Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude outage

status.claude.com

Post reply on HN
Print view View original

131–140 of 204 posts

Re: Claude outage

#131

Post by SXX »

jgalt212 wrote: post #118
iberator wrote: post #116

Earlier quoted context omitted.

Gemini have 100% uptime and zero limits of any kind in their free models :) Highly recommended

> Try Gemini to see how bad it can really get. Most of the time 2.5 pro requests fail for unknown reasons over the App. Another commenter lives in a different world than you.

I 99% sure they are talking of using it via AI Studio. It is very reliable.

Re: Claude outage

#132

Post by cpursley »

siva7 wrote: post #98
CapsAdmin wrote: post #88

I didn't even know they had a status page. Claude (with pro subscription) is often so unreliable with regards to connectivity and performance that I'm looking for something more predictable. It randomly fails halfway through a response, sometimes very slow to start, hangs for long periods during a response, and so on. The Claude chat interface can also slow down with long sessions. I sometimes use Claude code which i…

Try Gemini to see how bad it can really get. Most of the time 2.5 pro requests fail for unknown reasons over the App. Claude and Chatgpt are way more reliable.

Gemini is embarrassingly bad. It outright doesn’t work. I mean, it actually goes out and does stuff but it’s 100% of the time random. Even third-party forks of it work better (like Qwen Code), which is just wild.

Re: Claude outage

#136

Post by brabel »

trvz wrote: post #99

curl -fsSL https://ollama.com/install.sh | sh ollama run smollm2:135m-instruct-q2_K --verbose

Use llama.cpp or llamafile , ollama just adds weight and does too much magic.

Re: Claude outage

#138

Post by brabel »

roman_soldier wrote: post #100
frotaur wrote: post #97

Earlier quoted context omitted.

[flagged]

Yes because anyone offering a positive opinion on Grok or one of Elon's companies must have been bought.

Isn’t it funny how people think? Say anything about Palestine not deserving being flattened and you are now anti-Semitic and a Nazi. Mention how NATO is not as innocent and harmless a a fly and perhaps being on the other side of their friendly border is definitely reason for concern and you are now a Putin puppet. You know, I lived in a dictatorship briefly. That’s exactly how thinking goes when they reason about opposition and free thought.

Re: Claude outage

#139

Post by siva7 »

SXX wrote: post #128
siva7 wrote: post #98

Earlier quoted context omitted.

Try Gemini to see how bad it can really get. Most of the time 2.5 pro requests fail for unknown reasons over the App. Claude and Chatgpt are way more reliable.

It almost never fails via AI Studio though. Also I doubt fails you see really have anything to do with LLM itself, capacity or backend. It's just Google own UIs and apps are almost comically bad.

I use Gemini over web app and mobile app. Both are very unreliable. Anthropic and openai don't have more resources than google but still get it right most of the time - the quality of product development is not even in a similar league
Post reply on HN