Stop Blaming LLMs: ZeroEntropy’s Founder on Killing AI Hallucinations & Building ‘Impossible’ Search

Revolutionizing AI Search: Ghita Alami's Vision for Zero Entropy

In this episode, we chat with Ghita Alami, founder of Zero Entropy, about her company's groundbreaking approach to rebuilding search systems for AI agents. Learn about early chaos theory, tactical product insights, and the future of intelligent retrieval systems.

Most retrieval systems today only hit ~60–70% recall. Zero Entropy is pushing that above 90%, giving LLMs the relevant data they actually need to answer correctly — and cutting hallucinations at the root.

We get into:
• Why search is still broken — and how Zero Entropy is fixing it
• How poor retrieval fuels LLM hallucinations
• Real-world use cases in healthcare, legal, compliance & more
• Lessons from YC, building deeptech, and structuring high-performance co-founder teams
• Contrarian bets on small models, voice interfaces & the path to AGI

If you’re building with RAG, embedding-based search, or deploying LLMs in production — this is required listening.

Watch now and subscribe for more fast, founder-first conversations at the edge of AI and infrastructure.

00:00 Introduction to AI Retrieval Systems
00:07 Meet the Founder of Zero Entropy
00:43 Building a High Accuracy Search Engine
01:17 Challenges and Solutions in AI Retrieval
02:34 Specific Use Cases and Examples
08:09 Evaluating and Improving AI Systems
12:02 YC Experience and Building Zero Entropy
20:35 Technical Deep Dive: Measuring Accuracy
22:50 Challenges of Internet Search
23:33 Evaluating Retrieval Systems
24:02 Improving Recall and Precision
25:03 Advice for Founders on Search Issues
26:11 Leveraging LLMs in Tech Stacks
27:28 Staying Relevant in a Rapidly Changing AI Landscape

🔗 LINKS
Zero Entropy → https://zeroentropy.dev

🎧 Listen on the go:
Spotify: https://open.spotify.com/show/1u1JyRK...
Apple: https://apple.co/4cZ8RMm