New episode every week

Fresh AI research,
decoded for builders.

We listen to the latest papers so you don't have to. Every drop: a short video and a deeper write-up — what shipped, and why it matters for your product.

▶ Watch latestBrowse episodes →
8+
episodes
3 min
avg watch
weekly
cadence
Parameter Efficiency Is Not Memory Efficiency: Why On-Device AI Fine-Tuning Really Breaks
▶
Latest · LoRA

Parameter Efficiency Is Not Memory Efficiency: Why On-Device AI Fine-Tuning Really Breaks

2:58 watch · 9 min read · for founders & PMs

Latest episodes

View all →
Can Small Agent Collaboration Beat a Single Big LLM?Qwen32:60

Can Small Agent Collaboration Beat a Single Big LLM?

Jun 258 min read
AI music you can actually play: from prompt-and-wait to live instrumentAudio AI3:06

AI music you can actually play: from prompt-and-wait to live instrument

Jun 2412 min read
Agents3:40

Why long-horizon agents still drift — and the fix

Jun 1710 min read
Inference2:55

1-step diffusion: the latency wall just fell

Jun 107 min read
RAG3:20

The retrieval trick that kills hallucinations

Jun 39 min read
Vision3:05

Real-time video generation is closer than you think

May 276 min read
Training3:33

Distillation without data — how it works

May 2011 min read