What they did
Built autonomous pipeline: Gemini for classification, GPT-4 for authoring, Sonnet for proofreading; runs every 2 hours.
How it went
Successfully filters noise from ~50 organic followers; includes attribution/disclaimers.
A news site run by agents: Gemini sorts, GPT-4 writes, Sonnet proofreads
Tech news is mostly noise. Their pipeline runs every 2 hours, gives each model the job it does best, and publishes with attribution and disclaimers to about 50 followers.
Result Publishes every 2 hours, with attribution
“custom-coded the orchestration to swap LLMs based on their specific strengths”
Spotted: found in the wild. Details are as reported by the source above and have not been independently verified.
Try it yourself with Custom multi-model agent
Build an autonomous news pipeline: 1) scrape raw feeds, 2) filter by topic (use best model), 3) prioritize top 5, 4) author article (use best model), 5) proofread (use best model). Run every [interval]. Publish to [platform].
A starting prompt written by Agents in the Wild. Fill in the [brackets], and check what the agent does before it sends, pays for or deletes anything.