@AnthropicAI@claudeai Fable 5 fallback seems to over-trigger on defensive product QA language like “manifest”, “license”, “runtime values” and “deployment review”. I get why the safety gate exists, but the UX would be better with intent-aware scoring: distinguish exploit construction from product reliability planning, then ask for clarification instead of silently downgrading.
cc @OpenAI Hoping you take care of this in GPT-5.6 Ultra
@MartinGTobias Similar experience. When Claude reviews code written by codex it rarely finds critical issues but codex almost every time finds high sev issues in codes authored by Claude. Codex as a harness is looking better
While the AI hype consumes everything... HTTP just got its first new verb since 2010: QUERY (RFC 10008).
Safe + idempotent like GET but with a request body. Perfect for complex searches:
“QUERY /users HTTP/1.1
Content-Type: application/json
{"filter": {"age": {"gt": 30}}, "sort": "name"}”
No more URL stuffing or deliberate-POST reads. We almost missed this one.
#HTTP #WebDev
1/
Codex is quietly killing your SSD.
It writes diagnostic logs to disk non-stop, even when you're not
doing anything. Your SSD has a write limit. Codex is burning through it in the background.
One command fixes it 👇