Is fine-tuning still worth it?

I built a “fine-tuned” model at OLX about 3 years ago. Foundation models have evolved a lot, and so have RAG and other “tools”. As far as I know, fine-tuning a model is no longer worth it.

I’m writing this because I want to hear other people’s experiences. And what I’ve read out there confirms what I’m seeing. I think using other techniques is much more interesting.

Some texts on the subject:

• https://lnkd.in/gk8x3jkF
• https://lnkd.in/gUUApt5E

I really like the phrase “fine-tuning isn’t injecting knowledge, it’s overwriting”.

Links

• Fine-Tuning or Fine-Failing? Debunking Performance Myths in Large Language Models

Originally written in Portuguese and published on LinkedIn on August 21, 2025. I translated it into English with the help of AI (Claude): it let me bring my LinkedIn posts to people who don’t read Portuguese without rewriting each one from scratch. I reviewed the translation to keep my original meaning and tone.