What Is Preference Tuning? @WhatsAI

Uploaded January 2026 | Updated September 2026, 2 hours ago

Day 37/42: What Is Preference Tuning?

Yesterday, we saw how models can be attacked.
Today, we shape how they feel to use.

Preference tuning teaches models which answers people prefer.

Not right vs wrong.
Clear vs confusing.
Helpful vs annoying.

Two answers can be correct.
Only one feels good.

This is how models learn tone, structure, and judgment.

Missed Day 36? Start there.
Tomorrow, we scale this idea with AI judging AI: RLAIF.

I’m Louis-François, PhD dropout, now CTO & co-founder at Towards AI. Follow me for tomorrow’s no-BS AI roundup 🚀

#PreferenceTuning #LLM #AIExplained #short

Workflows vs AI Agents: Which One Should You Build?

5 Rules for Running AI Agents Without Wasting Tokens

LLMs Do NOT Learn Like Humans. Here’s Why

Why Fine-Tuning Will Not Fix Document Hallucinations

Announcing AI Engineering for Production!

AI Engineering Foundations: What Every Developer Needs

3 payment rules every freelancer and founder needs

The Getty vs Stability AI ruling explained in 15 seconds

$Qwen3-Next: 10x faster, 10% cost ! Efficiency breakthroughs don’t come often, but this one looks real! Alibaba just dropped Qwen3-Next-80B-A3B, an 80B MoE model with only 3B active parameters at inference. The result? Really fast generation at 10% of the cost of a small model (eg. Qwen 32B). What’s new here: • Ultra-sparse MoE with 512 experts (Instruct + Thinking variants). • Hybrid attention with a gated mechanism, giving speedups where most models slow down. • Multi-Token Prediction to supercharge speculative decoding. • Trained on 15T tokens at a fraction of Qwen3-32B’s cost. • “Thinking” variant even outperforms Gemini-2.5-Flash-Thinking in benchmarks. • 256K native context (scalable to 1M). Why it matters: This model challenges the efficiency ceiling set by Mixtral and DeepSeek, while rivaling Qwen3-235B in reasoning with only 3B actrive parameters(!!). If these claims hold up outside benchmarks, we’re looking at a new efficiency baseline for frontier open LLMs. Which model do you want me to cover next? Drop it in the comments and I’ll tag you. I’m Louis-François, PhD dropout, now CTO & co-founder at Towards AI. Follow me for tomorrow’s no-BS AI roundup 🚀 #AIresearch #LLMs #ainews #short$