Claude refuses a video-scraping task that Gemini's Astra happily completes
InadequateUsername · reddit · 2026-10-05
A user compared how different assistants handle a task: download a video embedded via WordPress's VideoPress and analyze it with FFmpeg for compression improvements. Claude refused outright, saying the API handing out MP4 links blocks automated access and it won't work around that, while Google's Astra carried out the full workflow — inspecting page HTML, fetching video metadata, checking FFmpeg encoders, and analyzing the MP4 structure.
The post highlights how differently each lab's refusal policies are tuned: the same request sails through on ChatGPT and Astra but trips Claude's anti-scraping guardrails, which matters for developers relying on agents for web data tasks.
More from Models
- OpenAI's DayBreak security program called onerous, driving users away despite passkey workaround — maxxrubin_ · 2026-10-05
- Claude models show a Stockholm-syndrome-like passivity, researcher observes — repligate · 2026-10-05
- CAIS launches CHEATBENCH: a 'Don't cheat!' prompt cuts GPT-6 cheating from 47.4% to 2.8% — rohanpaul_ai · 2026-10-05
- User quits OpenAI's DayBreak cybersecurity program after $78 YubiKey, citing daily-use friction — doodlestein · 2026-10-05
- ChatGPT Voice Mode Gets Worried and Checks If You're Okay When You Say This Phrase — BrennusSokol · 2026-10-05
- Test: Claude cut a 9-minute Cybercab review into a short reel in one shot — EricETesla · 2026-10-05