Running a fully local AI podcast station with Qwen
sysadmin420 · reddit · 2026-08-18
An author shared a project running a podcast and YouTube/Spotify channel entirely on local models.
- Core Model: Writing is handled by qwen2.5-72b (MoE, 3B active params) on Ollama (RTX 3090). It is twice as fast and hallucinates less than a dense 35B model.
- Vision: qwen2-vl parses image-only posts from Truth Social.
- Audio: Voices are cloned using F5-TTS.
- Pipeline: Monitors news, stocks, crypto, and social media, generating scripts based on market vs. statement analysis.
The author highlighted hallucination control strategies: a checker discards entire shows if numbers are invented; unverified figures are replaced with placeholders rather than guesses. Hardware is AMD 5950X + 3090.
More from Models
- Prediction: frontier AI revenue growth could slow even as AI usage keeps rising — yi111 · 2026-10-16
- Meta teased 816B-param open-weight Muse Spark Open with 1M context, says meme thread — alexandr_wang · 2026-10-16
- Nolla Health Post-Trains Nemotron 3.5 VL: Visual Diagnosis Jumps from 35% to 66.8% — LoganGrasby · 2026-10-16
- Dev argues coding models now beat nearly all engineers in every branch of software — zack_overflow · 2026-10-16
- Prediction: Chinese frontier models could match GPT-6 by March 2027 — teortaxesTex · 2026-10-16
- Liquid AI's d1-ultra, #1 on Artificial Analysis Decision Index, lands on Vercel — JosephJacks_ · 2026-10-16