517: Plan First, Think Less: Save Tokens, Improve Code
About this episode
The episode discusses AI avatars, text-to-speech deepfakes, and provides development advice on LLM thinking modes and token costs.
Episode 517 starts with a light chat about AI avatars and new text‑to‑speech deepfakes before diving into LLM “thinking” modes—what baked‑in planning actually does, why it multiplies token costs, and when it helps or hurts. James and Frank give concrete dev advice: try low‑thinking settings, use big models for creative planning then smaller ones to execute, leverage harnesses/system prompts, and beware quantized local models often do better without thinking. Follow Us Frank: Twitter , Blog , GitHub James: Twitter , Blog , GitHub Merge Conflict: Twitter , Facebook , Website , Chat on Discord Music : Amethyst Seer - Citrine by Adventureface ⭐⭐ Review Us ⭐⭐ Machine transcription available on http://mergeconflict.fm Support Merge Conflict
Topics covered
- AI avatars
- text-to-speech deepfakes
- LLM thinking modes
- token costs
- development advice
- creative planning
- local models
More episodes of Merge Conflict
- 516: Evolving Agent Session Management
- 515: Mini-LED, IPS, and the Great Monitor Rabbit Hole
- 514: Running Local LLMs in VS Code
Explore listener stats, chart rankings, contacts and more on the Merge Conflict podcast page.