In this episode, Patrick McKenzie (patio11) recorded with Zvi Mowshowitz (TheZvi) live at the LessOnline conference. They explore practical strategies for getting better results from large language models. Zvi explains how to customize AI behavior through thoughtful system prompts, while Patrick shares techniques for using LLMs as writing partners and research assistants. They discuss the evolving relationship between content creators and AI training data, touching on the emerging field of "generative engine optimization" (GEO). The conversation also covers multimodal capabilities, recursive AI use, and strategies for avoiding common failure modes like hallucination and sycophancy.
–
Full transcript: www.complexsystemspodcast.com/getting-better-at-llms-with-zvi-mowshowitz/

–

Sponsor:  Vanta

Vanta automates security compliance and builds trust, helping companies streamline ISO, SOC 2, and AI framework certifications. Learn more at https://vanta.com/complex

–

Links:

• Don’t Worry About the Vase https://thezvi.wordpress.com/ 

–

Timestamps:

(01:08) Understanding system prompts

(02:04) Customizing LLM behavior

(05:58) Memory features in LLMs

(10:21) Generative Engine Optimization (GEO)

(15:59) Sponsor: Vanta

(17:17) Art and AI: Enhancing creativity

(20:36) Recursive use of AIs

(25:22) Addressing LLM frustrations

(27:05) Checking for hallucinations in AI outputs

(28:11) Experimenting with AI models

(29:44) Optimizing AI prompts and outputs

(31:19) Using AI for writing and editing

(32:32) AI as a research and writing partner

(33:26) Prompting AI and humans effectively

(39:39) Balancing AI assistance with personal voice

(51:03) Wrap