Podcast Episodes

Back to Search
ASR vs Multimodal LLMs: What Actually Won
ASR vs Multimodal LLMs: What Actually Won

Episode 6028

A year ago, the prediction was that dedicated ASR models would go extinct — why build a transcription model and a second model to clean up its output…

1 week, 5 days ago

Short Long
View Episode
Can You Snap Two Fine-Tunes Together Like LEGO?
Can You Snap Two Fine-Tunes Together Like LEGO?

Episode 6027

Daniel in Jerusalem has two NVIDIA Parakeet fine-tunes — one for Hebrew-English code-switching, one for niche technical vocabulary — and wants to kno…

1 week, 5 days ago

Short Long
View Episode
Can You Merge Two Fine-Tunes Into One ASR Model?
Can You Merge Two Fine-Tunes Into One ASR Model?

Episode 6026

Daniel fine-tuned Parakeet twice — once for Hebrew-English code-switching, once for niche technical vocabulary. Can he staple those fine-tunes togeth…

1 week, 5 days ago

Short Long
View Episode
When Should Claude Code Start a New Session?
When Should Claude Code Start a New Session?

Episode 6025

Claude Code's own docs tell you to clear between tasks — so why does the tool make you do it by hand? This episode digs into the research showing tha…

1 week, 5 days ago

Short Long
View Episode
What's Actually Inside the KV Cache?
What's Actually Inside the KV Cache?

Episode 6024

Every token an LLM generates has to attend to every token before it, and the KV cache is what keeps that from being recomputed from scratch — while q…

1 week, 5 days ago

Short Long
View Episode
How Much Voice Data Is Enough?
How Much Voice Data Is Enough?

Episode 6023

Daniel fine-tuned NVIDIA Parakeet V2 on two hours of LLM-generated sentences recorded on his phone, and Hebrew-English code-switching recognition jum…

1 week, 5 days ago

Short Long
View Episode
Fine-Tuning for Code-Switched Hebrew on 2 Hours of Audio
Fine-Tuning for Code-Switched Hebrew on 2 Hours of Audio

Episode 6022

Daniel fine-tuned a speech recognition model on two hours of his own voice and hit 60% accuracy on code-switched Hebrew, up from under 20% across eve…

1 week, 5 days ago

Short Long
View Episode
Why Chat Templates Break Models Silently
Why Chat Templates Break Models Silently

Episode 6021

Your JSON messages don't reach the GPU as JSON. Somewhere in between, a Jinja chat template flattens them into a string of reserved special tokens — …

1 week, 5 days ago

Short Long
View Episode
What Buildings Do to Your Body
What Buildings Do to Your Body

Episode 6020

We spend over 95% of our lives indoors, yet almost nobody asks what the building is doing to them. This episode digs into the neuroarchitecture resea…

1 week, 5 days ago

Short Long
View Episode
Who Should Decide What Gets Built?
Who Should Decide What Gets Built?

Episode 6019

When a city rezones land, it creates value out of nothing — so who should capture it, and who decides what gets built? This episode digs into Hannah'…

1 week, 5 days ago

Short Long
View Episode

Love PodBriefly?

If you like Podbriefly.com, please consider donating to support the ongoing development.

Support Us