AI Letters #04 - The ML Skills That Didn’t Get Commoditized
· 3 min read
The LLM transition didn’t kill the ML engineer. It killed the ML engineer who was doing what LLMs now do for free. Embeddings, attention, fine-tuning, evals - the engineers who understood these at depth came out ahead. The ones operating at the abstraction layer that foundation models just ate did not.
This letter was first published on AI Letters →
Continue reading on Substack →
Interactive Data Stories
Interactive Chart
Domain Timeline →
How ML skills evolved through each major model release - from BERT to GPT-4 to today. Click any event to see what changed for engineers.
Interactive Chart
The Two Paradigms →
Commoditized vs still critical - click any skill to see exactly why it ended up in that column and what its demand curve looks like now.
Interactive Chart
The Evidence →
MTEB embedding benchmarks, Lost in the Middle accuracy drop, LoRA efficiency curves - the data behind every claim in this letter.