God finally showed up and said "ENOUGH!".
Doomsday.
90 tokens to midnight.
Qwen3.8-27B and Qwen3.6-35B-A3B are working from my machine. Anyone else?
Cloud is just other peoples computers - they can and will go down as well.<p>And even worse if its just a few computers run by a few people - as they will bring down many others depending on them.
I assumed it was an AWS outage, and AWS is experiencing problems, but Gemini is also experiencing outages and I assume Google is not using AWS for Gemini.<p>But, also, Claude has been working fine for me all morning.
Good opportunity for a Natural Experiment to study the productivity impact of LLMs.
Not just these three. OP mentions also Cloudflare, and additionally Downdetector also has AWS, Azure, and Google (both search and Gemini) listed as having spikes about the same time: <a href="https://downdetector.com/" rel="nofollow">https://downdetector.com/</a>
Time for Gemini 3.8 Flash to shine?!<p>It's much cheaper and has replaced Sonnet 5 for me.
Is this finally it?
Claude Code seems totally fine for me.
What's the single point of failure across providers?
Users perceiving the products as largely interchangeable and quickly DDoS'ing the other providers when one is down. So much for the possibility of a moat.
The entire industry runs on IOUs for compute and bills paid in cloud credits, could be anywhere. Could of course also be a a plain old DDoS.
They all have to rely on each other's LLMs to solve their own internal problems now.
Both clause and codex through bedrock seem to be working fine.
It's probably Azure infra that's the problem.
and thats why boys and gals you buy own gpus
Cannot replicate.
Gemini seems fine.
Good. Now to see what frauds are unable to work.
opus 5.0 is down, other models are working fine