be that Google recently switched a large number of Androids from using Google assistant to using Gemini? In terms of either sheer numbers of clients, or more mobile clients poisoning the well? Just guessing really. Not every vendor has pushed the update yet.
I've been using it since August '24 and it has never been reliable enough for production. My app from then is broken because it doesn't follow instructions, especially when you make it output JSON without some middle layer to "repair" the broken bits. There was a point where it simply broke all the JSON arrays and I had to restructure it to not use arrays.
It seems to work okay for something like Cursor, where you clearly realize that you're using a certain model and are willing to take on the risk for higher quality output.