Season 2 · Episode 1907
Why We Still Fine-Tune in 2026
Despite million-token context windows, fine-tuning remains essential. Here’s why behavior, not just facts, matters.
My Weird Prompts · Daniel Rosehill
April 2, 202629m 23s
Audio is streamed directly from the publisher (dts.podtrac.com) as published in their RSS feed. Play Podcasts does not host this file. Rights-holders can request removal through the copyright & takedown page.
Show Notes
In an era of massive context windows, why are companies still fine-tuning models? This episode explores the shift from teaching facts to shaping behavior. We discuss domain expertise, style alignment, and Text-to-SQL optimization, plus how Parameter-Efficient Fine-Tuning (PEFT) makes it accessible. Learn why fine-tuning creates specialized "neural highways" that outperform general models in production.