NVFP4

topic on 3 shows · 2 statements across 2 episodes · said 8 times in 3 episodes since 2026

the MAD Podcast 5 Latent Space 2 the Y Combinator Startup Podcast 1

Mentions by year, every show

tap a year for its mentions
0042832026episodesmentions
0232026episodes it came up in
001.51.5332026episodesmentions per episode

the MAD Podcast 5Latent Space 2the Y Combinator Startup Podcast 1

2026 8 mentions in 3 episodes 3 per episode

every mention on every show, scene by scene, with the transcript →

2 statements about NVFP4, every show

Quantization, speculative decoding, and prefill-decode disaggregation each yield 2x speedups
“Going from BF-sixteen to NVFP four is, it's not quite a two X, right? It's like, I think it's about like 30, 30 to 40% from 16 to eight, and then another 30 to 40% multiplied from eight to four. So that doesn't quite get you a two X, but like roughly a two X. …”
Philip Kiely Aug 3, 2026 ▶ 40:03 Next 100x in AI: Inference, Networking, & Self-Optimizing Models — Philip Kiely & Ali Taha, Baseten
MAD Assertion Supported
NVIDIA pre-trained Nemotron Ultra and Super natively using 4-bit floating point
“NemoTron Ultra and Super, ah, were pre-trained using four-bit arithmetic. We pre-trained those in MVFP four”
Bryan Catanzaro Jul 2, 2026 ▶ 35:41 Inside Nemotron & NVIDIA’s AI Lab | Bryan Catanzaro

← every entity, every show

Made with StarZero

Turn any episode into a week of clips.

This entire site, thousands of episodes across every show transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.